AI Robots with Purpose with Jake Loosararian of Gecko Robotics | E1947

01:13:37
https://www.youtube.com/watch?v=7PwNdz16MGE

Summary

TLDRThe video features a conversation about the use of robotics and artificial intelligence in solving current industrial issues. The key point is that developing specific robots for particular tasks can build the right business model and foundation to eventually develop more advanced robots, like humanoids. However, the use of AI often falters due to grounding on false or incorrect models which don't produce expected results. Business models should therefore clarify and prove the value of integrating advanced technologies into operational strategies, making it compelling for corporate executives to invest in them. The accurate and practical use of data is essential to success, and ground truth checks must support AI implementations to truly benefit businesses.

Takeaways

  • 🤖 Building robots for specific tasks can earn the right to pursue more ambitious robotic projects.
  • 🚀 There's a desire to innovate with autonomous technology despite the current existing problems.
  • 💡 Business models need to justify AI and robotic integration by demonstrating clear value.
  • 😱 Current AI solutions often fail because they rely on flawed data, leading to unfulfilled promises.
  • 📊 Data integrity and correct ground truth are critical for AI success in industry applications.
  • 🎯 Industry needs to understand and use AI effectively to ensure sustainable, productive innovation.
  • 📉 Many AI contracts end unsuccessfully due to a lack of concrete improvement results.
  • 🏗 Large industrial sectors are slow to adopt new tech due to complexity and potential disruption.
  • 🧑‍💼 Making CEOs and CFOs see the value in robot integration is crucial for change.
  • 👷‍♂️ Understanding the health of existing infrastructure is key to preventing costly failures.

Timeline

  • 00:00:00 - 00:05:00

    Building innovative robots is appealing, but there are issues with business models that don't prove value, leading to contract expirations. Geco Robotics, led by CEO Jay Lucari, focuses on solving current issues by creating specific purpose robots, emphasizing the importance of understanding the health of infrastructure.

  • 00:05:00 - 00:10:00

    Jay Lucari started his company after recognizing issues in infrastructure maintenance at a power plant, where dangerous jobs were required to prevent shutdowns. He founded Geco Robotics to build wall-climbing robots for critical infrastructure inspections, enabling safer and more intelligent infrastructure management.

  • 00:10:00 - 00:15:00

    The critical need for infrastructural maintenance in the U.S is discussed, highlighting its significant costs and the lack of data regarding infrastructure health. Geco’s robots are engineered for specific inspection tasks, making informed predictions about structural integrity, especially for facilities not designed for modern demands.

  • 00:15:00 - 00:20:00

    The discussion on Geco’s robots' capabilities to inspect various critical infrastructures continues. Their robots do the work once performed by humans, offering a safer and more detailed data collection method, like using ultrasonic sensors to predict structural weaknesses, reducing risk and improving efficiency.

  • 00:20:00 - 00:25:00

    Vanta makes regulatory compliance achievable faster, allowing businesses to focus on essential functions without hindrance. Geco Robotics doesn't just focus on selling robots but extends their lifespan and usefulness, stressing efficient operation and reduced environmental impact to avoid catastrophic failures.

  • 00:25:00 - 00:30:00

    Vanta simplifies regulatory demands. Geco’s robots perform inspections, assessing structural integrity and overseeing repair needs more comprehensively than traditional human inspections, saving businesses from potential catastrophic failures and offering strategic business models beyond direct equipment sales.

  • 00:30:00 - 00:35:00

    Jay discusses the importance of transitioning from selling robots to providing comprehensive maintenance solutions, emphasizing software development that tailors solutions to specific organizational needs, enhancing operational efficiency and resolving unique customer demands.

  • 00:35:00 - 00:40:00

    Jay shares the importance of understanding customer problems and optimizing robots' utility through embedded customer interactions. This approach emphasizes the importance of distribution through existing contracts, solving larger client problems by integrating solutions with positive business outcomes.

  • 00:40:00 - 00:45:00

    Creating digital twins enhances infrastructure management by assessing structural health and extending asset life. Geco Robotics collects extensive data, enabling prediction and efficient asset operation. The conversation extends to developing integration with military infrastructures, ensuring integrity and longtime viability.

  • 00:45:00 - 00:50:00

    Jay explains using sensors to establish digital twins, allowing real-time monitoring of potential structural failures in infrastructure. These digital twins enable informed repair decisions, prolonging infrastructure life, and decreasing costly downtimes for clients while ensuring continued operational integrity.

  • 00:50:00 - 00:55:00

    Robotics and AI are explored regarding structural asset management, emphasizing the importance of integrating new robotics solutions with traditional systems to mainstream operational management. Emis reduction through predictive maintenance and optimized operations is highlighted, showing environmental benefits.

  • 00:55:00 - 01:00:00

    Jay discusses developing AI insights from vast data collections, which helps predict structural failures, optimize efficiency, and inform manufacturers on how to build better infrastructure. The importance of real-world damage analysis and accurate data integration is a focus area for enhancing operational sustainability.

  • 01:00:00 - 01:05:00

    The conversation covers reducing overhead with high-value robotics and continuous monitoring, underlining the future potential for AI analysis within infrastructure environments, possibly informing insurance and manufacturing improvements. The approach advocates for a broader acceptance and integration.

  • 01:05:00 - 01:13:37

    Robotics in hazardous environments like deep sea welding could cut risks and costs. Business model hypotheticals consider robotics' uptime value and operational outcome focus, looking to maximize efficacy in improving productivity and infrastructure assessment.

Show more

Mind Map

Video Q&A

  • What are the challenges of using AI in industry according to the discussion?

    The challenges include inaccurate data and models not grounded in reality, leading to AI solutions failing to produce results.

  • Why is there a sex appeal to building new technologies?

    There's excitement around innovative technology like autonomous robots and the potential they have to solve future problems.

  • How must business models adapt according to the discussion?

    Business models must convince executives, such as CEOs and CFOs, about the usefulness and integration of robotics and AI into Industry 4.0 measures.

  • What often happens when AI companies intervene in business operations?

    AI companies promise radical changes but often fail as their contracts expire without delivering significant improvements due to flawed insights from bad data.

View more video summaries

Get instant access to free YouTube video summaries powered by AI!
Subtitles
en
Auto Scroll:
  • 00:00:00
    there's so much sex appeal to building
  • 00:00:02
    new things and like in 10 years we'll
  • 00:00:04
    have this really cool new autonomous
  • 00:00:06
    thing drone walking you know humanoid
  • 00:00:08
    that's going to solve all these problems
  • 00:00:10
    but the problem is there's a lot of
  • 00:00:11
    issues going on today and so the
  • 00:00:13
    approach to solving and using you know
  • 00:00:15
    specific robots for specific jobs is
  • 00:00:17
    actually just to earn the right to begin
  • 00:00:19
    building really cool robots that are
  • 00:00:20
    able to do like you know more
  • 00:00:22
    interesting things but you got to get
  • 00:00:22
    the business model right and the
  • 00:00:24
    business model has to incentivize and
  • 00:00:26
    make a CEO or CFO give a about you
  • 00:00:28
    know how useful these industry you know
  • 00:00:31
    4.0 principles and tools are cuz right
  • 00:00:33
    now that's not true I see this time and
  • 00:00:35
    time again where like I won't name like
  • 00:00:37
    the AI companies but like these AI
  • 00:00:39
    companies come in and say we'll like
  • 00:00:40
    completely turn on your head the way
  • 00:00:41
    you're operating you know your entire
  • 00:00:43
    business and they'll come in for some
  • 00:00:44
    contract that ends up expiring because
  • 00:00:46
    it just did not produce and that's the
  • 00:00:48
    problem you think you have all the
  • 00:00:49
    information and data but you're building
  • 00:00:50
    your AI and your Solutions off of ground
  • 00:00:53
    truth that's actually not ground truth
  • 00:00:56
    this weekend startups is brought to you
  • 00:00:58
    by vanta compliance and security
  • 00:01:01
    shouldn't be a deal breaker for startups
  • 00:01:04
    to win new business vanta makes it easy
  • 00:01:07
    for companies to get a sock 2 report
  • 00:01:09
    fast twist listeners can get $1,000 off
  • 00:01:13
    for a limited time at
  • 00:01:15
    v.com twist Dev Squad most Dev agencies
  • 00:01:19
    only offer developers why because
  • 00:01:23
    product management is hard get an entire
  • 00:01:26
    product team for the cost of one US
  • 00:01:28
    developer plus 10% off at Dev squad.com
  • 00:01:32
    twist and HubSpot podcast Network
  • 00:01:36
    HubSpot podcast Network presents a new
  • 00:01:38
    podcast called the next wave the next
  • 00:01:41
    wave is your personal Chief AI officer
  • 00:01:43
    bringing fresh takes industry insights
  • 00:01:46
    and a trustworthy perspective on how to
  • 00:01:48
    implement AI to grow your business just
  • 00:01:51
    search for the next wave on YouTube or
  • 00:01:54
    in your favorite podcast app that's the
  • 00:01:57
    next wave all right welcome back to
  • 00:01:59
    another episode this week in startups we
  • 00:02:02
    like to talk about Innovation here and
  • 00:02:04
    AI has been on everybody's Minds for the
  • 00:02:06
    last two years you know it's it's simply
  • 00:02:08
    brilliant What AI can do and we see it
  • 00:02:10
    improving every week of course there has
  • 00:02:13
    been this dread of oh my God what if AI
  • 00:02:17
    plus robotics gets put together and we
  • 00:02:20
    have the Terminator films the truth is
  • 00:02:22
    autonomous robots are coming and they
  • 00:02:24
    will have ai built in you've probably
  • 00:02:26
    seen figure or what elon's working on
  • 00:02:28
    with Optimus over a test this is going
  • 00:02:30
    to change the world in my belief and
  • 00:02:33
    today we have a company that's been
  • 00:02:35
    working on it for a little while it's
  • 00:02:37
    called geco robotics we have the CEO
  • 00:02:39
    here his name is Jay
  • 00:02:42
    lucari so uh tell me about the robots
  • 00:02:45
    you're building and for those of you not
  • 00:02:48
    watching this weekend startups on
  • 00:02:50
    YouTube or the video version on Spotify
  • 00:02:53
    you can go over to YouTube just type in
  • 00:02:55
    this week and startups hit the Subscribe
  • 00:02:56
    button you'll find this video there
  • 00:02:57
    under the videos Tab and you can
  • 00:02:59
    actually see what we're talking about
  • 00:03:00
    here yeah tell me what you're building
  • 00:03:02
    with gecko thanks for having me on I'm
  • 00:03:03
    really excited to to dive in on one of
  • 00:03:06
    my favorite topics with robotics and
  • 00:03:08
    artificial intelligence how it impacts
  • 00:03:09
    the world but um started in college um I
  • 00:03:12
    started a robotics company out of
  • 00:03:13
    college when I saw firsthand actually
  • 00:03:16
    the state and how the physical world
  • 00:03:17
    that we rely on every single day
  • 00:03:19
    collapses and isn't always there for you
  • 00:03:21
    and this happened at a power plant where
  • 00:03:23
    uh I got to see firstand where power
  • 00:03:25
    plant was having these massive shutdowns
  • 00:03:27
    and the best way to stop it was sending
  • 00:03:28
    a human into a environment and um trying
  • 00:03:32
    to predict when these built structures
  • 00:03:34
    in particular this boiler was going to
  • 00:03:35
    fail and the best way to do that was
  • 00:03:37
    sending a human into a dangerous
  • 00:03:39
    environment that same year I I'd gone
  • 00:03:41
    there um someone had fallen and died
  • 00:03:43
    doing this job sturdy dangerous and uh
  • 00:03:46
    and not typically talked about and so I
  • 00:03:48
    built a wall climbing robot in college
  • 00:03:50
    to solve for the uh the K of critical
  • 00:03:53
    infrastructure that we care now that we
  • 00:03:54
    rely on so deeply to you know live our
  • 00:03:57
    lives every single day so you know 11
  • 00:03:59
    years after after that now here I am um
  • 00:04:01
    still working on the same critical
  • 00:04:02
    mission of protecting and helping to
  • 00:04:05
    build uh new infrastructure but more
  • 00:04:07
    intelligently so these are purpose
  • 00:04:09
    designed robots to do very specific
  • 00:04:12
    tasks you're not taking the approach uh
  • 00:04:15
    that elon's taking a Tesla or the figure
  • 00:04:17
    robot is taking of specifically a
  • 00:04:20
    humanoid robot these are robots that are
  • 00:04:23
    designed for a specific function like
  • 00:04:26
    climbing up and inspecting a building
  • 00:04:28
    correct the whole premise was
  • 00:04:30
    um it seems like we don't care that
  • 00:04:31
    deeply or at least know that much about
  • 00:04:33
    like the built world that we rely on and
  • 00:04:35
    that was the the thought you know when I
  • 00:04:37
    was in college hey we go over a bridge
  • 00:04:39
    every day um hey we rely on power plants
  • 00:04:41
    We R on manufacturing facilities um
  • 00:04:44
    ships to carry supplies all around the
  • 00:04:45
    world um how do we know if those things
  • 00:04:48
    are are going to be around or there for
  • 00:04:50
    us um is it the right assumption to
  • 00:04:52
    believe that the bridge I'm crossing is
  • 00:04:53
    you know it's going to be structually
  • 00:04:54
    sound and not going to collapse and
  • 00:04:56
    that's where you started the journey you
  • 00:04:57
    said hey infrastructure is the ideal
  • 00:05:00
    customer profile for your startup and
  • 00:05:03
    for this product robotics your customer
  • 00:05:06
    is essentially infrastructure and
  • 00:05:09
    specifically infrastructure in the
  • 00:05:10
    United States which for whatever reason
  • 00:05:12
    we seem to have not allocated enough
  • 00:05:15
    resources towards yeah it's um you know
  • 00:05:17
    I in in in 2013 when I was in college
  • 00:05:20
    designing the first robot you know I
  • 00:05:22
    read this report it said 3.34% of GDP
  • 00:05:25
    around the world was spent on fighting
  • 00:05:26
    corrosion I was like wow that's a crazy
  • 00:05:28
    three and a half trillion dollar number
  • 00:05:31
    um I wonder like what that is then you
  • 00:05:32
    look into you know these uh these
  • 00:05:34
    interesting reports that show the US is
  • 00:05:37
    that like a degrade in terms of its
  • 00:05:39
    infrastructure and you know it costs you
  • 00:05:41
    know trillions of dollars just to keep
  • 00:05:42
    it there um and not to just not to
  • 00:05:44
    improve it but just to keep it there
  • 00:05:46
    maintain it yeah maintain it and that's
  • 00:05:48
    you know regardless of building new
  • 00:05:50
    things so it started with the critical
  • 00:05:51
    Industries infrastructure but it was
  • 00:05:53
    mostly this like thought that was wow we
  • 00:05:56
    we seemed like we seem like we talk as
  • 00:05:58
    if we know a lot lot and have a lot of
  • 00:06:00
    data about how the built World works and
  • 00:06:02
    how to make it better but that's
  • 00:06:03
    actually like very far from true for the
  • 00:06:05
    physical world for example we don't know
  • 00:06:07
    if a concrete um structure like a bridge
  • 00:06:10
    is going to be um is going to be sound
  • 00:06:12
    and going to be there and how long will
  • 00:06:13
    it last you know the bridges and
  • 00:06:14
    infrastructure that we rely on was not
  • 00:06:16
    built for the kind of like traffic and
  • 00:06:17
    and loads that we currently are
  • 00:06:20
    demanding on today so we're stressing
  • 00:06:22
    the infrastructure on top of it it was
  • 00:06:24
    built for right you know the Golden Gate
  • 00:06:26
    Bridge was built at a time when a
  • 00:06:28
    certain amount of Vehicles would go over
  • 00:06:30
    it a certain amount of weight of those
  • 00:06:32
    vehicles and obviously uh you know we've
  • 00:06:34
    induced a lot more usage of that with a
  • 00:06:36
    lot heavier Vehicles so maybe you could
  • 00:06:39
    show us uh in sports cast one of these
  • 00:06:42
    robots doing inspections and I know that
  • 00:06:44
    you're not just doing infrastructure
  • 00:06:46
    you've got energy defense manufacturing
  • 00:06:48
    other robots and other verticals you're
  • 00:06:50
    playing in but I would love to see uh
  • 00:06:53
    what these robots are and then get into
  • 00:06:56
    you know the business model uh because
  • 00:06:58
    it is this week in startups of how you
  • 00:07:00
    make money with these robots yeah yeah
  • 00:07:02
    absolutely yeah when I was in college
  • 00:07:04
    looked around and like I was describing
  • 00:07:06
    there seemed to be like this world that
  • 00:07:08
    you know technologists and startups like
  • 00:07:10
    didn't really pay that much attention to
  • 00:07:12
    uh it's the world of energy it's the
  • 00:07:14
    world of manufacturing that's the world
  • 00:07:16
    of defense and public infrastructure and
  • 00:07:19
    you know I saw I saw this like up close
  • 00:07:21
    and personally with the power sector and
  • 00:07:23
    it was just this idea of man we don't
  • 00:07:25
    really have that much data on the built
  • 00:07:26
    world and um and thus it makes it really
  • 00:07:29
    hard to know and understand like how to
  • 00:07:31
    predict how it's going to perform and
  • 00:07:33
    what you're showing on the screen here
  • 00:07:35
    is the Golden Gate Bridge I assume a
  • 00:07:37
    nuclear reactor and then it looks like a
  • 00:07:40
    really either another type of bridge and
  • 00:07:44
    inspectors literally repelling up and
  • 00:07:47
    down them which is dangerous and I'm
  • 00:07:49
    sure quite expensive I don't know what
  • 00:07:51
    those individuals get paid but yeah
  • 00:07:53
    they're getting paid half as much as
  • 00:07:55
    they should what does that person get
  • 00:07:57
    paid to repel off a a nuclear power
  • 00:07:59
    plant or the Golden Gate Bridge what do
  • 00:08:01
    they make a 100 bucks an hour 50 bucks
  • 00:08:03
    an hour do you know you must know um
  • 00:08:05
    yeah it's it's about it depends on the
  • 00:08:07
    level um but it's about in between like
  • 00:08:09
    30 and 70 bucks an hour um that's it oh
  • 00:08:13
    yeah it's probably it's even lower
  • 00:08:15
    $60,000 a year just times it by 2,000
  • 00:08:19
    low I mean yeah over time you can get a
  • 00:08:21
    little higher but yeah it's exactly
  • 00:08:22
    right this is a spherical tank um for
  • 00:08:24
    example at a orland gas Refinery um but
  • 00:08:27
    just like this you enter like this world
  • 00:08:29
    you know like most most like folks who
  • 00:08:31
    are starting technology companies or in
  • 00:08:32
    robotics or AI like have never stepped
  • 00:08:34
    foot on at a Refinery or don't really
  • 00:08:36
    know the first thing about like
  • 00:08:37
    structural and material science or
  • 00:08:39
    what's what what are the hundreds of
  • 00:08:41
    different types of corrosion instead of
  • 00:08:42
    Steels or instead of concretes but these
  • 00:08:45
    all are like really important not just
  • 00:08:47
    like to predict and and and ensure that
  • 00:08:49
    we're not not suffering from some sort
  • 00:08:51
    of catastrophic failure which actually
  • 00:08:53
    has environmental as well as just you
  • 00:08:55
    know functional um implications but also
  • 00:08:57
    how do you actually modulate how you're
  • 00:09:00
    operating the infrastructure to actually
  • 00:09:03
    get more out of the let's say the power
  • 00:09:05
    plant or the refinery um while also
  • 00:09:07
    reducing the amount of greenhouse uh
  • 00:09:09
    emissions that are being that are being
  • 00:09:11
    released by this by the company because
  • 00:09:12
    whenever there's a catastophic failure
  • 00:09:14
    of a pipeline guess what a lot of like
  • 00:09:16
    explosion leads to unfiltered um carbon
  • 00:09:20
    emitted right into the atmosphere and
  • 00:09:21
    like the worst environmental you know
  • 00:09:23
    recorded environmental incident was the
  • 00:09:25
    nordstream pipe exploding for example uh
  • 00:09:27
    or you know these uh um these deep water
  • 00:09:30
    Horizon events so you know ensuring that
  • 00:09:32
    you don't have these C failures is
  • 00:09:33
    actually really important as it relates
  • 00:09:34
    to as zero and those things but yeah so
  • 00:09:37
    the story was 10 years ago in college
  • 00:09:40
    and basically came across this like
  • 00:09:41
    weird problem of power plants having
  • 00:09:43
    these shutdowns and someone had died
  • 00:09:46
    listen a strong sales team can make all
  • 00:09:48
    the difference for a B2B startup but if
  • 00:09:50
    you're going to hire sharks you need to
  • 00:09:52
    let them hunt and you can't slow them
  • 00:09:54
    down with compliance hurdles like sock 2
  • 00:09:57
    what is sock 2 well any company that
  • 00:09:58
    stores customer data in the cloud needs
  • 00:10:00
    to be sock to compliant if you don't
  • 00:10:03
    have your sock too tight your sales team
  • 00:10:04
    can't close major deals it's that simple
  • 00:10:07
    but thankfully vanta makes it really
  • 00:10:08
    easy to get and renew your sock 2
  • 00:10:10
    compliance on average vanta customers
  • 00:10:12
    are compliant in just two to four weeks
  • 00:10:14
    without vanta it takes 3 to 5 months
  • 00:10:17
    vanta can save you hundreds of hours of
  • 00:10:19
    work and up to 85% on compliance costs
  • 00:10:22
    and vanta does more than just sock 2
  • 00:10:24
    they also automate up to 90% compliance
  • 00:10:26
    for gdpr Hippa and more so here's your
  • 00:10:28
    call to action stop slowing your sales
  • 00:10:30
    team down and use vanta get $1,000 off
  • 00:10:32
    at v.com twist that's v.com twist for
  • 00:10:35
    $11,000 off your sock too a Charming
  • 00:10:38
    backstory if there ever was one show us
  • 00:10:40
    the robot okay all right show us the
  • 00:10:43
    robot we want to see this thing in
  • 00:10:44
    action okay you know every Founder's got
  • 00:10:47
    a Charming story do you have a PR team
  • 00:10:49
    that helped you craft that or that's the
  • 00:10:51
    authentic story this is me no this is me
  • 00:10:52
    the authentic story okay yeah so believe
  • 00:10:56
    the first robot was a um was one that
  • 00:10:57
    was climbing up wall and and gathering
  • 00:11:00
    information Visual and ultrasonic
  • 00:11:02
    basically what we're looking at um what
  • 00:11:04
    I was looking at these this you know 10
  • 00:11:05
    years ago was what is the structural
  • 00:11:07
    Integrity of the pressure vessel and um
  • 00:11:10
    how do you ensure that you're
  • 00:11:12
    understanding just like you're doing a
  • 00:11:13
    cat cat scan or doing a sonogram you use
  • 00:11:16
    high frequency sound wave to look inside
  • 00:11:18
    of a material without needing to like
  • 00:11:19
    open it up and destroy the so what we're
  • 00:11:21
    seeing on the screen is a robot that's
  • 00:11:24
    about the size of a pool cleaner with a
  • 00:11:26
    tether and it's zipping up and down some
  • 00:11:30
    pipes and it's open so you could see all
  • 00:11:33
    the innards of it it looks like an
  • 00:11:35
    insect crawling along the pipes that's
  • 00:11:37
    the size of a pool cleaner like I said
  • 00:11:39
    am I about right the robots it's the
  • 00:11:41
    size of a briefcase it comes in a couple
  • 00:11:43
    different forms forms but basically
  • 00:11:46
    heing yeah because it's upside down and
  • 00:11:50
    it's gripping is it suction cups magnets
  • 00:11:53
    what is it doing there so it's climbing
  • 00:11:55
    up the surfaces whether it be an outside
  • 00:11:58
    of a ship or a let say a PR um some sort
  • 00:12:00
    of like piping or um a dam even we'll
  • 00:12:03
    use um uh neodium rare earth magnets
  • 00:12:06
    arranged in a Hallock array and that
  • 00:12:09
    maximizes pull Force into a surface to
  • 00:12:11
    allow for payloads to be added onto the
  • 00:12:13
    robots and they're collecting different
  • 00:12:15
    kinds of data layers one of the data
  • 00:12:17
    layers for example is this Ultrasonics
  • 00:12:19
    um ultrasonic data layer that's looking
  • 00:12:21
    at what's the structural Integrity
  • 00:12:23
    corrosion erosion um of the surface to
  • 00:12:26
    get generalized idea what is the health
  • 00:12:27
    of this just like you would do like a
  • 00:12:29
    picture of a belly using a sonogram test
  • 00:12:32
    for prancy wow so do humans do this when
  • 00:12:36
    they're climbing up and down when we saw
  • 00:12:37
    them repelling do they have some device
  • 00:12:39
    that they do this manually with yeah
  • 00:12:41
    they do so the best so so basically the
  • 00:12:43
    Our Savior today is Joe um Joe and a
  • 00:12:46
    rope um but basically it's it's it's
  • 00:12:48
    These Guys these guys are our best our
  • 00:12:50
    best defense um the guys who are hanging
  • 00:12:52
    off of off ropes or climbing on
  • 00:12:54
    scaffolding or on ja G's and they're
  • 00:12:56
    armed with single um probes that you use
  • 00:12:58
    some Jael squirt the gel on a Surface
  • 00:13:00
    let's say on kilometers of pipeline
  • 00:13:02
    you'd squirt gel every 10 meters every 1
  • 00:13:04
    meter depending on this the uh
  • 00:13:05
    criticality and then you use the
  • 00:13:07
    ultrasonic sensor and you record the
  • 00:13:09
    waveform and then because you know if if
  • 00:13:12
    you understand the the speed of sound
  • 00:13:14
    through that material you can actually
  • 00:13:15
    understand what's the thickness of that
  • 00:13:17
    material and then you could use um and
  • 00:13:19
    then you record that down on a piece of
  • 00:13:21
    paper or in an Excel sheet and basically
  • 00:13:23
    that's the way that we understand how
  • 00:13:25
    the works they're taking a sample but
  • 00:13:28
    you're
  • 00:13:29
    taking so you have the full picture it's
  • 00:13:33
    possible in
  • 00:13:34
    fact that the humans are going to most
  • 00:13:38
    issues am I correct that they're going
  • 00:13:40
    to miss most or some they're gonna miss
  • 00:13:43
    um a fair amount or there's actually
  • 00:13:45
    human eras that relates to interpreting
  • 00:13:47
    the the the waveforms but there's other
  • 00:13:49
    kinds of techniques that you either are
  • 00:13:51
    or not using so visuals one just like
  • 00:13:53
    hey this thing looks like it's leaking
  • 00:13:55
    that's bad um or um we know with around
  • 00:13:58
    like well like you have to do
  • 00:13:59
    certifications of welds on critical
  • 00:14:01
    pipelines for example and you're using
  • 00:14:03
    x-rays um interpreting the X-ray is
  • 00:14:05
    actually pretty difficult and um it's
  • 00:14:07
    also super dangerous because us using
  • 00:14:09
    something that can cause cancer um if
  • 00:14:11
    you're not appropriately like operating
  • 00:14:13
    it um so use so so we actually will'll
  • 00:14:15
    put on the robots something called
  • 00:14:16
    phased array which is basically um
  • 00:14:19
    ultrasound but just like hundreds of
  • 00:14:21
    different um sound waves going into like
  • 00:14:23
    a very small area you can apply
  • 00:14:25
    basically these different payloads onto
  • 00:14:27
    the robots to look at rosion look at
  • 00:14:29
    cracking look at uh generalized erosion
  • 00:14:32
    but then you can also add other kinds of
  • 00:14:33
    information you can use electromagnetics
  • 00:14:35
    to look at what's the um what's the
  • 00:14:37
    damage um over top of some substrates um
  • 00:14:41
    that um you have to like remove um like
  • 00:14:44
    some sort of insulation so got it anyway
  • 00:14:46
    what you're trying to solve for the
  • 00:14:47
    customer is how do you reduce the
  • 00:14:48
    downtime or the or the um the time I'm
  • 00:14:51
    spending not making my product um and so
  • 00:14:54
    that's what you're trying to First help
  • 00:14:56
    the customer understand is how do you
  • 00:14:58
    ensure that you are solving this problem
  • 00:15:01
    um of ensuring that there's not going to
  • 00:15:02
    be some catastrophic event um but
  • 00:15:04
    limiting the amount of time you're not
  • 00:15:05
    making your your product and so the
  • 00:15:08
    robots are going into these like missile
  • 00:15:10
    silos for example or on top of flight
  • 00:15:12
    decks on destroyers um it's climbing
  • 00:15:14
    inside of power plants at boilers um
  • 00:15:16
    it's going on to dams using suction and
  • 00:15:19
    aesan and um there but basically we've
  • 00:15:23
    we've gone from like what you just saw
  • 00:15:24
    in terms of the robot climbing up a wall
  • 00:15:26
    looking at corrosion and erosion um and
  • 00:15:28
    we've now like combine that into a bunch
  • 00:15:30
    of different robots um some of which are
  • 00:15:33
    doing this climbing some of which are
  • 00:15:35
    using um just like you know drones that
  • 00:15:37
    are looking at using photogrametry to
  • 00:15:40
    understand what is like in general um
  • 00:15:43
    let me do a quick analysis of potential
  • 00:15:45
    damaged areas over like large
  • 00:15:47
    geographical area or maybe integrating
  • 00:15:49
    like um a walking dog or and then then
  • 00:15:53
    um you can use fixed sensors to
  • 00:15:54
    continually
  • 00:15:56
    monitor this reminds me of the pranovo a
  • 00:15:59
    full body scan uh which a lot of doctors
  • 00:16:02
    will say hey you don't need it it's
  • 00:16:03
    going to cause you to find things
  • 00:16:06
    nodules little things grows in your body
  • 00:16:08
    you're not going to know what they are
  • 00:16:10
    and you might panic and get anxiety and
  • 00:16:12
    I'm like well wait but what if it is
  • 00:16:14
    something and you live longer because
  • 00:16:17
    you found you know some God for a big
  • 00:16:19
    cancer or tumor early or something with
  • 00:16:22
    your brain I would much rather have that
  • 00:16:25
    therefore uh you are going to inspect
  • 00:16:28
    these things and have an image in time
  • 00:16:31
    and then you can look for the Deltas and
  • 00:16:32
    what's changed between the two Imaging
  • 00:16:35
    so if you were to do this every year on
  • 00:16:39
    the Golden Gate Bridge what what would
  • 00:16:40
    be the frequency that the Golden Gate
  • 00:16:43
    Bridge or you know a submarine should
  • 00:16:46
    have this done to it so we're actually
  • 00:16:48
    working on I'm so I'm in Pittsburgh
  • 00:16:50
    Pennsylvania right now which is uh um
  • 00:16:52
    which where I you know so I started the
  • 00:16:54
    company did three and a half years of
  • 00:16:56
    boot trapping it um down to left like a
  • 00:16:59
    100 bucks Mega count ended up choosing
  • 00:17:00
    to go to YC opposed to an acquisition
  • 00:17:02
    offer uh went out to California against
  • 00:17:04
    all investors like desires came back to
  • 00:17:06
    Pittsburgh close to customers was able
  • 00:17:08
    to grind closer to there in Pittsburgh
  • 00:17:10
    though it's interesting you know there's
  • 00:17:11
    so many bridges it was where we you know
  • 00:17:13
    69% of the world steel was like uh was
  • 00:17:16
    built here um and now it's kind of
  • 00:17:18
    Reinventing itself in terms of like this
  • 00:17:19
    Robotics and AI Hub but um what's what's
  • 00:17:22
    exciting is actually it's actually a
  • 00:17:23
    really great state as it relates to the
  • 00:17:26
    political support to try and utilize
  • 00:17:29
    Technologies like geckos to do things
  • 00:17:31
    like create um the uh the most
  • 00:17:33
    sophisticated Bridge um evaluation and
  • 00:17:36
    infrastructure process there's like you
  • 00:17:37
    know we're we're uh so working with the
  • 00:17:39
    governor actually on on an initiative
  • 00:17:41
    with Bridges But to answer your question
  • 00:17:43
    you want how often you got to inspect a
  • 00:17:45
    bridge you want to be able to look at a
  • 00:17:47
    bridge you you'd want to look at it with
  • 00:17:49
    uh a deep scan like we would do like a
  • 00:17:51
    full health like here's exactly what's
  • 00:17:52
    going on with the entire Bridge maybe
  • 00:17:54
    like once through five years you don't
  • 00:17:56
    want to like look at it every year you
  • 00:17:57
    actually though like once you understand
  • 00:18:00
    um the general the General Health um you
  • 00:18:03
    know similar to how you would do like
  • 00:18:04
    with a human um you would you would then
  • 00:18:07
    um use fixed sensors that are enabled by
  • 00:18:09
    Wi-Fi or 5G and then those are
  • 00:18:12
    constantly updating a digital twin and
  • 00:18:15
    um and that actually like this explain
  • 00:18:17
    what a digital twin is for people yeah
  • 00:18:19
    so digital twin is it's represented in
  • 00:18:22
    in software it's three-dimensional you
  • 00:18:24
    can manipulate it but it needs to update
  • 00:18:26
    itself so it needs to be continually um
  • 00:18:28
    dating in with information whether
  • 00:18:30
    information is the health of the asset
  • 00:18:32
    or how the asset is performing so an
  • 00:18:34
    example for a bridge might be a real
  • 00:18:35
    world example you've come across might
  • 00:18:37
    be yes a real world example is uh um a
  • 00:18:40
    tank um so a tank at let's say Pulp and
  • 00:18:43
    Paper manufacturing place the place
  • 00:18:45
    where we all get our our toilet paper um
  • 00:18:47
    so we have really big contracts with
  • 00:18:48
    this company that's interested in um
  • 00:18:50
    extending the useful life of their tanks
  • 00:18:53
    um but what they want to do is instead
  • 00:18:54
    of um you know the tank is 20 years old
  • 00:18:57
    you have to you know it's pass useful
  • 00:18:59
    life so we have to build a new one and
  • 00:19:00
    we come in and say actually you don't
  • 00:19:02
    need to um we'll take you know this tank
  • 00:19:04
    plus 50 other tanks that look similar to
  • 00:19:06
    this and we'll tell you how to make it
  • 00:19:08
    last 10 years longer or even 20 years
  • 00:19:10
    longer we'll tell you exactly what to
  • 00:19:11
    repair and Jason we're actually because
  • 00:19:13
    we now have this information uh uh on
  • 00:19:17
    this health and structure structural
  • 00:19:18
    Integrity of the world's some of the
  • 00:19:20
    world's most critical infrastructure
  • 00:19:22
    like 500,000 assets that's where we use
  • 00:19:25
    AI hey take a moment picture the
  • 00:19:27
    ultimate All-Star team for your startup
  • 00:19:28
    you got that image maybe you're thinking
  • 00:19:30
    about the Avengers huh maybe the X-Men
  • 00:19:31
    Wolverine Cyclops Iron Man well let's be
  • 00:19:34
    realistic here until you've raised your
  • 00:19:36
    series A or B you're going to need help
  • 00:19:38
    building this a team right you may not
  • 00:19:39
    have all the resources to do it finding
  • 00:19:41
    Talent managing all these timelines and
  • 00:19:43
    ensuring quality is a huge task for
  • 00:19:46
    anybody building a product and scouting
  • 00:19:47
    for the perfect all-in-one 10x developer
  • 00:19:50
    is going to take a lot of time well
  • 00:19:52
    there's somebody out there who does this
  • 00:19:53
    for a living skip the hassle and power
  • 00:19:56
    up with Dev Squad okay they're like
  • 00:19:58
    professor from the Xmen they just know
  • 00:20:00
    where all those great mutant developers
  • 00:20:01
    are those really powerful ones Dev Squad
  • 00:20:04
    offers a complete product team loaded
  • 00:20:06
    with top tier Talent straight from Latin
  • 00:20:07
    America same time zone as us right your
  • 00:20:10
    team is going to have two to six full
  • 00:20:11
    stack developers depending on what you
  • 00:20:13
    need a technical product manager keep
  • 00:20:15
    everything tight and a bunch of
  • 00:20:16
    Specialists maybe you need UI and ux
  • 00:20:18
    design maybe you need devops how about
  • 00:20:20
    quality assurance they're all going to
  • 00:20:21
    be synced up in your time zone and it's
  • 00:20:23
    going to cost 75% less than the
  • 00:20:25
    equivalent team in the USA we all know
  • 00:20:27
    there's a an incredible arbit there's
  • 00:20:29
    Talent everywhere in the world and now
  • 00:20:30
    that we all work online remote hybrid
  • 00:20:33
    whatever it is you know that you can
  • 00:20:34
    integrate this amazing Dev squat team
  • 00:20:36
    into your team if you're looking for
  • 00:20:38
    seamless collaboration and the
  • 00:20:39
    flexibility of monthly payments without
  • 00:20:41
    long-term commitments well it's time to
  • 00:20:43
    take action head over to Dev squad.com
  • 00:20:44
    twist and snag an exclusive 10% off your
  • 00:20:47
    engagement that's Dev Squad d vsq a.com
  • 00:20:51
    twist it's time to Squad up and
  • 00:20:53
    transform your startup with the allstar
  • 00:20:55
    team you deserve so let's take a look at
  • 00:20:58
    a digital twin then you I'm assuming you
  • 00:21:00
    can show me one or yeah cuz this is
  • 00:21:02
    fascinating you inspect this container
  • 00:21:05
    right um and let's say it's got you
  • 00:21:08
    didn't give the exact example of what
  • 00:21:10
    you would continuously monitor in the
  • 00:21:11
    Dig wi but I'm assuming maybe there's
  • 00:21:13
    some area where you think it might get
  • 00:21:15
    fractured or be compromised and so you
  • 00:21:17
    put a sensor on that permanently yeah
  • 00:21:19
    that sends a continuous continuous
  • 00:21:23
    reading to let you know if it's getting
  • 00:21:25
    worse and at what point you think it's
  • 00:21:26
    going to explode or crack or fou which
  • 00:21:29
    would be the equivalent of like in a
  • 00:21:31
    human body just monitoring some you know
  • 00:21:35
    tumor that might be benign or might not
  • 00:21:37
    be benign am I correct in my that's
  • 00:21:39
    right framing here that's right but you
  • 00:21:40
    also okay so this Orient shows a
  • 00:21:42
    business model too it's like you know we
  • 00:21:44
    we started building robots and they were
  • 00:21:45
    really exciting and cool um what we
  • 00:21:47
    ended up finding was that just building
  • 00:21:49
    robots and using a robot as a service we
  • 00:21:51
    don't actually sell the robots we're
  • 00:21:52
    going out to site with the robots and
  • 00:21:54
    getting data and then giving it to the
  • 00:21:55
    customer got it what we ended up finding
  • 00:21:58
    was that that wasn't uh that wasn't a
  • 00:22:00
    model that actually oriented towards
  • 00:22:02
    value creation um so we were creating
  • 00:22:04
    outsid returns in some cases like nine
  • 00:22:07
    and even um like there was even a case
  • 00:22:09
    of a half a billion dollar value
  • 00:22:10
    creation because we stopped this like
  • 00:22:12
    crazy explosion at this Refinery the
  • 00:22:14
    biggest refinery in the US because they
  • 00:22:16
    had gotten B information from the guy on
  • 00:22:18
    the r Joe what we ended up doing was you
  • 00:22:20
    know charging them like a couple hundred
  • 00:22:22
    thousand bucks for this and that was
  • 00:22:23
    crazy because how outsized the value
  • 00:22:25
    creation was so what we ended up also
  • 00:22:28
    finding was there was lack of of uh
  • 00:22:30
    ability to take action on the data to
  • 00:22:32
    improve how the the customer was was um
  • 00:22:35
    was operating their assets um so let me
  • 00:22:37
    walk you through that so what we'll do
  • 00:22:39
    is we'll we'll have a suite of different
  • 00:22:40
    robotics that um that we offer and work
  • 00:22:43
    with the customer to try and solve for
  • 00:22:45
    around the problem in this case it was
  • 00:22:46
    how do I manage you know 50 of my most
  • 00:22:48
    important pressure vessels and tanks for
  • 00:22:49
    this customer and about five years ago
  • 00:22:51
    we started developing Canever which is
  • 00:22:53
    our Enterprise software so when a
  • 00:22:54
    customer buys gecko they're buying um uh
  • 00:22:56
    Enterprise software and that's called
  • 00:22:58
    can and what they're getting from that
  • 00:23:00
    is um a solution oriented towards a very
  • 00:23:03
    you know specific problem for that
  • 00:23:04
    customer set so we actually not only had
  • 00:23:06
    to become experts on Robotics and and Ai
  • 00:23:09
    and software we also had experts on like
  • 00:23:11
    our customers's actual problem both
  • 00:23:12
    upstream and downstream you know I'll
  • 00:23:14
    take you through an example of a which
  • 00:23:16
    I'm assuming you got by asking them
  • 00:23:18
    questions in customer
  • 00:23:20
    interviews and saying well what are you
  • 00:23:22
    going to do with this data we've now
  • 00:23:24
    given you the data and they told you hey
  • 00:23:27
    well we need to make this decision when
  • 00:23:29
    to retire this tank and then it became
  • 00:23:32
    your business becomes not selling a
  • 00:23:34
    robot or selling an inspection your
  • 00:23:36
    business is now extending the life of
  • 00:23:38
    tanks that is one of the important value
  • 00:23:41
    outcomes yes and but it also was like in
  • 00:23:43
    the beginning like I had to like spend
  • 00:23:45
    all my days and time at the customer
  • 00:23:48
    sites and just like living and
  • 00:23:49
    understanding their problems like better
  • 00:23:51
    than they could um and the that for not
  • 00:23:53
    just power plants but you know
  • 00:23:55
    manufacturing facilities like like
  • 00:23:57
    places that are making steel places are
  • 00:23:58
    making aluminum uh places that are
  • 00:24:00
    refining oil that are uh that are
  • 00:24:02
    operating a hydroelectric Dam like you
  • 00:24:05
    you we have to end up going into these
  • 00:24:07
    these industries and understanding
  • 00:24:08
    exactly what they're trying to produce
  • 00:24:10
    and and from my first principles what
  • 00:24:12
    goes into both the like good and poor
  • 00:24:14
    outcomes and then also understanding
  • 00:24:16
    where they get um where they are getting
  • 00:24:19
    value and can pull value like from a
  • 00:24:20
    regulation standpoint it gets really
  • 00:24:22
    complicated so so for this customer we
  • 00:24:25
    call this technique by the way in the
  • 00:24:26
    business a bear hugging uh so when you
  • 00:24:28
    have a customer who's like a key
  • 00:24:30
    customer yeah you give them that big
  • 00:24:32
    bear hug which means you get on location
  • 00:24:34
    you spend time with them I learned this
  • 00:24:36
    from a company we're investors in called
  • 00:24:37
    density. that does people counting and
  • 00:24:40
    when you are on site you will overhear
  • 00:24:43
    things you know and you're going to have
  • 00:24:45
    the customer just through the course of
  • 00:24:48
    hanging out with them give you insights
  • 00:24:50
    that you're not going to get in a
  • 00:24:52
    20-minute customer interview you might
  • 00:24:53
    get them but in all likelihood just
  • 00:24:55
    hanging out at the facility or maybe
  • 00:24:57
    having giv a drink or having lunch with
  • 00:24:59
    Folks at some point you're going to have
  • 00:25:01
    these Epiphany moments yeah then that's
  • 00:25:03
    what happened for you yeah 100% but but
  • 00:25:05
    also like you can codify that into a
  • 00:25:06
    business model so um you know one of our
  • 00:25:09
    one of our so our series a investor was
  • 00:25:11
    Founders fund um Trey Stevens um was our
  • 00:25:15
    as our partner and board member there
  • 00:25:16
    but what was interesting is while I was
  • 00:25:18
    we went we took a trip to the UAE in
  • 00:25:21
    like 2020 it was literally right before
  • 00:25:24
    covid um we almost got stuck actually in
  • 00:25:26
    uh Oman I think it was we ended up just
  • 00:25:29
    he ended up just talking through um
  • 00:25:31
    paler in the early days and how you know
  • 00:25:33
    he's helped set up the paler office in
  • 00:25:35
    the Middle East we ended up talking
  • 00:25:36
    deeply about um for deployed engineering
  • 00:25:39
    and I was like wow this is like this is
  • 00:25:41
    so cool um that they have like taken
  • 00:25:44
    this approach for deployed engineering
  • 00:25:45
    is you send out your engineers on
  • 00:25:47
    deployments uh as an implementation team
  • 00:25:49
    of the software usehold um and then you
  • 00:25:51
    work alongside customers to understand
  • 00:25:52
    their problems to help create a soft the
  • 00:25:55
    software modules that are oriented
  • 00:25:57
    towards the solution that the customer
  • 00:25:59
    is actually trying to solve because in
  • 00:26:00
    reality you know when you deal with
  • 00:26:02
    these industries they're so complex
  • 00:26:03
    these problems are so complex and they
  • 00:26:05
    are so hesitant to either communicate or
  • 00:26:07
    even to talk about like the different
  • 00:26:09
    problems that exist in these like you
  • 00:26:11
    know Manhattan siiz environments like
  • 00:26:13
    the their size refineries the size of
  • 00:26:15
    Manhattan and so these are then there's
  • 00:26:17
    like so many different things B like you
  • 00:26:20
    know variable frequency drives that like
  • 00:26:22
    need to be like you know looked at and
  • 00:26:24
    wrench turned in this way and all these
  • 00:26:26
    like nuances and this is actually one of
  • 00:26:27
    the big issues as well is that there's
  • 00:26:29
    you know these these people that we rely
  • 00:26:31
    on every single day are are completely
  • 00:26:34
    um uh eaching this point of phasing out
  • 00:26:38
    um whether they're dying or or retiring
  • 00:26:40
    and there's a huge knowledge Gap so
  • 00:26:42
    anyway they began to think about um what
  • 00:26:44
    what if actually um we took like the
  • 00:26:46
    early learnings of for deploying our
  • 00:26:48
    roboticists um in combination and
  • 00:26:50
    concert with for deployed um you know
  • 00:26:52
    software Engineers to actually build a a
  • 00:26:54
    vertically integrated stack of data
  • 00:26:57
    collection um of various types and a lot
  • 00:26:59
    of it um so we call them data layers and
  • 00:27:01
    then pull all that into a single um
  • 00:27:04
    source of Truth um a data warehouse and
  • 00:27:07
    then deliver the the modules and
  • 00:27:10
    software to solve customer problems but
  • 00:27:11
    do so um located actually alongside
  • 00:27:14
    customers because you know um you have
  • 00:27:17
    to to convert someone into using a
  • 00:27:19
    different system you actually have to
  • 00:27:20
    help build it alongside of them yeah and
  • 00:27:23
    the current system was probably pencil
  • 00:27:25
    and paper pictures and you know stuff
  • 00:27:27
    scattered across disparate systems I'm
  • 00:27:29
    assuming yeah that's right um and
  • 00:27:31
    inconsistent like you know per site so
  • 00:27:34
    like Marathon you know they might have
  • 00:27:35
    um seven refineries and each of those
  • 00:27:37
    refineries operates completely
  • 00:27:38
    differently because they're both
  • 00:27:39
    producing you know $20 billion do each
  • 00:27:41
    or something like that you didn't show
  • 00:27:43
    us the actual digital twin let's get it
  • 00:27:44
    make sure we show that yeah of
  • 00:27:48
    course yeah it's so fascinating what
  • 00:27:50
    you're doing it's easy in an interview
  • 00:27:52
    like this to get sidetracked into all
  • 00:27:53
    the different nuggets of what you're
  • 00:27:55
    discovering as a Founder but I I did
  • 00:27:57
    want to see the digital twin okay let's
  • 00:27:58
    do it yeah so you start with like a okay
  • 00:28:01
    what what problem you trying to solve
  • 00:28:02
    well we're trying to solve for um you
  • 00:28:04
    know increasing life extension or
  • 00:28:05
    understanding like how to fix like 50
  • 00:28:07
    tanks and manage 50 tanks all right
  • 00:28:08
    sounds good so the outcomes we were able
  • 00:28:10
    to solve for I'll just like skip that to
  • 00:28:12
    the end basically customer will send you
  • 00:28:15
    say like okay customer I need I need
  • 00:28:17
    your metadata as it relates to the
  • 00:28:18
    structures that we're going to go out
  • 00:28:19
    and try to evaluate so they'll send us
  • 00:28:22
    the metadata and then we'll incorporate
  • 00:28:23
    that inside of can lever as we build out
  • 00:28:25
    their profile and so we're we're
  • 00:28:27
    delivering um using just drawings you
  • 00:28:29
    know what is a very rudimentary digital
  • 00:28:31
    twin um and so this is an example of a
  • 00:28:35
    3D representation of a tank using the
  • 00:28:37
    dimensions of the customer then you send
  • 00:28:39
    out your robot Fleet um and so the
  • 00:28:41
    robots go out there and they're climbing
  • 00:28:43
    all over these structures and they're
  • 00:28:44
    trying to evaluate what is the health um
  • 00:28:47
    of this tank and doing so as as quickly
  • 00:28:49
    as possible while a tank is actually an
  • 00:28:51
    operation so you don't have to shut the
  • 00:28:52
    thing down and then you understand what
  • 00:28:54
    the health of that structure is this one
  • 00:28:55
    was pretty bad um red is good for
  • 00:28:58
    example green is bad it Lally pixelated
  • 00:29:00
    because your as the robot's climbing
  • 00:29:02
    it's pulsing the the area it's climbing
  • 00:29:04
    over um hundred of times every single
  • 00:29:07
    inch and then you can either look at
  • 00:29:09
    what's the what's the mean in terms of
  • 00:29:11
    the um uh how um structurally sound or
  • 00:29:14
    how healthy that that area is like you
  • 00:29:16
    know an inch by inch grid or you can
  • 00:29:17
    look at the data in other ways but you
  • 00:29:19
    want to label all that data set because
  • 00:29:21
    it'll be very helpful as relates to what
  • 00:29:22
    kind of
  • 00:29:24
    corrosion uh is going on for whatever
  • 00:29:26
    reason here when you're looking at T
  • 00:29:27
    shell the bottom 25% of the tank is
  • 00:29:30
    green yeah so the bottom is actually
  • 00:29:32
    super healthy is what you're seeing so
  • 00:29:34
    green is healthy red is notth green yeah
  • 00:29:37
    all right AI is moving really fast and
  • 00:29:40
    we're seeing it in all aspects of our
  • 00:29:43
    lives in business personally and you
  • 00:29:46
    know what it's okay to be confused by
  • 00:29:47
    all of this I found a great podcast for
  • 00:29:50
    you that's perfect at helping you
  • 00:29:52
    navigate all things AI it's called the
  • 00:29:54
    next wave and it's brought to you by the
  • 00:29:56
    HubSpot podcast Network the host bat
  • 00:29:58
    wolf and Nathan L both have an insane
  • 00:30:01
    level of AI expertise because they both
  • 00:30:04
    founded and invested in AI companies so
  • 00:30:07
    think of the next wave as your personal
  • 00:30:09
    Chief AI officer that's right serving up
  • 00:30:12
    fresh perspectives industry insights
  • 00:30:14
    tools tips and tricks on how to use AI
  • 00:30:17
    to grow your business that's the most
  • 00:30:19
    important part tactical advice to
  • 00:30:21
    actually put AI to use in your business
  • 00:30:23
    check out the episode the large language
  • 00:30:25
    model race with Peter Wang found of the
  • 00:30:28
    neuron Peak covers the latest AI
  • 00:30:30
    developments Trends and research
  • 00:30:32
    including a fresh take on where Ai and
  • 00:30:34
    your creativity Collide its impact on
  • 00:30:36
    Industries like music and
  • 00:30:38
    Entrepreneurship the balance between
  • 00:30:40
    Innovation and ethics and the rising
  • 00:30:43
    capabilities of llms so here's your qual
  • 00:30:46
    to action super easy I want you to go
  • 00:30:48
    search for the next wave on YouTube or
  • 00:30:51
    your favorite podcasting app And
  • 00:30:53
    subscribe and discover how to leverage
  • 00:30:55
    AI to unlock new opportunities for your
  • 00:30:57
    bus business brought to you by the
  • 00:30:59
    HubSpot YouTube Network so the base is
  • 00:31:02
    healthy but the top 75% for some reason
  • 00:31:05
    is really weathered I don't know if
  • 00:31:07
    that's caused by wind and the or the
  • 00:31:09
    ground is painted or covered what what
  • 00:31:11
    would cause that weird pattern so what
  • 00:31:12
    we what we what we found is like it it's
  • 00:31:14
    very correlated with how you operate the
  • 00:31:15
    tank um and so this is a really
  • 00:31:17
    important tank and they were typically
  • 00:31:19
    filling their tank halfway up um and
  • 00:31:21
    this was like a very common method you
  • 00:31:22
    know in in in fuel storage tanks for
  • 00:31:24
    example you're filling you're filling up
  • 00:31:26
    your fuel storage tanks if you're EX on
  • 00:31:28
    like 40% of the way up because you just
  • 00:31:29
    don't want everything to like collapse
  • 00:31:31
    and um you spew like bunch of uh oil
  • 00:31:34
    into a river somewhere but which is
  • 00:31:36
    actually not good because you're
  • 00:31:37
    actually destroying the asset um in a
  • 00:31:39
    certain way and you're reducing the
  • 00:31:40
    capacity you can run at um and how much
  • 00:31:43
    oil uh so they were scared of a disaster
  • 00:31:46
    so they made a decision that increased
  • 00:31:48
    the likelihood of a disaster so good
  • 00:31:50
    intent bad outcome that's right but also
  • 00:31:53
    you know for one of our customers um
  • 00:31:55
    adock which is the um um the uae's
  • 00:31:58
    national oil company and we just send a
  • 00:32:00
    $30 million deal with them um but the
  • 00:32:03
    big thing they're trying to solve for is
  • 00:32:04
    how do I produce more oil and oil per
  • 00:32:06
    day barrels per day go from three and a
  • 00:32:08
    half million to to five million barrels
  • 00:32:10
    per day um but what we're trying to also
  • 00:32:12
    show is that you can reduce the
  • 00:32:15
    potential carbon emissions while also
  • 00:32:17
    increasing your throughput um if you
  • 00:32:19
    just operate your assets more
  • 00:32:20
    intelligently so we we'll get into like
  • 00:32:22
    exactly what that means um but you don't
  • 00:32:24
    have to actually build new things
  • 00:32:25
    necessarily which is uh a huge deal then
  • 00:32:28
    you send in robots that can do
  • 00:32:30
    evaluations again while the tank is in
  • 00:32:32
    operations um uh in a submersed way and
  • 00:32:35
    so you're again you're looking at what's
  • 00:32:37
    the structural Integrity of the floor
  • 00:32:39
    because the floor is actually um you
  • 00:32:41
    know one of the most compromised areas
  • 00:32:43
    and so the green again is good the red
  • 00:32:45
    is bad and you want to try and evaluate
  • 00:32:46
    where to fix things uh and then you use
  • 00:32:48
    lar to look at what's the depression of
  • 00:32:50
    the tank because it's really heavy and
  • 00:32:53
    so um it'll begin to depress um in
  • 00:32:56
    certain locations which can also lead to
  • 00:32:57
    a bunch of issues so lighter is really
  • 00:32:59
    important and and there's like different
  • 00:33:01
    kinds of like rules and regulations that
  • 00:33:03
    different bodies set out and then we'll
  • 00:33:05
    we'll create these repair plans and so
  • 00:33:08
    that's what that's trying to do is help
  • 00:33:10
    the customer understand how much Capital
  • 00:33:11
    deploy and then like how many years you
  • 00:33:13
    get from the life extension so what
  • 00:33:15
    we'll do is we'll use the data that
  • 00:33:17
    we're collecting from this tank the
  • 00:33:18
    other 50 tanks at that site and then
  • 00:33:20
    also the thousands of other assets that
  • 00:33:22
    look just like this to try and evaluate
  • 00:33:25
    where is the areas that you want to fix
  • 00:33:26
    right now to extend useful life of the
  • 00:33:28
    asset um but then once you do that you
  • 00:33:31
    have to send it back to the real world
  • 00:33:32
    so it goes physical digital physical so
  • 00:33:35
    it has to be an output where there's an
  • 00:33:37
    action being taken from the insights
  • 00:33:39
    that digital twin is actually helping to
  • 00:33:41
    be used by folks that are welding and
  • 00:33:44
    doing repairs for example or uh or like
  • 00:33:46
    trying to make like an actual like
  • 00:33:47
    functional decision around how to
  • 00:33:49
    operate the asset um in this case the
  • 00:33:51
    tank but then we'll install fixed
  • 00:33:53
    sensors that are pinging the digital T
  • 00:33:56
    these are those black little CES around
  • 00:33:59
    the compromised area one of the Rings of
  • 00:34:01
    this you think of it like a barrel is
  • 00:34:04
    compromised so you're putting sensors on
  • 00:34:05
    it that tell you what it'll look at the
  • 00:34:08
    the structural Integrity um so it'll
  • 00:34:10
    ping you every day because corrosion
  • 00:34:12
    actually is doesn't funny enough doesn't
  • 00:34:14
    happen linearly it happens typically um
  • 00:34:17
    in these like weird moments of large
  • 00:34:19
    Decay over short period of time
  • 00:34:20
    typically that's related to like what
  • 00:34:22
    kind of chemical is in the um let's say
  • 00:34:24
    the into the tank that maybe out in
  • 00:34:27
    abnormal or maybe there's some sort of
  • 00:34:28
    you know this really windy and rainy
  • 00:34:30
    like that month or there's a lot of
  • 00:34:32
    sodium in the air and that's like
  • 00:34:33
    causing a lot of like increased
  • 00:34:34
    corrosion like by a golf for example it
  • 00:34:37
    also like people say how did it happen
  • 00:34:40
    slowly then all at once slow yeah it's
  • 00:34:43
    slowly degrading and then some event
  • 00:34:45
    happens and all at once it gets
  • 00:34:47
    compromised is generally Aviation
  • 00:34:50
    bankruptcy yeah and structural failures
  • 00:34:53
    all seem to go in that direction
  • 00:34:54
    suddenly slowly then all at once that's
  • 00:34:57
    exactly right and uh there are some root
  • 00:34:59
    causes that you can begin to understand
  • 00:35:01
    so you might get a shipment of like new
  • 00:35:03
    fuel or a new chemical um and that you
  • 00:35:07
    there might be actually a compromise in
  • 00:35:08
    the quality of that yeah this is
  • 00:35:10
    actually an issue as well with you
  • 00:35:12
    manufacture new things like windmills
  • 00:35:14
    are falling over in Germany right now
  • 00:35:15
    because of poor steel quality um and you
  • 00:35:18
    know so like you know you end up you end
  • 00:35:20
    up having this like issue where you
  • 00:35:22
    might be getting like really inefficient
  • 00:35:24
    process um for some reason but you can
  • 00:35:26
    actually tell when that inefficiency is
  • 00:35:28
    occurring or when there's like imbalance
  • 00:35:30
    of chemicals in your um you know in your
  • 00:35:32
    processing um batch um that you want to
  • 00:35:35
    be able to react to and that's like
  • 00:35:36
    something that's re that's not
  • 00:35:38
    predictable it's it's reactive people
  • 00:35:40
    started coming to you saying hey we
  • 00:35:42
    we're installing this new thing we want
  • 00:35:44
    to have a day one inspection so we have
  • 00:35:47
    a benchmark and so if it was installed
  • 00:35:49
    improperly we can you know before we
  • 00:35:52
    make the final payment to the group
  • 00:35:54
    Construction Group we want you to do the
  • 00:35:56
    inspection of the work done has that
  • 00:35:58
    started to happen it has and actually
  • 00:36:00
    has happened with um the
  • 00:36:03
    $32 billion Columbia class um nuclear
  • 00:36:06
    sub program as well as um other sub um
  • 00:36:09
    subw work so we're working with the Navy
  • 00:36:11
    actually on um new builds and
  • 00:36:14
    Manufacturing and so what's happen
  • 00:36:16
    insting them when they're in dry so
  • 00:36:17
    inspecting them while they're being
  • 00:36:18
    built actually so what ended up
  • 00:36:20
    happening was um so on the on the
  • 00:36:23
    government side the defense side
  • 00:36:24
    schedule adherence is like a really big
  • 00:36:26
    problem as well as like if you're pay
  • 00:36:27
    $132 billion for new Subs 12 new ones
  • 00:36:30
    you want to make sure your you know your
  • 00:36:31
    tax dollars are actually being used for
  • 00:36:33
    building good things with high integrity
  • 00:36:35
    so we're actually building out digital
  • 00:36:37
    twins of the sub as it's being
  • 00:36:38
    constructed and looking at the quality
  • 00:36:41
    of um of welds because that just
  • 00:36:44
    recently caused like a six-month delay
  • 00:36:46
    in a process where they had to take take
  • 00:36:48
    sections of the sub apart so anyway this
  • 00:36:49
    is like Priority One right now um you
  • 00:36:52
    know on the on the on the manufacturing
  • 00:36:53
    of new subside and then we also do some
  • 00:36:55
    work with the Navy we just got um we've
  • 00:36:59
    got a pretty a pretty large you have a
  • 00:37:01
    digital twin of the sub you have a
  • 00:37:02
    digital twin of the battleship um not
  • 00:37:05
    that I can show you but um but we we're
  • 00:37:07
    doing this as well with um actually it's
  • 00:37:09
    an interesting program I can show you
  • 00:37:11
    real quick but um we're beginning to use
  • 00:37:13
    the same kind of tech philosophy as it
  • 00:37:16
    relates to concrete so what what we're
  • 00:37:18
    doing for the US Air Force is um we're
  • 00:37:21
    sending our robot you know different
  • 00:37:23
    form factor robot with um leveraging the
  • 00:37:26
    stack that we we developed to climb up
  • 00:37:28
    nuclear missile silos so there's 450 in
  • 00:37:30
    the US and what's going on right now
  • 00:37:32
    with the Sentinel program is about $125
  • 00:37:35
    are being depl billion dollars are being
  • 00:37:36
    deployed to upgrade the Cold War era um
  • 00:37:40
    nuclear um deterrent system of you um of
  • 00:37:44
    um uh of these icbms in Silo um but the
  • 00:37:48
    what's happening is the concrete's begin
  • 00:37:49
    to Decay and um crumble it's actually
  • 00:37:52
    causing these um and in Oklahoma there
  • 00:37:54
    was this oxidation explosion actually
  • 00:37:56
    one of the ICBM missile which you don't
  • 00:37:58
    want oxidation explosions inside of a
  • 00:37:59
    nuclear chamber so yeah generally
  • 00:38:02
    speaking explosions plus nuclear weapons
  • 00:38:04
    not a good combination not a good
  • 00:38:05
    combination so we're um got soul sourced
  • 00:38:08
    uh on on work it'll be about $250
  • 00:38:10
    million um project but it's um but what
  • 00:38:13
    you're looking at is you want to
  • 00:38:15
    understand there's a steel liner and
  • 00:38:16
    then a concrete liner like 5et of
  • 00:38:18
    concrete you want to understand what's
  • 00:38:19
    the structural Integrity of the concrete
  • 00:38:21
    and like where all the dam where the
  • 00:38:23
    issues are occurring and then what's the
  • 00:38:24
    structural Integrity of the steel and
  • 00:38:26
    then you can figure out okay now I can
  • 00:38:28
    create a plan to fix all this stuff
  • 00:38:30
    amazing so anyway there's uh there's
  • 00:38:32
    there's like these interesting
  • 00:38:33
    applications like new builds as you
  • 00:38:35
    refering to and you're exle yeah no it's
  • 00:38:40
    incredible um and I know you have in the
  • 00:38:42
    deck the the the the Destroyer as well
  • 00:38:45
    and I guess looking at the hull of that
  • 00:38:48
    is all of this going to culminate in
  • 00:38:51
    permanent Robotics and permanent sensors
  • 00:38:54
    being put onto these things or is that
  • 00:38:56
    cost prohibited in some way or just too
  • 00:38:59
    bulky and too much maintenance in and of
  • 00:39:02
    itself because based on what you're
  • 00:39:04
    learning yeah maybe the sensors should
  • 00:39:07
    just be built into everything in the
  • 00:39:09
    same way I know this is like a minor
  • 00:39:10
    analogy here but you know like air tags
  • 00:39:13
    the act of finding your stuff is going
  • 00:39:15
    to be built into other devices I think
  • 00:39:18
    like the Apple remote controls have air
  • 00:39:19
    tags built into them essentially and so
  • 00:39:21
    it does seem to me that based on all
  • 00:39:23
    you're learning and the stuff man they
  • 00:39:25
    should just be putting these sensors in
  • 00:39:26
    a lot of different places permanently or
  • 00:39:28
    have these robots permanently installed
  • 00:39:30
    because the robots currently have an
  • 00:39:32
    inspector working with them correct they
  • 00:39:34
    have to be supervised they're not like
  • 00:39:35
    we're not at the point where like these
  • 00:39:37
    robots just exist in a little cabin and
  • 00:39:39
    go up every week and inspect and go back
  • 00:39:41
    like a a Droid in Star Wars right we're
  • 00:39:43
    not at that level yet no and I don't
  • 00:39:45
    know if you need to be um actually you
  • 00:39:47
    want to do what you said which is once
  • 00:39:48
    you build something you want to
  • 00:39:50
    understand what's the health of that
  • 00:39:51
    structure so J what I what I believe is
  • 00:39:53
    like in the next like five to seven
  • 00:39:54
    years like you won't be able to build
  • 00:39:55
    new things without first understanding
  • 00:39:57
    that the health of that asset on its
  • 00:39:59
    construction and you create a digital
  • 00:40:01
    twin um that is able to be updated as
  • 00:40:03
    well with sensors that you build into
  • 00:40:04
    these structures especially structures
  • 00:40:06
    that are really important like the
  • 00:40:06
    nuclear sub and what you want to do is
  • 00:40:08
    instead of the sub saying it's going to
  • 00:40:10
    last you know 40 years it's going to
  • 00:40:12
    last 50 years that actually can be
  • 00:40:14
    doubled so you want to be able to you
  • 00:40:15
    know you want to make sure that you can
  • 00:40:17
    create something that can be updated
  • 00:40:18
    every time that you're doing some
  • 00:40:20
    turnaround and then eventually maybe
  • 00:40:21
    even don't need to spend 18 months in
  • 00:40:23
    Dry Dock to do an evaluation of the
  • 00:40:26
    health of the structure which is
  • 00:40:27
    currently the state of a lot of our Navy
  • 00:40:30
    so like a third of our Navy is currently
  • 00:40:31
    in drydoc um trying to do its
  • 00:40:33
    maintenance Cycles um which means that a
  • 00:40:35
    third of our Navy is not out there
  • 00:40:36
    patrolling the the the Seas and ensuring
  • 00:40:39
    that conflict is being deterred so it's
  • 00:40:41
    actually a pretty large problem that um
  • 00:40:43
    secretary Navy Del tur and I have talked
  • 00:40:45
    about like a lot actually is this
  • 00:40:47
    schedule adherence and also
  • 00:40:49
    understanding um what is the state of
  • 00:40:51
    the structures as we build them and how
  • 00:40:53
    do we ensure that we're having in
  • 00:40:54
    creating these living U models um of the
  • 00:40:58
    glass sets you could cut that dry dock
  • 00:41:00
    time in all of these cases down by 50%
  • 00:41:04
    you think ultimately 90% well the goal
  • 00:41:07
    should be actually like don't spend any
  • 00:41:10
    time um if you can help it continu
  • 00:41:12
    moding so that the if you're inspecting
  • 00:41:14
    a battleship or a submarine you could
  • 00:41:17
    have it at the surface if it's a
  • 00:41:18
    submarine obviously the the ship is
  • 00:41:20
    already at the surface you could have
  • 00:41:21
    underwater robots inspecting the hull
  • 00:41:23
    while it's out in the ocean yeah you can
  • 00:41:25
    I mean there's like a 2% gain efficiency
  • 00:41:28
    if you can like scrub a hole while
  • 00:41:29
    you're like um while a ship is going
  • 00:41:31
    from like one place to the other just
  • 00:41:32
    because of like you know barnacle and
  • 00:41:34
    and buildup on the on the whole of a
  • 00:41:36
    ship it's there's a bunch of things you
  • 00:41:37
    can do to improve the efficiencies of
  • 00:41:39
    existing um critical infrastructure but
  • 00:41:41
    I think the big thing we should be
  • 00:41:42
    orienting to is just like how do you how
  • 00:41:44
    do you not be so reactive and and
  • 00:41:47
    incentivize a model which is currently
  • 00:41:49
    incentivized for time materials right so
  • 00:41:51
    whenever you're doing whenever you have
  • 00:41:52
    like these large maintenance primes they
  • 00:41:54
    are incentivized to have the maintenance
  • 00:41:56
    cycle last as long as long as it's could
  • 00:41:58
    possibly last right show me an incentive
  • 00:42:00
    I'll show you the outcome exactly the
  • 00:42:02
    longer it's in Dry Dock the more the the
  • 00:42:04
    meter is running right you are a big
  • 00:42:06
    threat to maintenance companies because
  • 00:42:09
    you'll tell them you only need to do
  • 00:42:11
    maintenance on this 20% the other 80%
  • 00:42:13
    fine well I think it's not a not a
  • 00:42:15
    threat it's like a it's it's it's
  • 00:42:17
    orienting the outcome towards like
  • 00:42:19
    improved performance and so it's it's a
  • 00:42:22
    how do you actually you know you know
  • 00:42:25
    like Power by the hour was like the old
  • 00:42:27
    Rolls-Royce model it's like how do you
  • 00:42:29
    um get paid for the amount of up time
  • 00:42:30
    you're producing that should be the
  • 00:42:32
    orientation it should be like you know
  • 00:42:34
    that's what should be the incentive is
  • 00:42:35
    how do you keep this thing in service
  • 00:42:39
    not how much service do you do and
  • 00:42:41
    that's really hard to do because then
  • 00:42:42
    you have an incentive the other way hey
  • 00:42:44
    we got to keep this thing flying and
  • 00:42:46
    maybe you put something up there in the
  • 00:42:48
    air that shouldn't be flying and should
  • 00:42:50
    be in dry talk and should be inspection
  • 00:42:52
    and what you're trying to do is get to
  • 00:42:53
    the truth and the truth shall make you
  • 00:42:55
    free if you actually have the truth
  • 00:42:57
    truth you don't need to if you can get
  • 00:42:59
    to ground truth here first principles
  • 00:43:00
    you're going to not have to try to game
  • 00:43:02
    an incentive but also yeah exactly but
  • 00:43:05
    you when you build new things like think
  • 00:43:07
    about it this is the way what I get
  • 00:43:08
    excited about when you when you build
  • 00:43:10
    new things um you know you want to be
  • 00:43:12
    able to learn from the experience of the
  • 00:43:15
    you know billions of iterations of that
  • 00:43:18
    thing being in you know in use every
  • 00:43:19
    single day we don't do that right now we
  • 00:43:21
    don't you know we can model as much as
  • 00:43:23
    we want about how to build the best sub
  • 00:43:24
    or how to build the best Destroyer how
  • 00:43:26
    to build the best refinery
  • 00:43:27
    or new hydrogen conversion power plant
  • 00:43:31
    but we haven't learned from what's the
  • 00:43:33
    impact of the equipment in operations
  • 00:43:35
    and use that's what we have to figure
  • 00:43:37
    out because you can't you can't build
  • 00:43:39
    new infrastructure unless you're
  • 00:43:40
    learning from the the how it's How the
  • 00:43:42
    old ones are working and this is why
  • 00:43:43
    it's so important to we get we get
  • 00:43:45
    there's a there's so much sex appeal to
  • 00:43:48
    building new things and like in 10 years
  • 00:43:50
    we'll have this really cool new
  • 00:43:51
    autonomous thing drone walking you know
  • 00:43:54
    humanoid um that's going to solve all
  • 00:43:56
    these problems but the problem is
  • 00:43:58
    there's a lot of issues going on today
  • 00:43:59
    and so the approach to solving and using
  • 00:44:02
    you know specific robots for specific
  • 00:44:04
    jobs is actually just to earn the right
  • 00:44:06
    to begin building really cool robots
  • 00:44:08
    that are um you know that are able to do
  • 00:44:10
    like you know more interesting things
  • 00:44:11
    but you got to get the business model
  • 00:44:12
    right and the business model has to
  • 00:44:14
    incentivize and uh and um and make a CEO
  • 00:44:17
    or CFO give a about you know how
  • 00:44:19
    useful um these industry you know 4.0 uh
  • 00:44:23
    principles and tools are because right
  • 00:44:25
    now that's not true and you can hire
  • 00:44:27
    I hear I see this time and time again
  • 00:44:29
    where like I won't name like the AI
  • 00:44:32
    companies but like these AI companies
  • 00:44:33
    come in and say we like you know
  • 00:44:35
    completely turn on your head the way
  • 00:44:36
    you're operating you know your entire
  • 00:44:38
    business and they'll come in for some
  • 00:44:39
    contract that ends up expiring because
  • 00:44:41
    it just did not produce and that's the
  • 00:44:43
    problem it's like you you you think you
  • 00:44:44
    have all the information and data but
  • 00:44:45
    you're building your AI and your
  • 00:44:48
    Solutions um off of ground truth it's
  • 00:44:51
    actually not ground truth there's
  • 00:44:52
    actually a low amount of Integrity if
  • 00:44:54
    you're not if you're not interrogating
  • 00:44:56
    the data all the way to you to the
  • 00:44:57
    ground level and so for for us like we
  • 00:44:59
    are building um Ai and software but off
  • 00:45:02
    of like data sets that robots and and
  • 00:45:05
    smart sensors are actually collecting in
  • 00:45:07
    order to affect some large business
  • 00:45:09
    outcome eida and cash flow is what we
  • 00:45:11
    Orient to um or it could be schedule
  • 00:45:13
    adherence or it can be environmental
  • 00:45:15
    impacts but you have to be able to
  • 00:45:17
    interrogate the the impact from the
  • 00:45:19
    solutions all the way down to like
  • 00:45:20
    what's actually causing the the change
  • 00:45:22
    and for us it's very clear like if you
  • 00:45:24
    can start with a with a core Foundation
  • 00:45:26
    of what is the health of everything um
  • 00:45:28
    of of my built structures then what
  • 00:45:30
    we've actually found is that we don't
  • 00:45:32
    have to ask our customers for data sets
  • 00:45:33
    they'll give it to us and so the end of
  • 00:45:36
    that of that case study I was going to
  • 00:45:37
    show you was those 50 tanks we actually
  • 00:45:40
    were able to extend the useful life of
  • 00:45:42
    that one tank um by 10 years and and
  • 00:45:45
    scrap a an 8 million capex expense and
  • 00:45:48
    we were able to Across the 50 assets
  • 00:45:50
    that the site said and did did an
  • 00:45:52
    analysis that predicted because of a a
  • 00:45:55
    modulation in fil height we were able to
  • 00:45:58
    actually impact gross margin by uh by
  • 00:46:00
    about 4% and so there's like these it
  • 00:46:03
    has to be oriented towards the outcome
  • 00:46:05
    so the customer has to buy that outcome
  • 00:46:08
    they they can't buy the robots and we'll
  • 00:46:10
    be happy to use and integrate we do um
  • 00:46:13
    other sorts of robots because I don't
  • 00:46:14
    want to build all the robots um but it
  • 00:46:16
    has to be again oriented towards the big
  • 00:46:18
    outcome and problem um for the customer
  • 00:46:20
    otherwise like you know it's not going
  • 00:46:22
    to get funed Point yeah I mean we there
  • 00:46:24
    there's there's a use for doing deep te
  • 00:46:27
    uh and there's a use for just trying to
  • 00:46:29
    make things work in the world but at a
  • 00:46:32
    certain point they have to solve a
  • 00:46:33
    problem and I think that's I think what
  • 00:46:36
    you learned was you know the robot was
  • 00:46:38
    one way to solve the problem but the
  • 00:46:40
    sensors is another way to do it right
  • 00:46:42
    and the Those sensors being on there and
  • 00:46:44
    yeah wow it was incredible progress
  • 00:46:46
    you've made let's talk a little bit
  • 00:46:48
    about AI we'll open the aperture here as
  • 00:46:50
    you collect all this data and over the
  • 00:46:53
    next 10 years you'll have systems fail
  • 00:46:55
    you'll have things you got right have
  • 00:46:57
    things you got wrong you know weird
  • 00:46:59
    things will happen random things will
  • 00:47:00
    occur ships will lose power and run into
  • 00:47:03
    Bridges all kinds of events are going to
  • 00:47:05
    occur and you're going to be collecting
  • 00:47:06
    all this data about these things and
  • 00:47:09
    then AI will be able to process all that
  • 00:47:12
    and maybe give us some insights when do
  • 00:47:14
    you think you'll start having insights
  • 00:47:17
    powered by AI a human wouldn't have
  • 00:47:20
    gotten to and in a reasonable amount of
  • 00:47:22
    time and what do you think the insights
  • 00:47:24
    might look like what what what might you
  • 00:47:25
    figure out collecting all this data and
  • 00:47:28
    then you know running algorithms machine
  • 00:47:31
    learning other things against a new
  • 00:47:33
    we've collected now um and own uh data
  • 00:47:35
    sets um on the health and structal
  • 00:47:37
    integrity of over 500,000 of the world's
  • 00:47:39
    most critical assets and what we're
  • 00:47:40
    doing is we're capturing this immense
  • 00:47:42
    amount of information as it relates to
  • 00:47:43
    what is what is going on as it relates
  • 00:47:45
    to why do things why are things damage
  • 00:47:47
    um why are they what kind of damage
  • 00:47:49
    mechanism is occurring and um and also
  • 00:47:53
    um building out machine learning to
  • 00:47:55
    interpret what is a a soundwave
  • 00:47:56
    attenuation indicative of what kind of
  • 00:47:58
    like issues um and so we've been able to
  • 00:48:00
    train actually on um what is causing
  • 00:48:03
    certain sort of damage mechanisms
  • 00:48:05
    because we've been labeling for the past
  • 00:48:06
    11 years ah and so so it's like it's the
  • 00:48:10
    AI that we've that we're we we believe
  • 00:48:13
    very very much in is like let's be
  • 00:48:15
    Masters and very excellent at being the
  • 00:48:17
    best in the world at understanding why
  • 00:48:19
    things um are damaged um what kinds of
  • 00:48:21
    materials what kinds of repair
  • 00:48:22
    techniques what kinds of um inputs as it
  • 00:48:25
    relates to what kind of variables you
  • 00:48:27
    know are um leading to certain kind of
  • 00:48:29
    damage mechanisms um to be able to begin
  • 00:48:31
    to inform and and inform what to expect
  • 00:48:35
    as it relates to how to predict when
  • 00:48:37
    something would fail um what are the and
  • 00:48:39
    then also um how do you increase the
  • 00:48:41
    efficiency maybe thermal efficiency or
  • 00:48:43
    that throughput or even you know like a
  • 00:48:45
    motor efficiency detecting when a motor
  • 00:48:47
    is about to fail and how to make it make
  • 00:48:49
    sure you're adjusting it to be Optimum
  • 00:48:50
    but basically you want to create
  • 00:48:52
    efficiencies um off of this like Poe
  • 00:48:54
    information and data sets that we have
  • 00:48:56
    that no does which is so you'll be able
  • 00:48:59
    to go back to the people who manufacture
  • 00:49:00
    these tanks the integration firms that
  • 00:49:03
    actually install them in the
  • 00:49:04
    construction companies and say you know
  • 00:49:06
    what what we've seen when you're you
  • 00:49:08
    know within a 100 miles or 50 miles of
  • 00:49:10
    the coastline salt water is X Y and Z
  • 00:49:13
    these tanks should be built in the
  • 00:49:14
    following way or this is what happens in
  • 00:49:16
    extreme heat this is what happens in
  • 00:49:17
    extreme cold this is what happens from
  • 00:49:19
    sun damage you there might be silly
  • 00:49:21
    things like a certain coat of some
  • 00:49:23
    sealant in one area might solve a
  • 00:49:26
    problem s and you can even start AB
  • 00:49:28
    testing this right you could you could
  • 00:49:30
    tell this person with 50 tank hey we
  • 00:49:33
    multivar test it we want to have 10
  • 00:49:36
    layers put on these five layers put on
  • 00:49:38
    these three layers put on these and want
  • 00:49:39
    these other ones to be in the shade I'm
  • 00:49:41
    coming up with stupid ideas here but
  • 00:49:43
    just there's no bad ideas when it comes
  • 00:49:45
    to testing extending the life of
  • 00:49:47
    critical infrastructure right and yeah
  • 00:49:50
    that's right that's going to be super
  • 00:49:51
    powerful have you started to give
  • 00:49:53
    manufacturers like notes or installation
  • 00:49:55
    people notes like on how to do things
  • 00:49:57
    better from the get we've not opened up
  • 00:49:59
    um that um uh products or or Services as
  • 00:50:03
    relates to helping improve the OEM
  • 00:50:05
    process and what materials to choose and
  • 00:50:06
    otherwise but um but you're you're
  • 00:50:09
    correct in assuming like that's where
  • 00:50:10
    heads are at as well as assuming who
  • 00:50:12
    cares a lot about this stuff while
  • 00:50:13
    insurance companies do right because
  • 00:50:15
    they're ensuring all these assets
  • 00:50:16
    they're ensuring the downtime from these
  • 00:50:18
    assets and so these data sets are
  • 00:50:19
    actually quite interesting as it relates
  • 00:50:21
    to um the carrot and stick um of um
  • 00:50:24
    adopting these kinds of um these kinds
  • 00:50:26
    of tools
  • 00:50:27
    because the insights of the how well is
  • 00:50:30
    my um billion dollars of infrastructure
  • 00:50:32
    being taken care of you right now is
  • 00:50:34
    being informed by Joe in a row so
  • 00:50:36
    there's a lot if you can measure it and
  • 00:50:38
    you can manage it and you can Ure it
  • 00:50:41
    everybody says if you measure it you can
  • 00:50:42
    manage it managing it in a lot of cases
  • 00:50:45
    means ensuring it actually freeberg who
  • 00:50:46
    I think was the one who brought you guys
  • 00:50:47
    up on a recent all-in pod and which is
  • 00:50:50
    why I invited you um he did Metro mile
  • 00:50:52
    which was also measuring hey how many
  • 00:50:54
    miles are you doing we should only
  • 00:50:55
    charge you for that and then you start
  • 00:50:57
    thinking about Teslas they have a
  • 00:50:58
    driving score I don't know if it's still
  • 00:50:59
    in the app but you know one of the
  • 00:51:02
    people who was driving my car is quite
  • 00:51:04
    an aggressive driver at times and uh it
  • 00:51:08
    was like whoa you're driving pretty
  • 00:51:09
    close to the person in front of you it
  • 00:51:10
    has the distance it has the speed you
  • 00:51:12
    know Zip Zip Zip and you could just make
  • 00:51:15
    insurance for people who are Zippy in
  • 00:51:16
    their cars and people who are slow MOS
  • 00:51:19
    in the the right hand lane and you could
  • 00:51:21
    just right siiz Insurance what you're
  • 00:51:23
    saying is hey with these tanks if we're
  • 00:51:24
    inspecting them and we're doing you know
  • 00:51:26
    this um remediation yeah maybe we should
  • 00:51:30
    have a different Insurance profile than
  • 00:51:31
    somebody who does none of that and if
  • 00:51:33
    we're putting these sensors on here boom
  • 00:51:35
    we should have a different level of
  • 00:51:36
    insurance that's exactly what happens in
  • 00:51:38
    journalism by the way when I first
  • 00:51:40
    started my first magazine they were like
  • 00:51:42
    do you had do factchecking do you check
  • 00:51:44
    quotes with the folks who did it do you
  • 00:51:47
    record your calls you know and they went
  • 00:51:49
    through all this stuff and I was like oh
  • 00:51:51
    wow this is really interesting I'm like
  • 00:51:53
    why does this matter like well we're
  • 00:51:54
    going to make different levels of
  • 00:51:55
    insurance based on your factchecking so
  • 00:51:58
    media insurance people don't know this
  • 00:52:00
    you know if you're I don't know and I
  • 00:52:02
    don't even know if some people like Alex
  • 00:52:04
    Jones take like an extreme example who
  • 00:52:05
    like does conspiracy theories and
  • 00:52:07
    whatever like yeah uninsurable and then
  • 00:52:10
    you go to people like I don't know New
  • 00:52:11
    York Times And if you've ever been in a
  • 00:52:13
    New York Times story like do they check
  • 00:52:15
    the facts do they call you and confirm
  • 00:52:17
    the quotes no yeah New Yorker at least
  • 00:52:20
    in my experience I haven't had a New
  • 00:52:21
    York Times fact Checker check but I have
  • 00:52:23
    had the New Yorker check have had Vanity
  • 00:52:25
    Fair check so con does a really good job
  • 00:52:28
    with that and the insurance I think
  • 00:52:30
    works out being proportional to the
  • 00:52:32
    effort you put into getting your your
  • 00:52:33
    facts correct and here it's you know the
  • 00:52:35
    effort you put into getting your sensus
  • 00:52:36
    correct has have insurance companies uh
  • 00:52:39
    uh started collaborating with you yet or
  • 00:52:41
    they've reached out um and they've come
  • 00:52:43
    inbound but we we've basically just held
  • 00:52:45
    to the approach of look we're we're very
  • 00:52:47
    focused on like helping improve the
  • 00:52:48
    state of our our customers largest
  • 00:52:50
    problems and and Orient to Value
  • 00:52:52
    creation we'll be more interested in
  • 00:52:53
    those kinds of models that you're we
  • 00:52:55
    talking about as relates to OEM and
  • 00:52:57
    insurance um at some point in the future
  • 00:53:00
    but right now it's just you know we want
  • 00:53:02
    to build out the infrastructure and a
  • 00:53:03
    good architecture to begin um
  • 00:53:06
    implementing this like four this
  • 00:53:07
    industry 4.0 um type of like talk um and
  • 00:53:12
    in a in a pragmatic way that's also
  • 00:53:14
    trying to meet the customers where
  • 00:53:15
    they're at I mean a lot of these a lot
  • 00:53:16
    of these customers have a hard have a
  • 00:53:18
    hard time and are very adverse to
  • 00:53:21
    technologists um software and Robotics
  • 00:53:23
    like folks coming in because they just
  • 00:53:25
    have not seen the towards like helping
  • 00:53:28
    them like fight every the fires that
  • 00:53:29
    they fight every single day and so
  • 00:53:31
    they're not actually that willing to
  • 00:53:33
    give you like much information to help
  • 00:53:35
    you build a good product stack I think
  • 00:53:37
    this is like this is why the death of so
  • 00:53:39
    many you know drone companies or
  • 00:53:42
    robotics companies or software companies
  • 00:53:44
    um occur in this sector is because one
  • 00:53:47
    Venture capitalists have no idea about
  • 00:53:48
    these sectors and what they're talking
  • 00:53:50
    about and so like you we're very much of
  • 00:53:52
    Black Sheep because we were this like
  • 00:53:53
    Pittsburgh robotics company focused on
  • 00:53:55
    energy these are all the wrong things
  • 00:53:57
    back in 2016 when I went through IC in
  • 00:53:59
    2016 yes and in 2024 now everybody's got
  • 00:54:02
    the bug right after they've seen What's
  • 00:54:04
    happen with Tesla and SpaceX and that
  • 00:54:06
    opened the wedge up to hey some of these
  • 00:54:09
    boring Industries or you know Real World
  • 00:54:11
    Industries might be worth going after
  • 00:54:13
    and it was also Uber and um Airbnb were
  • 00:54:16
    also real world businesses I remember
  • 00:54:18
    when they were raising their their
  • 00:54:19
    funding people were like I don't want to
  • 00:54:20
    be in a real world business it's too
  • 00:54:22
    dangerous what if somebody trashes your
  • 00:54:23
    apartment it's like well people FR
  • 00:54:27
    hotels every weekend yeah kind of what
  • 00:54:30
    hotels are for at least amongst addicts
  • 00:54:33
    and rockstars is for trashing them and
  • 00:54:36
    like Hotel have figured out how to deal
  • 00:54:38
    with a trashed hotel room they just
  • 00:54:39
    throw everything in they charge the
  • 00:54:40
    person money for trashing it the end
  • 00:54:43
    yeah part of the game again here in
  • 00:54:45
    Pittsburgh there's like there's so many
  • 00:54:47
    robotex companies that like star and die
  • 00:54:48
    all the time and it's beg because the
  • 00:54:49
    it's not because they they are really
  • 00:54:51
    dumb at building like great robots and
  • 00:54:52
    solutions they're actually like really
  • 00:54:54
    smart but the problem is what are the
  • 00:54:56
    robot like useful for yes and and that's
  • 00:54:59
    like the big issue and that's why we
  • 00:55:01
    spend so that's why I spent so much time
  • 00:55:03
    we spend so much time like trying to dig
  • 00:55:05
    in with the customers in an embedded way
  • 00:55:07
    because be hug is so critical if if a
  • 00:55:10
    Founder gets anything out of our our
  • 00:55:12
    together it's the be hug Works being on
  • 00:55:15
    location and it was a famous story I
  • 00:55:17
    think Paul Graham told of telling or Joe
  • 00:55:19
    jebbia told it on this podcast the
  • 00:55:21
    co-founder of Airbnb said you know all
  • 00:55:24
    the customers were in New York and Paul
  • 00:55:26
    Graham told him go to New York and he
  • 00:55:28
    said you know all the places with good
  • 00:55:30
    pictures get rented the places without
  • 00:55:32
    pictures don't get rented he said go to
  • 00:55:34
    New York and take pictures and get a
  • 00:55:35
    good camera and they literally bought a
  • 00:55:37
    digital SLR and started taking great
  • 00:55:39
    pictures I think yeah literally ryot and
  • 00:55:42
    Joe took the pictures themselves as the
  • 00:55:45
    co-founders and this is like the closer
  • 00:55:46
    you get to the customers the closer you
  • 00:55:48
    get to the truth it should seem obvious
  • 00:55:49
    but it's scary to talk to customers for
  • 00:55:51
    some introverted Builders Engineers
  • 00:55:54
    whatever you just got to be right there
  • 00:55:57
    at their desk sitting side by side with
  • 00:55:59
    them solving the problem together I
  • 00:56:00
    think that's what you learned and some
  • 00:56:02
    customers don't want that right but you
  • 00:56:03
    only need one or two to say yes and then
  • 00:56:05
    they get the benefit so if you're on the
  • 00:56:07
    customer side of this if you let a
  • 00:56:09
    startup ined with you not ined embed
  • 00:56:13
    with an a if you embed a startup in your
  • 00:56:15
    company you get all the gains years
  • 00:56:18
    before you're competitors so if you're
  • 00:56:19
    in a big company embed those startups
  • 00:56:21
    and take a risk with them and help them
  • 00:56:22
    build the future with you that's what
  • 00:56:24
    you were able to do which is just so
  • 00:56:26
    brilliant I think the thing that that
  • 00:56:28
    like it's important for listeners to
  • 00:56:29
    understand too is like in a very
  • 00:56:30
    regulated environment where especially
  • 00:56:33
    there's a lot of monopo monopolies that
  • 00:56:34
    play whether it be the government sector
  • 00:56:36
    or it be the energy sector like change
  • 00:56:39
    is like very disincentivized you don't
  • 00:56:41
    want to like change the way you're
  • 00:56:43
    you're you know you're maintaining
  • 00:56:45
    something that could go boom and uh kill
  • 00:56:47
    people and take down you know a Refinery
  • 00:56:50
    that's making 50 million bucks a day so
  • 00:56:52
    that's actually not that intuitive um or
  • 00:56:55
    accepted for you know folks to to um Say
  • 00:56:58
    Hey Kid come on and give your best shot
  • 00:57:00
    that where that was effective was
  • 00:57:02
    actually in the power sector and
  • 00:57:03
    specifically the fossil fuel power
  • 00:57:06
    sector which we just like hell like I
  • 00:57:09
    need I need help because um uh you know
  • 00:57:12
    I've got less funding I've got less
  • 00:57:13
    people I've got less expertise and my
  • 00:57:16
    demand is actually like pretty high
  • 00:57:17
    still and um and like I'm having
  • 00:57:20
    shutdowns of my power plants 50% of the
  • 00:57:22
    year because pressure vessels just keep
  • 00:57:24
    exploding so way to look at this is
  • 00:57:27
    what's at stake you know I always tell
  • 00:57:29
    Founders like how much is at stake here
  • 00:57:30
    and if you're doing the uh family trip
  • 00:57:34
    planning app every time we get pitches
  • 00:57:36
    in like every H 100th one or every 200th
  • 00:57:39
    one is I'm making an app that takes your
  • 00:57:41
    group chat and let you plan a trip in an
  • 00:57:43
    app and you're like not a lot at stake
  • 00:57:46
    and the solution of doing it in iMessage
  • 00:57:49
    or whatever WhatsApp you're into it
  • 00:57:52
    works out just fine it's like enough
  • 00:57:54
    like there's not much at stake here like
  • 00:57:57
    splitting the bill it's like it's $100
  • 00:57:59
    in Mexican food you got to split it four
  • 00:58:00
    ways nobody cares it's not enough at
  • 00:58:02
    stake then you start looking at hey
  • 00:58:05
    getting to space putting stuff in space
  • 00:58:07
    SpaceX lot at stake self-driving you
  • 00:58:09
    know getting from point A to point B
  • 00:58:11
    those there actually a lot at stake in
  • 00:58:13
    you know Uber's business or airbnbs like
  • 00:58:16
    I'm going off vacation I need to place a
  • 00:58:17
    state there's a lot at stake there and
  • 00:58:19
    what you found is like man if if one of
  • 00:58:21
    these things fails that's a half billion
  • 00:58:23
    dollars and that's insurance companies
  • 00:58:26
    people lose their jobs at the company
  • 00:58:28
    people get sued I mean that's a lot at
  • 00:58:31
    stake and as you said in that one
  • 00:58:33
    example you extend that one tank you
  • 00:58:35
    save 8 million bucks and you probably
  • 00:58:37
    made what 800,000 off that customer or
  • 00:58:40
    880,000 off that customer uh yeah a a
  • 00:58:42
    lot a little more than that but yeah
  • 00:58:43
    it's it's a little more than 80 or a
  • 00:58:45
    little more than
  • 00:58:46
    800 a little more than 800 okay so
  • 00:58:49
    essentially if you made a little more
  • 00:58:51
    you you were 15% of the cost of the
  • 00:58:54
    other reality so they got 85% of the
  • 00:58:57
    benefit you got
  • 00:58:58
    15 pretty happy Park yeah yeah that's
  • 00:59:02
    where like I think technology is at its
  • 00:59:04
    best when the customer gets the bulk of
  • 00:59:05
    the gain and the company gets a small
  • 00:59:08
    portion of the gain it makes it a
  • 00:59:09
    no-brainer yeah and and also like you
  • 00:59:11
    have to understand that the these these
  • 00:59:13
    sectors are trying like hell to figure
  • 00:59:15
    out how to adopt technology um and not
  • 00:59:17
    be sold like a bag of goods that is is
  • 00:59:20
    is false and and so like you there has
  • 00:59:23
    to what ends up you end have to do
  • 00:59:26
    create a model that very clearly um you
  • 00:59:29
    can backtrack into where is the value
  • 00:59:30
    creation happening but also how do I
  • 00:59:33
    sort through like the 10 to 20 different
  • 00:59:35
    like um options um for Robotics and
  • 00:59:37
    drones and and AI companies like that's
  • 00:59:40
    really tough for these large
  • 00:59:41
    organizations and they really just want
  • 00:59:43
    like someone to come in and solve like a
  • 00:59:45
    bunch of their problems and so like if
  • 00:59:46
    you can create an environment where you
  • 00:59:48
    can bring in and vet technology um you
  • 00:59:50
    can vet different kinds of um robotics
  • 00:59:53
    tools um fix sensors that are enabled um
  • 00:59:56
    by some smart technology you end up
  • 00:59:58
    putting together the different pieces
  • 01:00:00
    that make up some large outcome that
  • 01:00:01
    you're trying to solve for the customer
  • 01:00:03
    packaged though in a u a software that
  • 01:00:06
    you know helps to centralize decision-
  • 01:00:07
    making and very clearly articulates um
  • 01:00:10
    where the value creation is coming from
  • 01:00:13
    um and you can interrogate how those
  • 01:00:14
    decisions were made and what inputs led
  • 01:00:16
    to the improved outcome so you actually
  • 01:00:19
    need to help you know in order to like
  • 01:00:21
    have a lot more startups enter the
  • 01:00:22
    sector you need to actually create a
  • 01:00:24
    model that very clearly articul Ates
  • 01:00:26
    like what the the product Market fit
  • 01:00:28
    needs to be or what the problem you have
  • 01:00:29
    to solve needs to be what the data layer
  • 01:00:32
    that needs to be added to the stack like
  • 01:00:34
    needs to end up looking like or or um
  • 01:00:36
    what kind of information you end up
  • 01:00:37
    collecting that's not currently out
  • 01:00:39
    there and so that model you yeah that
  • 01:00:41
    actually like it's pretty we got to do
  • 01:00:43
    this with now now half doen like other
  • 01:00:45
    robotics companies where we're like hey
  • 01:00:46
    come under our our contracts and um we
  • 01:00:49
    really love the solutions that you're
  • 01:00:50
    building um you can come under our
  • 01:00:52
    contracts and add these different um
  • 01:00:54
    Solutions wonderful yeah hey you've got
  • 01:00:55
    a great drone a walking drone underwater
  • 01:00:57
    drone we don't have it yeah we'll plug
  • 01:00:59
    it in here's the API and let's rock and
  • 01:01:01
    this is where I see humanoid robots
  • 01:01:03
    going it's like these are really
  • 01:01:04
    complicated problems to solve and the
  • 01:01:06
    data that robots collect in the real
  • 01:01:08
    world is interesting but it's not
  • 01:01:10
    actually super valuable to some customer
  • 01:01:12
    that's trying to solve um how do I
  • 01:01:13
    increase the efficiency of my um you
  • 01:01:16
    know of my like batch process of making
  • 01:01:18
    a roll of of of Steel so like the robots
  • 01:01:21
    can do interesting tasks and can
  • 01:01:23
    actually observe interesting data in the
  • 01:01:25
    real world and get get information
  • 01:01:26
    that's not previously available however
  • 01:01:28
    what is the use of that information um
  • 01:01:31
    as it relates to solving some large
  • 01:01:32
    outcome for a client so so that's how
  • 01:01:34
    like I'm excited about you know
  • 01:01:36
    humanoids and walking dog robots because
  • 01:01:38
    that offers like different data layers
  • 01:01:39
    but like they're one of a couple um
  • 01:01:41
    different data there'll be a thousand
  • 01:01:43
    flowers are going to bloom in robotics I
  • 01:01:45
    mean these little ones to carry your
  • 01:01:46
    burritos from point A to point B I mean
  • 01:01:48
    if you just watched all wars or any
  • 01:01:50
    modern science fiction you're going to
  • 01:01:52
    see a range of robots and and you know
  • 01:01:55
    science fiction author
  • 01:01:56
    and directors and creatives they they
  • 01:01:58
    really do think about human use cases
  • 01:02:01
    and sure enough these little robots that
  • 01:02:03
    would Scurry past Darth Vader's feet
  • 01:02:06
    look just like the ones that are
  • 01:02:07
    delivering burritos in a lot of major
  • 01:02:09
    cities and sure we'll have a c3p we'll
  • 01:02:11
    have an R2D2 we'll have everything in
  • 01:02:13
    between and and the bomb the build of
  • 01:02:17
    materials on your robots is a little bit
  • 01:02:19
    High because you have some I think some
  • 01:02:20
    really intense sensors yeah um so they
  • 01:02:23
    those look like those could be tens of
  • 01:02:25
    thousands of dollars I assume in terms
  • 01:02:27
    of the bomb yeah it's like upper it's
  • 01:02:30
    it's like close to Six Figure is about
  • 01:02:32
    where it is but it's not we're not
  • 01:02:33
    optimizing for the bomb but yeah that's
  • 01:02:35
    right right but when you look at the
  • 01:02:36
    general robot Optimus or figure or some
  • 01:02:39
    of these the bomb on those is going to
  • 01:02:41
    be what do you think you have to take if
  • 01:02:43
    you had to pick a number five years from
  • 01:02:45
    now what's the build of materials and
  • 01:02:47
    then you know we can we can extrapolate
  • 01:02:49
    pricing of consu for consumers after
  • 01:02:51
    that what do you think like a functional
  • 01:02:53
    robot that could walk your dog or I
  • 01:02:56
    don't know do your dishes or I don't
  • 01:02:58
    know you know tidy up around the house
  • 01:03:01
    or or work in a factory what do you
  • 01:03:02
    think the without the specialized
  • 01:03:03
    sensors but I think the bottom on one of
  • 01:03:05
    those is goingon to be it's gonna be
  • 01:03:06
    interesting because you you also like
  • 01:03:07
    have to think about like what kind of
  • 01:03:08
    certifications like the robot has to
  • 01:03:10
    like have or come under but yeah I think
  • 01:03:12
    I think it'll end up being it's going to
  • 01:03:14
    be hard for me to imagine it's below 40
  • 01:03:15
    in in five years um I think it's
  • 01:03:18
    actually a lot higher than that um and
  • 01:03:20
    because those are early on but
  • 01:03:22
    ultimately you think a $40,000 bomb yeah
  • 01:03:24
    but ultimately I think a $40,000 bomb
  • 01:03:25
    makes sense in the next like than
  • 01:03:27
    thought more 20 or so yeah in the next
  • 01:03:30
    seven years I think it'll end up going
  • 01:03:32
    down basically just based on like what
  • 01:03:34
    kind of like volume so I'm not assuming
  • 01:03:35
    like in seven years a lot of volume if
  • 01:03:37
    there's a lot of volume then I'd
  • 01:03:38
    probably estimate it's you know it sits
  • 01:03:40
    closer like the to the 20K I think it'll
  • 01:03:42
    get to 10 um and it'll get cheaper than
  • 01:03:44
    that coming out of China absolutely yeah
  • 01:03:46
    so 40 when they launch 10 ultimately
  • 01:03:49
    when they're commoditized everything in
  • 01:03:51
    between and what what are the major
  • 01:03:53
    costs you think in that robot what what
  • 01:03:56
    are the top two or three costs that
  • 01:03:57
    you're going to need the actuators are
  • 01:03:59
    the big
  • 01:04:00
    thing I think the compute is also going
  • 01:04:03
    to be like expensive I'm not sure how
  • 01:04:04
    that'll be dealt with yeah does it have
  • 01:04:05
    like the equivalent of you know an h100
  • 01:04:09
    powering it or does it have like a
  • 01:04:10
    Macbook and it's connected to the net
  • 01:04:13
    you know it's like a very interesting
  • 01:04:14
    question yeah it'll be there'll be a lot
  • 01:04:16
    of robots that like you know there's
  • 01:04:18
    certain robots that like won't be able
  • 01:04:19
    to go into certain environments that's
  • 01:04:20
    like certified for explosion proof and
  • 01:04:23
    it's like those are expensive yeah both
  • 01:04:25
    to get the certification and to ensure
  • 01:04:27
    that they like won't like combust for
  • 01:04:29
    example yeah battery life comes to mind
  • 01:04:31
    but the actuators are what make the move
  • 01:04:33
    their arms there but the the equivalent
  • 01:04:35
    of your joints essentially and the and
  • 01:04:37
    the the pulley system to move things
  • 01:04:39
    around those are not cheap right now
  • 01:04:41
    yeah and like fine dexterities um like
  • 01:04:43
    those are like really tricky the hands
  • 01:04:45
    yeah we we had a company root AI that
  • 01:04:47
    was picking strawberries that with the
  • 01:04:49
    MIT hand oh yeah you some of these MIT
  • 01:04:52
    hands are so incredible what they are
  • 01:04:54
    capable of doing then we have X picking
  • 01:04:56
    up coffee cups and making lattes and
  • 01:04:58
    putting foam on them and we thought it
  • 01:05:00
    would crush the cup and how does it do
  • 01:05:02
    it it's like oh no cups are easy like
  • 01:05:03
    we're working with berries really you're
  • 01:05:06
    working with berries yeah we're pulling
  • 01:05:07
    strawberries and off and raspberries
  • 01:05:10
    like we're talking about fragile berries
  • 01:05:12
    off of stems it's not an easy task when
  • 01:05:14
    you think about it but I guess in some
  • 01:05:16
    ways it is um and then you could you
  • 01:05:19
    could actually see these being rented
  • 01:05:22
    for 10 bucks a day 20 bucks a day you
  • 01:05:24
    know 10 bucks a day is $3,600 a year 20
  • 01:05:27
    bucks a day is 7,000 a year yeah 20
  • 01:05:30
    bucks a day is what people spend on
  • 01:05:32
    lunch now so 20 bucks a day to have a
  • 01:05:34
    robot's pretty dope in my
  • 01:05:36
    mind I I think maybe less about like the
  • 01:05:40
    commercial the BC um implications mostly
  • 01:05:44
    just because like the amount you have to
  • 01:05:46
    spend on making like getting that last
  • 01:05:48
    10% um for Robotics and the amount of
  • 01:05:51
    time oh edge cases it's it's yeah the
  • 01:05:54
    edge cases are just like so hard and
  • 01:05:55
    expensive so in in my opinion it it more
  • 01:05:58
    aligns to like what kind of value are
  • 01:06:00
    you creating from the robots and then if
  • 01:06:02
    you can create a lot of value and charge
  • 01:06:03
    a lot then you can then you can justify
  • 01:06:05
    like large amounts of spend um onto
  • 01:06:08
    making some really cool robotics I think
  • 01:06:10
    that's the key that most roboticists
  • 01:06:11
    haven't actually solved for is what is
  • 01:06:14
    the what is the value creating and how
  • 01:06:15
    much can you extract from the value
  • 01:06:17
    create um once you do that then you have
  • 01:06:18
    a a vicious cycle of being able to
  • 01:06:20
    optimize those robotics um to do some
  • 01:06:23
    really cool things um and I think that's
  • 01:06:25
    that's the at least the way that I'm
  • 01:06:27
    approaching it um because I don't have
  • 01:06:29
    you know5
  • 01:06:30
    billion you seem good at picking Market
  • 01:06:32
    where would you send the first humanoid
  • 01:06:34
    robotics to to maximize the business
  • 01:06:37
    model soldiers
  • 01:06:40
    welders soldiers um I mean soldiers come
  • 01:06:43
    to mind I mean think about how much
  • 01:06:44
    money we put into a soldier I mean I had
  • 01:06:46
    a friend who was a Green Beret and he
  • 01:06:48
    was like I'm like a f he told me he was
  • 01:06:49
    like a $5 million asset he said the
  • 01:06:51
    seals are like a $20 million asset each
  • 01:06:54
    you know accumulative training you know
  • 01:06:57
    I think robotics will not get used in in
  • 01:06:59
    Warfare unless there's like some large
  • 01:07:01
    conflict and then it'll be like then we
  • 01:07:04
    have bigger problems than robotics
  • 01:07:06
    Humanity problems yeah yeah but I think
  • 01:07:08
    it's like you can't send like robotics
  • 01:07:10
    into some Village because like the edge
  • 01:07:12
    cases right it's like there's so much
  • 01:07:14
    potential issue and then like then
  • 01:07:16
    you're dealing with like a large PR
  • 01:07:17
    problem large government right I I think
  • 01:07:20
    it's ultimately going to be where the
  • 01:07:22
    cost of um where the human exposure and
  • 01:07:25
    theost cost of potentially having like a
  • 01:07:27
    large issue because of like some oosha
  • 01:07:29
    violation or or something like that so I
  • 01:07:31
    really my mind just goes to like what is
  • 01:07:34
    the most deep sea welding dangerous deep
  • 01:07:37
    sea welding is like the one actually in
  • 01:07:38
    my head I was going to because it's you
  • 01:07:40
    know that's the most one of the most
  • 01:07:41
    dangerous jobs your life expectancy is
  • 01:07:42
    like four years and um is four years wow
  • 01:07:46
    that's wild like that it's it's like um
  • 01:07:48
    people get paid like hundreds of
  • 01:07:49
    thousands of dollars um to go do that
  • 01:07:52
    job I think it's like 500k um was like
  • 01:07:54
    was like a going rate for like under SE
  • 01:07:55
    welder but like your Lifey is like turns
  • 01:07:59
    out yeah yeah trees are dangerous very
  • 01:08:01
    dang construction workers truck drivers
  • 01:08:04
    miners I don't think roofers just
  • 01:08:06
    because it's
  • 01:08:07
    again firefighters also very dangerous
  • 01:08:10
    running into burning buildings I think
  • 01:08:12
    but I think it's like what's the what is
  • 01:08:14
    the cost like what is the value you can
  • 01:08:16
    create like from the information that
  • 01:08:17
    the the robots are collecting I think
  • 01:08:20
    the the your mind's going more to like
  • 01:08:22
    labor which I think makes sense but I
  • 01:08:24
    think I'm more interested in like what
  • 01:08:25
    where chat PT's mind went I just did a
  • 01:08:27
    chat PT what's the most dangerous
  • 01:08:29
    professions logging's up there I mean
  • 01:08:31
    you think about it logs falling
  • 01:08:32
    everywhere and like heavy duty Machinery
  • 01:08:34
    with chainsaws and blades and yeah man
  • 01:08:37
    if you ever seen those logs like roll
  • 01:08:39
    down a hill man you're dead if you get
  • 01:08:41
    hit by one of those man I think I think
  • 01:08:43
    that it'll end up just being oriented
  • 01:08:45
    towards yeah maybe on the Labor sidey
  • 01:08:46
    welding but I actually think it's more
  • 01:08:48
    just like walk downs that refineries
  • 01:08:50
    it's you know um it's like it's welding
  • 01:08:53
    um um um and perfecting the weld
  • 01:08:56
    speeding up the time of the weld but
  • 01:08:58
    it's more oriented towards like what
  • 01:08:59
    kind of new information am I getting in
  • 01:09:02
    a way that's um that helps to improve
  • 01:09:04
    the overall state of a let's say like of
  • 01:09:07
    the organism that is like making a role
  • 01:09:10
    of uh steel or um or paper products or
  • 01:09:14
    um refining petroleum or making power
  • 01:09:16
    those of things fact work comes to mind
  • 01:09:19
    as a good first step too right I think
  • 01:09:20
    that's what Elon is thinking is if I can
  • 01:09:21
    get him to work in the Tesla Factory on
  • 01:09:23
    repetitive tasks in a controlled zone so
  • 01:09:26
    no humans to get run into right what
  • 01:09:29
    what I'm thinking more of is like what
  • 01:09:30
    kind of information is being collected
  • 01:09:32
    by the robot that helps improve an
  • 01:09:34
    overall process that a human you know is
  • 01:09:37
    is not constantly streaming data
  • 01:09:39
    somewhere right it's it's constantly
  • 01:09:40
    streaming to your head and that never
  • 01:09:42
    best example of that best example of
  • 01:09:44
    that so you're walk let's say like
  • 01:09:46
    you're walking down um a Refinery and
  • 01:09:49
    you're looking at and trying to listen
  • 01:09:50
    for uh different kinds of noises that
  • 01:09:53
    might be indicative of some like leak
  • 01:09:54
    somewhere or you're looking at you know
  • 01:09:56
    you're you're trying to like look at
  • 01:09:58
    like temperature transmitters and see if
  • 01:09:59
    there's like some kind of inconsistency
  • 01:10:01
    of temperature that like could lead to
  • 01:10:02
    something going boom you can use like
  • 01:10:04
    things like thermal cameras you can use
  • 01:10:06
    things like um you can use things like
  • 01:10:08
    liars as well to like constantly update
  • 01:10:10
    like what is the what's the um process
  • 01:10:13
    of um refining um petroleum um you can
  • 01:10:17
    begin to like incorporate different
  • 01:10:19
    kinds of pieces of information that can
  • 01:10:21
    tell you how efficient is your facility
  • 01:10:23
    operating at and then update whatever
  • 01:10:25
    model you're using and change the way
  • 01:10:27
    that you're actually operating the
  • 01:10:29
    facility because lessons they would
  • 01:10:31
    learn in the field that could be
  • 01:10:32
    incorporated into a better process that
  • 01:10:34
    you're saying basically it's like we we
  • 01:10:35
    we know very little about what's going
  • 01:10:37
    on in the in like in the real world and
  • 01:10:39
    so like robots or like cars that are
  • 01:10:42
    going around with light are spinning all
  • 01:10:43
    the time they like very interesting
  • 01:10:44
    information and data monetized in like
  • 01:10:47
    interesting ways that we don't even know
  • 01:10:49
    hear about or talk about but it's that
  • 01:10:51
    same sort know population density they
  • 01:10:53
    know how popular rway is in your town at
  • 01:10:57
    1:00 on a Sunday measuring emissions is
  • 01:11:00
    a big thing too like it's like if you're
  • 01:11:02
    like walking if you have a robot walking
  • 01:11:03
    around um and doing tests of of how much
  • 01:11:07
    uh you know how much CO2 is coming out
  • 01:11:09
    of my of my stack it's like these these
  • 01:11:11
    are like interesting you know different
  • 01:11:14
    kinds of information that can drive you
  • 01:11:16
    know certain sorts of large outcomes
  • 01:11:17
    maybe it's like some sort of Premium you
  • 01:11:18
    can get from the inflation reduction act
  • 01:11:20
    or something like that it's fascinating
  • 01:11:22
    I think it's going to be like a Brave
  • 01:11:23
    New World can't wait for these things
  • 01:11:25
    come out uh all right listen gecko
  • 01:11:27
    robotics Jake another overnight success
  • 01:11:30
    11 years in the making congratulations
  • 01:11:33
    keeping us safe I mean I just I was just
  • 01:11:35
    thinking about that building in Miami
  • 01:11:37
    remember the pool and that building
  • 01:11:38
    collapsed in Miami yeah me they had just
  • 01:11:42
    been inspect and they and they kind of
  • 01:11:43
    knew that it was messed up but they just
  • 01:11:45
    didn't take it seriously they didn't
  • 01:11:46
    inspect it man you get a couple of those
  • 01:11:48
    happening and and there are other
  • 01:11:50
    countries where the building standards
  • 01:11:52
    are not like the us and that happened in
  • 01:11:54
    the US and I don't know what developing
  • 01:11:57
    nations now are getting rich and have a
  • 01:11:59
    lot of buildings that were built maybe
  • 01:12:00
    when they weren't as rich and there
  • 01:12:02
    weren't as much regulation going back
  • 01:12:04
    and figuring out hey these buildings
  • 01:12:05
    built in you know I'm thinking of
  • 01:12:07
    emerging countries that are now we don't
  • 01:12:10
    use the term first in third world
  • 01:12:12
    anymore but Frontier markets turning
  • 01:12:13
    into Emerging Markets turning into
  • 01:12:15
    primary markets they're going to need to
  • 01:12:17
    inspect some of that previous
  • 01:12:18
    infrastructure and make sure it's what
  • 01:12:21
    yeah it's a there's interesting stats
  • 01:12:22
    like there's 700 there 1700 or 17 , 555
  • 01:12:26
    bridges in New York and uh I think six
  • 01:12:29
    was the latest or not need of immediate
  • 01:12:31
    repairs it's like this stuff's old and
  • 01:12:34
    um I the arrows and it was rusted and
  • 01:12:38
    gross that's like pre one of the Premier
  • 01:12:41
    bridges in New York it's like it's
  • 01:12:42
    really sad to see and then also just
  • 01:12:44
    like you don't think about it in the US
  • 01:12:46
    you can reduce row row um does this
  • 01:12:49
    interesting study um 18% is the um is
  • 01:12:52
    the reduction in US emissions by 2030 if
  • 01:12:55
    can stop um critical assets from um
  • 01:12:58
    failing and exploding within the oil and
  • 01:13:00
    gas manufacturing sector so it's like
  • 01:13:01
    these are like pretty interesting you
  • 01:13:03
    know connection points into how
  • 01:13:05
    important it is to understand the health
  • 01:13:06
    of the built World um that most people
  • 01:13:08
    don't think about yeah and it's that's a
  • 01:13:10
    hard one to sell on unless there's just
  • 01:13:12
    been something terrible that's happened
  • 01:13:14
    on an infrastructure basis and people
  • 01:13:15
    are highlighted to because people don't
  • 01:13:16
    want to talk about the reality of
  • 01:13:19
    another BP oil spill in the Gulf or
  • 01:13:21
    another Bridge collapsing it's just it's
  • 01:13:23
    it's dark to think about it but that
  • 01:13:25
    there are people like you out there
  • 01:13:26
    solving these problems so the rest of us
  • 01:13:28
    can feel safer great job J wish you
  • 01:13:30
    continued success and we'll see you all
  • 01:13:31
    next time on this weekend startups
  • 01:13:33
    bye-bye
Tags
  • Robotics
  • AI
  • Industry 4.0
  • Business Models
  • Data Integrity
  • Automation
  • Industrial Innovation
  • Ground Truth
  • Technology Integration