AI Leaders Reveal the Next Wave of AI Breakthroughs (At FII Miami 2025) | EP #150

00:31:13
https://www.youtube.com/watch?v=kbG2CWrmKTw

Sintesi

TLDRThe panel discusses the potential of AI technology in transforming industries, emphasizing its impact on creativity, business operations, and healthcare. Industry leaders from Stability AI, Liquid AI, Sandbox AQ, and others highlight advancements in generative AI, the significance of quantum technologies, and the importance of open-source models. They stress the necessity of executive buy-in and innovative leadership in effectively integrating AI into companies. The future of AI appears promising, with experts predicting rapid advancements and wider adoption within various sectors.

Punti di forza

  • 🌟 AI is reshaping industries and daily lives.
  • 🎥 Stability AI's Stable Diffusion changed image generation.
  • 🔍 Liquid AI enables private AI on devices without heavy computing power.
  • 🧬 Quantum AI aims to revolutionize drug discovery.
  • 💻 Open-source AI fosters collaboration and innovation.
  • 🏢 Executive buy-in is essential for successful AI implementation.
  • 📈 Small teams can drive significant change in AI development.
  • 📚 AI can enhance education and learning experiences.
  • 🤝 Collaboration is key to harnessing AI's full potential.
  • 🔮 The future of AI holds immense possibilities across sectors.

Linea temporale

  • 00:00:00 - 00:05:00

    The panel opens with a discussion about transformative technologies shaping lives, businesses, and nations. The speaker poses the question of readiness for these advancements, emphasizing the significance of AI in these domains.

  • 00:05:00 - 00:10:00

    Prem, CEO of Stability AI, discusses the impact of their product, Stable Diffusion, on image-based AI generation and how they are fine-tuning AI for content creation in film, TV, and advertising. He indicates that significant video generation advancements are expected to occur within a year.

  • 00:10:00 - 00:15:00

    Alexander from Liquid AI highlights the unique capabilities of their liquid neural networks, which prioritize local processing for privacy and efficiency. The technology allows AI applications on personal devices, addressing privacy concerns in enterprise settings.

  • 00:15:00 - 00:20:00

    Jack from Sandbox AQ explains the power of AI and quantum computing in various sectors, especially in drug discovery and materials science. He emphasizes the need for companies to adopt AI to remain competitive and solve pressing societal challenges effectively.

  • 00:20:00 - 00:25:00

    Jim Keller discusses the importance of hardware in AI development, advocating for the opening of software stacks to enhance collaboration and democratization of technology. He stresses the need for affordable, efficient systems that can simplify the programming of AI models.

  • 00:25:00 - 00:31:13

    Alexander from Quantum Black describes the company's efforts to bridge the gap between advanced AI technologies and practical applications in enterprises, underscoring the critical role of leadership commitment in successfully implementing AI solutions.

Mostra di più

Mappa mentale

Video Domande e Risposte

  • What is Stability AI known for?

    Stability AI is the creator of Stable Diffusion, an AI model for image generation that has significantly impacted the field.

  • How far are we from realistic video generation with AI?

    Experts believe realistic video generation will become increasingly available within the next 6 to 12 months.

  • What is Liquid AI’s unique technology?

    Liquid AI utilizes liquid neural networks, allowing AI applications to run on less powerful devices while ensuring privacy.

  • How is Quantum AI different from standard AI?

    Quantum AI focuses on quantitative models for applications like drug discovery, rather than just language processing.

  • What is the significance of open-source AI models?

    Open-source AI promotes accessibility and innovation, allowing more developers to contribute and collaborate on projects.

  • What is Quantum Black's approach to AI deployment in companies?

    Quantum Black focuses on helping major businesses successfully integrate AI into their operations.

  • Why is executive buy-in essential for AI initiatives?

    Successful AI implementation relies heavily on support and commitment from the highest levels of leadership in organizations.

  • What challenges do companies face when adopting AI?

    Many companies struggle with data management, changing internal politics, and obtaining the necessary skills for AI implementation.

  • What role will AI play in education according to the panel?

    AI can serve as a powerful educational tool, enhancing learning opportunities for students.

  • What do the panelists foresee for the future of AI?

    They predict significant advancements and integration of AI technologies across various industries in the coming years.

Visualizza altre sintesi video

Ottenete l'accesso immediato ai riassunti gratuiti dei video di YouTube grazie all'intelligenza artificiale!
Sottotitoli
en
Scorrimento automatico:
  • 00:00:00
    welcome welcome uh we're about to have a
  • 00:00:02
    conversation and I want you to listen up
  • 00:00:05
    this is the technology that's going to
  • 00:00:07
    reshape your families your lives your
  • 00:00:11
    businesses your industries your nation
  • 00:00:13
    states and the question I'm going to be
  • 00:00:15
    asking at the end here is are you ready
  • 00:00:18
    so let's dive in um I'm going to be
  • 00:00:22
    going through two rounds of questions
  • 00:00:24
    I'm embarrassed that we should in fact
  • 00:00:26
    have a three-hour session for our panel
  • 00:00:29
    uh but we got minutes so you'll excuse
  • 00:00:31
    me as we we run through this um Prem uh
  • 00:00:36
    I love what your company's done and
  • 00:00:38
    you're an example of a CEO who takes a
  • 00:00:41
    company and like doesn't 10 exit you 100
  • 00:00:44
    Exit stability AI uh what are you doing
  • 00:00:49
    and how are you going to impact the
  • 00:00:51
    world thank you for that so stability is
  • 00:00:53
    the creator of stable diffusion which
  • 00:00:56
    launched August 22 which changed
  • 00:01:00
    everything in image based AI generation
  • 00:01:03
    it was the chat GPT moment for for image
  • 00:01:07
    uh there's been over 270 million
  • 00:01:09
    downloads of stable division stable
  • 00:01:12
    diffusion to date uh give you a sense of
  • 00:01:15
    scale the number two most popular uh
  • 00:01:18
    model has been downloaded 9 million
  • 00:01:20
    times so it it is it is by far the
  • 00:01:22
    market leader what we're using it now
  • 00:01:24
    for is my background is in professional
  • 00:01:27
    film and television and we're now fine
  • 00:01:30
    tuning what I call Ultra narrow AI
  • 00:01:33
    fine-tuning our model to uh to work in a
  • 00:01:37
    professional content creation so film TV
  • 00:01:39
    gaming and marketing and advertising so
  • 00:01:41
    you brought Jim Cameron onto your board
  • 00:01:43
    you have investors I'm an investor of
  • 00:01:45
    full disclosure you brought in Eric
  • 00:01:47
    Schmidt incredible group of individuals
  • 00:01:50
    um how far are we from uh creating
  • 00:01:53
    reality given the technology that exists
  • 00:01:56
    right now we are we're already there
  • 00:01:59
    with with certain workflows now what
  • 00:02:00
    we're doing with um to make this a full
  • 00:02:03
    reality we're doing exactly as the
  • 00:02:06
    artist creates a film so what we've seen
  • 00:02:09
    in other textto video models are one
  • 00:02:11
    text and one video that's not how
  • 00:02:14
    professional content is created
  • 00:02:15
    professional content is created in shot
  • 00:02:17
    elements and then they're composited
  • 00:02:19
    together to make the shot so what we're
  • 00:02:21
    doing is going step by step in each one
  • 00:02:24
    of those processes whether that be rig
  • 00:02:26
    removal or paint and rotoscope or camera
  • 00:02:29
    match plate construction and doing hyper
  • 00:02:31
    narrow AI models around each and every
  • 00:02:34
    step we're probably about two dozen uh
  • 00:02:37
    models in and about 50 to 60 models so
  • 00:02:40
    how how far are we from me starring in
  • 00:02:43
    my favorite episode Star Trek you should
  • 00:02:46
    be in that now I it was up to
  • 00:02:50
    me but but in all seriousness in terms
  • 00:02:52
    of video generation on the Fly where I
  • 00:02:54
    have a request to create something
  • 00:02:57
    extraordinary that looks real
  • 00:03:00
    I would say we're probably that's going
  • 00:03:02
    to happen this year I think within six
  • 00:03:04
    12 months yeah year right so how does
  • 00:03:05
    your reality change when you're not sure
  • 00:03:09
    if you've created it or if someone else
  • 00:03:11
    has created it what's possible for you
  • 00:03:13
    in your businesses in your
  • 00:03:16
    lives all right we'll come back to you
  • 00:03:18
    in a moment um
  • 00:03:20
    Alexander uh ran excuse me rine uh again
  • 00:03:24
    a investor in your company so full
  • 00:03:26
    disclosures over here liquid AI uh you
  • 00:03:30
    came out of the gate uh Z to2 billion
  • 00:03:32
    valuation in just about two years uh at
  • 00:03:36
    the end of the day you're
  • 00:03:38
    enabling private AI capability with your
  • 00:03:42
    liquid models uh so I know a number of
  • 00:03:45
    companies that are fearful they don't
  • 00:03:46
    allow their employees to use chat GPT
  • 00:03:49
    because they're concerned that openi has
  • 00:03:50
    access to all the data so what's
  • 00:03:53
    possible using uh liquid AI absolutely
  • 00:03:56
    so um yeah we are a foundation model
  • 00:03:58
    company we are building
  • 00:04:00
    um AI system generative AI systems for
  • 00:04:04
    um Enterprises and we're providing these
  • 00:04:07
    systems we're powering these systems by
  • 00:04:09
    um a new technology not the Transformer
  • 00:04:11
    architecture that enabled the new wave
  • 00:04:13
    of AI we have created something on top
  • 00:04:16
    of something that we invented at MIT uh
  • 00:04:18
    liquid neural networks at Daniel laus's
  • 00:04:21
    lab these are warm inspired AI that
  • 00:04:25
    transferred into like something we
  • 00:04:26
    evolve these kind of systems into
  • 00:04:28
    something that is more uh tangible and
  • 00:04:30
    now we can create value off of this new
  • 00:04:33
    type of AI the very special thing about
  • 00:04:35
    this technology is that the amount of
  • 00:04:37
    compute that is needed for packing a lot
  • 00:04:40
    of intelligence into a device is very
  • 00:04:43
    very minimal so as opposed to other type
  • 00:04:45
    of AI you can get chat gbt experience on
  • 00:04:48
    a phone locally on a phone on a laptop
  • 00:04:51
    and um in places where privacy matters
  • 00:04:54
    you know for as a from a product
  • 00:04:57
    perspective what we did today with
  • 00:04:59
    Enterprise
  • 00:05:00
    wherever there is sensitivity on data or
  • 00:05:02
    security issues that you cannot actually
  • 00:05:05
    use a cloud kind of solution we can or
  • 00:05:08
    you don't have access to GPU based
  • 00:05:10
    infrastructures this is the places where
  • 00:05:12
    liquid can immediately come in and
  • 00:05:14
    expand at at large like putting like our
  • 00:05:17
    our type of enabling Enterprises to use
  • 00:05:20
    generative AI I mean you just raised a
  • 00:05:22
    killer round uh with g42 is one of the
  • 00:05:25
    leads a quarter of a billion dollars
  • 00:05:26
    congratulations on that thank you very
  • 00:05:28
    much um you're tracking on revenues this
  • 00:05:30
    year which are spiking which is which is
  • 00:05:32
    fantastic I remember you used your
  • 00:05:34
    liquid neural networks to fly fighter
  • 00:05:37
    jets can you take one second about that
  • 00:05:40
    yeah well I mean this was one of the
  • 00:05:41
    only neural network architectures that
  • 00:05:43
    enabled safe applications of AI on
  • 00:05:47
    device so they trusted United States Air
  • 00:05:49
    Force trusted our technology to be the
  • 00:05:52
    the first version of a neural network
  • 00:05:54
    that can navigate autonomously a flyer
  • 00:05:56
    jet and U you can imagine like private
  • 00:05:59
    AI is is not just having an a chbt
  • 00:06:01
    experience on a phone it can power cars
  • 00:06:05
    you know it can go on a satellite it can
  • 00:06:07
    go on a on a on a um on a jet you know
  • 00:06:11
    and and the applications of these things
  • 00:06:13
    is phenomenal like recently I talked to
  • 00:06:15
    some CEOs of uh an education company
  • 00:06:18
    that they're providing tablets to
  • 00:06:20
    students and they want to have an
  • 00:06:21
    experience in the in this kind of sector
  • 00:06:24
    this is your educator your physician
  • 00:06:26
    your pilot your everything I mean it's
  • 00:06:28
    it's it's innate Intelligence on your
  • 00:06:31
    device that you own 100% amazing yeah U
  • 00:06:36
    we'll come back um uh a dear friend Jack
  • 00:06:38
    hit first of all I have to point out
  • 00:06:41
    Jack is dressed in Miami we're in Miami
  • 00:06:45
    Peter okay okay welcome everyone to our
  • 00:06:48
    home in Miami Richard thank you for
  • 00:06:51
    bringing us to Miami and I have a
  • 00:06:53
    request for Richard Peter if I could
  • 00:06:55
    yeah please of course for next year for
  • 00:06:57
    fi Miami could we request that the dress
  • 00:07:00
    code be Miami business does everyone
  • 00:07:04
    agree yes
  • 00:07:06
    okay and by the way Miami business means
  • 00:07:09
    wearing avocado dress socks we have
  • 00:07:11
    avocado dress socks there's many others
  • 00:07:14
    to choose from though I do want to do a
  • 00:07:16
    a commercial for Jack's us book AI or
  • 00:07:18
    die um it's very subtle Peter AI or die
  • 00:07:21
    but it is something that every person
  • 00:07:23
    should read it is literally what you
  • 00:07:25
    need to understand as a leader about Ai
  • 00:07:28
    and it's written in a very very readable
  • 00:07:30
    uh fashion so Jack uh you are the CEO of
  • 00:07:35
    sandbox AQ AI A is for AI Q is for
  • 00:07:38
    Quantum you've got incredible uh
  • 00:07:41
    chairman of your board Eric Schmidt you
  • 00:07:43
    spun out of MIT with uh
  • 00:07:46
    Google Google we also we're very
  • 00:07:49
    involved with MIT also though out of
  • 00:07:51
    Google with a incredible uh seed round I
  • 00:07:54
    think of how many uh we raised 850
  • 00:07:57
    million now 850 milon amazing yeah all
  • 00:07:59
    right so what do people need to know
  • 00:08:01
    about about sandbox AQ and the quantum
  • 00:08:06
    liquid networks the quantum networks
  • 00:08:07
    that you're producing yeah language
  • 00:08:09
    models Peter it's a very exciting moment
  • 00:08:11
    first of all it's great to see so many
  • 00:08:12
    friends on the panel here and uh in the
  • 00:08:14
    audience and it's it's a beginning
  • 00:08:16
    moment this is the incipient moment for
  • 00:08:18
    AI uh everyone's excited lots of
  • 00:08:21
    businesses looking at it but I think
  • 00:08:23
    we're pass hopefully the shiny object
  • 00:08:25
    phase of AI and now it's getting serious
  • 00:08:27
    everyone on this panel has very serious
  • 00:08:30
    offerings that really impact business
  • 00:08:32
    and impact how Hollywood Works how many
  • 00:08:35
    many parts of uh the major sectors of
  • 00:08:37
    the world work at sandbox AQ what we
  • 00:08:40
    realize is that language models
  • 00:08:42
    fundamental everyone its table Stakes
  • 00:08:44
    should be using language models to cut
  • 00:08:46
    costs if you have customer service
  • 00:08:48
    because your Delta your Herz your Hilton
  • 00:08:51
    any company with thousands of customers
  • 00:08:53
    must be using large language models to
  • 00:08:55
    cut those costs and actually deliver
  • 00:08:57
    better customer service I think we all
  • 00:08:59
    know though customer service cannot get
  • 00:09:00
    worse you know uh as it's delivered now
  • 00:09:03
    so it's only going to get better but we
  • 00:09:05
    at sandbox AQ decided Peter let's
  • 00:09:07
    actually go for a different part of the
  • 00:09:10
    economy let's go for the quantitative AI
  • 00:09:13
    not the language Ai and what do we mean
  • 00:09:14
    by that if you're santae if you're a
  • 00:09:16
    drug company if you want to create a new
  • 00:09:18
    medicine for cancer uh for Alzheimer's
  • 00:09:21
    for dementia each of our families here
  • 00:09:24
    in this room unfortunately will all be
  • 00:09:26
    impacted at some point in our lives by
  • 00:09:28
    these diseases language models can help
  • 00:09:31
    uh initially when they scour and look at
  • 00:09:33
    all the summaries of scientific
  • 00:09:34
    literature very very helpful uh to give
  • 00:09:36
    you some ideas about what's been done
  • 00:09:38
    before but ultimately Peter if we're
  • 00:09:40
    talking about building a molecule we
  • 00:09:42
    need an AI that is not trained on social
  • 00:09:45
    media and cap pictures but is trained on
  • 00:09:48
    molecules and atoms yes right that's
  • 00:09:50
    fundamental right and uh that is the AI
  • 00:09:54
    that sandbox AQ is the global pay Setter
  • 00:09:55
    in so very importantly you're not
  • 00:09:57
    talking about using Quant computers to
  • 00:10:00
    run these quantitative uh Networks
  • 00:10:03
    you're using you know quantum equations
  • 00:10:06
    on the gpus so what kind of quantum
  • 00:10:08
    equations are you using on the gpus
  • 00:10:09
    right so this is something that everyone
  • 00:10:11
    of course uh we all know from elementary
  • 00:10:13
    school Schrodinger's equation all the
  • 00:10:14
    equations that that everyone is familiar
  • 00:10:17
    with but what we realized the
  • 00:10:18
    Breakthrough that we had is that gpus
  • 00:10:21
    were getting so much better and we have
  • 00:10:23
    a hardware person on the paddle
  • 00:10:25
    representing token Hardware person thank
  • 00:10:27
    you on the paddle uh fundamental to gpus
  • 00:10:30
    is the ability to run in parallel Matrix
  • 00:10:33
    algebra imagine a spreadsheet like Excel
  • 00:10:35
    times another big spreadsheet uh a
  • 00:10:38
    million rows by a million columns and a
  • 00:10:40
    million rows here by a million columns
  • 00:10:42
    that magnitude of of uh Matrix algebra
  • 00:10:46
    we can actually convert the quantum
  • 00:10:48
    equations the equations of drugs of
  • 00:10:51
    treatments of new energy of battery
  • 00:10:52
    storage all that we can convert to the
  • 00:10:55
    language of that GPU and that's the
  • 00:10:57
    Breakthrough that we had now when
  • 00:10:59
    computers come and scale we just had a
  • 00:11:01
    great announcement from Microsoft people
  • 00:11:03
    may have seen that yesterday another
  • 00:11:05
    announcement from Google just a few
  • 00:11:06
    weeks ago these announcements you're
  • 00:11:08
    going to see come in a great Cadence
  • 00:11:10
    culminating in a crescendo Peter in
  • 00:11:13
    about 5 to seven years of having great
  • 00:11:16
    quantum computers We'll add those to the
  • 00:11:18
    Arsenal we'll have GPU qpu Quantum
  • 00:11:21
    Processing Unit in one mesh Cloud hybrid
  • 00:11:24
    but today Peter we use the gpus to get
  • 00:11:26
    the work done with drug companies with
  • 00:11:29
    ramco a ramco I see is a sponsor our
  • 00:11:31
    newest announced customer is a ramco in
  • 00:11:34
    Saudi why a ramco because they want to
  • 00:11:37
    take the hydrocarbons coming out of uh
  • 00:11:40
    the ground and convert them to higher
  • 00:11:42
    order chemicals using carbon and
  • 00:11:45
    hydrogen not low-grade fuels but carbon
  • 00:11:48
    composits as an example that could be
  • 00:11:49
    used to make a car lighter they could be
  • 00:11:51
    used to make a space ret lighter that
  • 00:11:53
    could be used to make an airplane
  • 00:11:54
    lighter for Airbus or Boeing this is the
  • 00:11:57
    kind of transformation that we focus on
  • 00:11:59
    with lqm large quantitative models
  • 00:12:02
    versus the the very necessary large
  • 00:12:04
    language models as amazing amazing uh
  • 00:12:06
    yeah Jack Jack is a nuclear power plant
  • 00:12:09
    behind that man um all right uh Jim
  • 00:12:12
    Keller uh T Toren uh you're a hardware
  • 00:12:15
    manufacturer you're our sole Hardware
  • 00:12:17
    manufacturer against all of these
  • 00:12:19
    software Geeks um congratulations on
  • 00:12:22
    your recent round uh you know pretty
  • 00:12:25
    good $700 million of a series D round so
  • 00:12:28
    you've got lot of capital ability
  • 00:12:29
    Hardware yep
  • 00:12:31
    [Applause]
  • 00:12:33
    I I'll take that so what kind of
  • 00:12:36
    Hardware are you building and when
  • 00:12:38
    someone says no I only do software how
  • 00:12:41
    important is Hardware versus software
  • 00:12:42
    today yeah
  • 00:12:44
    so so gpu's got a a real solid head
  • 00:12:47
    start on building AI because they had
  • 00:12:49
    parallel Computing but they're still
  • 00:12:52
    relatively complicated the program and
  • 00:12:55
    the way they do like handle tensors and
  • 00:12:58
    stuff actually wasn't to gpus now gpus
  • 00:13:00
    have evolved to add tensor processors
  • 00:13:03
    tense torrent builds a native tensor
  • 00:13:05
    processor that's simpler and easier to
  • 00:13:08
    program also we build it so that the
  • 00:13:10
    tensor processors natively talk to each
  • 00:13:13
    other really nicely and um and then we
  • 00:13:16
    last year we open sourced our software
  • 00:13:18
    stack and like the it's really
  • 00:13:20
    interesting the fundamental math AI is
  • 00:13:22
    simple a = b * C plus d like it couldn't
  • 00:13:25
    be simpler at some level but the the
  • 00:13:28
    scale of it is amazing when I started
  • 00:13:31
    building computers 40 years ago we were
  • 00:13:34
    doing millions of instructions a second
  • 00:13:36
    now we're doing trillions of trillions
  • 00:13:38
    of instructions a second and to scale
  • 00:13:41
    that takes a like a special
  • 00:13:43
    collaboration between the hardware and
  • 00:13:44
    the software so when you when your
  • 00:13:46
    machines are up and operating uh I guess
  • 00:13:48
    the question is what are they enabling
  • 00:13:51
    uh for people in the room here well well
  • 00:13:55
    right now there's a really large family
  • 00:13:57
    of models so our mission is run all the
  • 00:13:59
    models with really simple transparent
  • 00:14:02
    code so big llm people say oh the
  • 00:14:05
    software is huge actually it's 600 lines
  • 00:14:07
    of code it's not very complicated at the
  • 00:14:09
    program level but when you go down in
  • 00:14:11
    the software stack it can really explode
  • 00:14:14
    and that's where by building a native
  • 00:14:16
    software stack that is tensor based it's
  • 00:14:19
    communication based and it's open source
  • 00:14:22
    people can see exactly how that works
  • 00:14:24
    and how it runs and I think that's going
  • 00:14:26
    to unlock a lot of AI applications that
  • 00:14:29
    are currently hard to program with gpus
  • 00:14:31
    so there's been a lot of debate on open
  • 00:14:33
    source versus closed Source AI
  • 00:14:36
    models and we just saw the R1 model
  • 00:14:38
    being open source we're seeing a lot of
  • 00:14:40
    conversation where the leaders are
  • 00:14:42
    saying you know open source will win I
  • 00:14:45
    think there's been an extraordinary
  • 00:14:47
    velocity in open sourcing how important
  • 00:14:49
    is open source uh as far as you see yeah
  • 00:14:53
    it's it so I have personal experience
  • 00:14:55
    working with gpus where we're trying to
  • 00:14:58
    solve a hard software problem we
  • 00:15:00
    couldn't cuz the the math library was
  • 00:15:02
    encrypted or part of the software stack
  • 00:15:04
    was proprietary and because we couldn't
  • 00:15:06
    look all the way down the stack we
  • 00:15:08
    couldn't figure out the problem and
  • 00:15:10
    solve it now open source AI is really
  • 00:15:12
    wild because most of the high-end
  • 00:15:15
    research is published many of the models
  • 00:15:17
    are open- Source some of the weights not
  • 00:15:20
    that much of the infrastructure not that
  • 00:15:23
    much of the foundation Library so it
  • 00:15:25
    turns out to be kind of a mixed bag like
  • 00:15:27
    one thing we're going to do is we're
  • 00:15:29
    going to open source a whole software
  • 00:15:30
    stack and it's for our Hardware but I
  • 00:15:32
    encourage people if you have your own
  • 00:15:34
    hardware and you want software stack
  • 00:15:35
    that works you know steal our software
  • 00:15:37
    it's a beautiful thing and then I want
  • 00:15:40
    to make it so many of the foundation
  • 00:15:42
    models the environment the framework the
  • 00:15:45
    build and train your own models are also
  • 00:15:47
    open source and available and I think
  • 00:15:50
    it's really important to democratize the
  • 00:15:52
    hardware stack and the software stack so
  • 00:15:54
    it's not just a few very large players
  • 00:15:56
    that control the AI world so we have a
  • 00:15:58
    lot of people from around the world here
  • 00:16:00
    are the machines that you're building
  • 00:16:03
    likely to be used in the global South
  • 00:16:06
    more than in in North America no we're
  • 00:16:09
    going to sell to everybody we license a
  • 00:16:11
    small AI configuration to go on a
  • 00:16:13
    television chip and we're building
  • 00:16:15
    machines that can train large language
  • 00:16:16
    models and everything in between and the
  • 00:16:19
    other part of our business model and
  • 00:16:20
    again I think inov Innovation comes from
  • 00:16:23
    lots and lots of input so the software
  • 00:16:24
    is open it's been our best hiring
  • 00:16:27
    strategy by the way this is great our
  • 00:16:29
    programmers look at our software stack
  • 00:16:31
    they like it they send a resume or worse
  • 00:16:33
    or funnier they don't like it and they
  • 00:16:35
    send me a resume because they want to
  • 00:16:37
    come fix it and I think and I think
  • 00:16:39
    that's really great and then we've
  • 00:16:41
    licensed our Ai and our risk 5 CPU
  • 00:16:44
    technology to people and they like it
  • 00:16:46
    and they use it and they send us
  • 00:16:47
    feedback so we're going to license our
  • 00:16:49
    AI technology but also build and sell
  • 00:16:52
    systems fantastic um how many folks here
  • 00:16:55
    have heard of Mackenzie everybody right
  • 00:16:58
    how many folks here heard of quantum
  • 00:16:59
    black could you raise your
  • 00:17:01
    hand uh you need some publicity
  • 00:17:05
    Alexander uh so Alexander runs Quantum
  • 00:17:08
    black uh which is a 5,000 person
  • 00:17:12
    software and engineering team inside of
  • 00:17:14
    McKenzie on their AI Focus area um ex
  • 00:17:19
    your your story has been incredible tell
  • 00:17:21
    us about Quantum black please thanks P I
  • 00:17:24
    think there are basically Two Worlds
  • 00:17:26
    there is a beautiful and shiny world
  • 00:17:28
    that is on this stage and we all enjoy
  • 00:17:30
    the age of AI enjoy the evaluations
  • 00:17:33
    right and having great life and I think
  • 00:17:35
    we have a quite a confused audience that
  • 00:17:37
    kind of hears it but doesn't see any
  • 00:17:39
    impact in their lives be it a bottom
  • 00:17:41
    lines but be it also as human beings
  • 00:17:44
    right and kind of so the question is how
  • 00:17:46
    do you reconcile these two and if you
  • 00:17:48
    look at the numbers and we take the
  • 00:17:49
    technology companies aside the said
  • 00:17:52
    number in The Last 5 Years is
  • 00:17:55
    11% meaning only one out of 10 use cases
  • 00:17:59
    ever saw the light of production
  • 00:18:01
    everything else is kind of entertainment
  • 00:18:03
    we play with it but it's irrelevant now
  • 00:18:05
    while we inspired by
  • 00:18:07
    gen the success rate is maybe 7% at the
  • 00:18:11
    best so what we're trying to do in
  • 00:18:13
    Quantum black is basically bring these
  • 00:18:15
    two worlds together where we're trying
  • 00:18:18
    to move from 11% into 100% and what we
  • 00:18:23
    got today exactly like you alluded to is
  • 00:18:25
    5,000 people working in 50 countries
  • 00:18:28
    trying to transform Nations and Company
  • 00:18:32
    we have a five R&D centers that working
  • 00:18:35
    on the most precious products for
  • 00:18:37
    Humanity and we roughly 43 products that
  • 00:18:41
    we deploy globally and we clearly work
  • 00:18:43
    with a lot of colleagues to kind of try
  • 00:18:45
    to bring their Innovation into
  • 00:18:47
    day-to-day of Enterprises around the
  • 00:18:49
    globe and K indeed McKenzie is known for
  • 00:18:52
    producing slid and many other funny
  • 00:18:53
    things but also to be fair it's
  • 00:18:55
    reinvented the Consulting profession
  • 00:18:57
    already twice and this is the third
  • 00:18:59
    attempt and humbled to be here and try
  • 00:19:01
    to reinvent it you know one of the
  • 00:19:04
    things I say is that uh you know by the
  • 00:19:07
    end of this decade there going to be two
  • 00:19:09
    kinds of companies those that are fully
  • 00:19:11
    utilizing Ai and those that are out of
  • 00:19:15
    business uh do you agree with that I
  • 00:19:18
    agree with it but I think the problem is
  • 00:19:19
    slightly different that today most of
  • 00:19:22
    these transformation and I think
  • 00:19:24
    everybody here tried to do a digital
  • 00:19:25
    transformation right or at least
  • 00:19:27
    declared it to their boards and share
  • 00:19:28
    share holders now the success rate is
  • 00:19:30
    not high because what we're trying to do
  • 00:19:33
    is to infuse technology into by
  • 00:19:35
    definition broken
  • 00:19:37
    process into the old process and what I
  • 00:19:40
    truly believe we are instead of kind of
  • 00:19:42
    learning technology as all the
  • 00:19:44
    colleagues here on the stage and
  • 00:19:45
    stepping back and trying to reinvent
  • 00:19:47
    something and understand completely
  • 00:19:50
    different reality that operates on very
  • 00:19:51
    different cost structure different
  • 00:19:53
    social rules which tries to force it and
  • 00:19:55
    therefore it fails so if anything I
  • 00:19:57
    truly believe that we are at the end of
  • 00:19:59
    age of mediocracy whatever mediocre
  • 00:20:02
    whatever standard could be kind of
  • 00:20:04
    played by machine but actually beginning
  • 00:20:06
    of the age of creativity because the
  • 00:20:08
    notion of how do you create something
  • 00:20:10
    with technology that is a commodity
  • 00:20:11
    becomes much more
  • 00:20:13
    interesting Prem back to you most
  • 00:20:17
    important thing that uh the audience
  • 00:20:20
    here needs to take away from the work
  • 00:20:22
    stability is doing from your perspective
  • 00:20:26
    on AI as a Enterprise uh uh creativity
  • 00:20:31
    tool um speak to us about that sure so
  • 00:20:36
    probably everybody's seen so much
  • 00:20:38
    controversy around AI in the
  • 00:20:40
    entertainment industry and there was
  • 00:20:42
    even the industry went on strike for
  • 00:20:43
    over a year uh and then we've they've
  • 00:20:46
    settled and all the guilds have like cut
  • 00:20:48
    deals with the studios and this is
  • 00:20:51
    really no different uh than what really
  • 00:20:53
    happened in 1927 when the first uh when
  • 00:20:57
    movies went from Silent to talking what
  • 00:20:58
    they called before there was great
  • 00:21:00
    controversy at that point they thought
  • 00:21:02
    everybody in the Broadway they thought
  • 00:21:04
    talking was for Broadway and movies
  • 00:21:06
    needed to be silent and uh obviously
  • 00:21:08
    that was proven wrong color took forever
  • 00:21:10
    color took decades to actually catch on
  • 00:21:12
    and then finally in the 60s it did and
  • 00:21:14
    now it's Unthinkable you know it's other
  • 00:21:16
    it's like an artistic Choice obviously
  • 00:21:18
    to be in black and white and then of
  • 00:21:20
    course in digital transformation
  • 00:21:22
    everybody kind of fought it at first and
  • 00:21:24
    they and they they're confusing headwin
  • 00:21:26
    with Tailwind uh I think is probably the
  • 00:21:28
    best best way I can summarize it and of
  • 00:21:31
    course when digital you know kicked off
  • 00:21:33
    in 2000 instead of film by 2017 I think
  • 00:21:37
    like 98% of all films were were made on
  • 00:21:39
    on digital so the lasting statement for
  • 00:21:42
    the film industry is don't look at AI as
  • 00:21:44
    headwind look at it as tailwind and do
  • 00:21:47
    we see stability becoming a creative
  • 00:21:49
    agent so that every individual can
  • 00:21:52
    become a Creator absolutely and we're
  • 00:21:55
    going to see it we're going to see it in
  • 00:21:57
    Avatar 3 4 and 5
  • 00:21:59
    well Avatar 3 Jim's editing now so I
  • 00:22:01
    think it's it done but the uh and
  • 00:22:03
    hopefully it'll come out in December uh
  • 00:22:05
    but uh uh definitely I think in the
  • 00:22:08
    later avatars and others I think you're
  • 00:22:10
    going to hopefully see a lot of our
  • 00:22:11
    Tools in there all right well
  • 00:22:12
    congratulations on the success uh you
  • 00:22:15
    know Prem came in as a CEO how long ago
  • 00:22:18
    about eight months ago eight months ago
  • 00:22:20
    and just revolutionized the company had
  • 00:22:22
    a huge Legacy of models and capabilities
  • 00:22:25
    but it really has really driven it
  • 00:22:26
    extraordinary way thank you congrats
  • 00:22:29
    remmen
  • 00:22:31
    um the world needs another
  • 00:22:35
    llm why no so what the word needs is as
  • 00:22:41
    as uh I mean detailed out very nicely
  • 00:22:43
    it's it's the phase right now that we
  • 00:22:45
    have to make AI useful you know it
  • 00:22:47
    doesn't matter what runs AI I mean and
  • 00:22:49
    then we are in this amazing period of
  • 00:22:52
    time where at every scale AI can bring
  • 00:22:54
    value you know and um you put it nicely
  • 00:22:58
    like I I can see a future where
  • 00:23:00
    everything is going to be integ I is
  • 00:23:02
    going to be integrated in our society
  • 00:23:04
    it's not that we need a different type
  • 00:23:05
    of uh LM we need to do it right you know
  • 00:23:08
    we have a motto like in our offices that
  • 00:23:10
    you know like every engineer at liquid
  • 00:23:12
    AI we're designing AI we call it ml Done
  • 00:23:16
    Right machine learning done right so
  • 00:23:18
    that means we don't need to consume a
  • 00:23:20
    lot of energy to build AI systems we
  • 00:23:23
    don't need to use a lot of energy to
  • 00:23:25
    host AI systems so what we do we're
  • 00:23:28
    basically democratizing kind of access
  • 00:23:29
    to AI so if you think about it in a in a
  • 00:23:33
    in the cheapest possible way so hosting
  • 00:23:35
    a foundation model on a phone or on a
  • 00:23:38
    device today with liquid AI costs $0
  • 00:23:42
    because it doesn't run on a GPU anymore
  • 00:23:45
    is Everything is
  • 00:23:46
    Everything is every device in my home in
  • 00:23:49
    my car in my office going to be AI
  • 00:23:51
    enabled correct yes so so what's your
  • 00:23:53
    world look like when everything is
  • 00:23:55
    intelligent every single device you're
  • 00:23:57
    touching talking to thinking about is
  • 00:24:00
    intelligent I think it's going to be
  • 00:24:02
    amazing and the human humanoid robots
  • 00:24:04
    that runs in our homes in the future
  • 00:24:06
    they're not going to have access they're
  • 00:24:08
    not going to be connected to the cloud
  • 00:24:09
    you know they're going to have their own
  • 00:24:10
    kind of local AI Banks when they have
  • 00:24:13
    safe so Elon can't actually you know
  • 00:24:15
    sort of start the robot Revolution yes
  • 00:24:18
    Jack
  • 00:24:20
    uh you see the future and you're leading
  • 00:24:23
    it and it's just it's 100 Xing it's not
  • 00:24:27
    10 Xing as we start to see you know qlms
  • 00:24:30
    and uh and and uh as we see quantum
  • 00:24:34
    computers coming online is the world
  • 00:24:36
    ready for how much is going to change in
  • 00:24:40
    the next 5 years well I think it's going
  • 00:24:44
    to be a fascinating next five years but
  • 00:24:47
    Peter if I can just give maybe two ideas
  • 00:24:49
    that can share with the audience to help
  • 00:24:50
    absorb what's about to happen first is
  • 00:24:53
    the power of small
  • 00:24:54
    teams what I recommend to every one of
  • 00:24:57
    us and what we're practic in is that
  • 00:25:00
    armed with llms and lqm armed with these
  • 00:25:02
    new AI tools that are from the panelists
  • 00:25:04
    or from others at this conference and
  • 00:25:06
    elsewhere small teams can change the
  • 00:25:09
    world small teams if you're a big
  • 00:25:12
    company portion off a team of 10 people
  • 00:25:15
    and say you're going to be in this new
  • 00:25:16
    area you have this mission go so your
  • 00:25:19
    moonshot teams yeah and so at sandbox AQ
  • 00:25:21
    just to give you one practical example
  • 00:25:23
    there's a big issue when you try to fly
  • 00:25:25
    a plane now there's no more GPS if you
  • 00:25:26
    try to go to parts of Europe no more G
  • 00:25:28
    GPS if you go to anywhere in the Gulf
  • 00:25:30
    region Saudi you're Landing in Riyad or
  • 00:25:32
    damam if you're Landing in Abu Dhabi or
  • 00:25:34
    Dubai there is no more GPS it's being
  • 00:25:36
    jammed it's being spoofed it's out in
  • 00:25:39
    the Indo paccom area in no Pacific again
  • 00:25:41
    uh PRC China is blocking GPS there an
  • 00:25:44
    11p person team Peter that we armed with
  • 00:25:47
    this kind of AI and some Quantum sensors
  • 00:25:49
    solve the problem it's flying now on the
  • 00:25:51
    United States Air Force and and so small
  • 00:25:54
    teams now are the order of the day as
  • 00:25:57
    managers leaders this is what we're
  • 00:25:59
    doing this is what I think more people
  • 00:26:01
    will start to realize and the second
  • 00:26:03
    thing I would leave you with Peter is
  • 00:26:05
    that I know a lot of people are still
  • 00:26:06
    concerned about AI concerned about what
  • 00:26:09
    are the implications of AI let me also
  • 00:26:11
    say that what we're concerned about is
  • 00:26:13
    people not embracing a ey fast enough to
  • 00:26:16
    solve the big problems of our current
  • 00:26:18
    Society let's hit the big diseases that
  • 00:26:22
    have plagued us and challenged us for 40
  • 00:26:24
    plus years let's bring battery storage
  • 00:26:26
    to a new level going Beyond lithium I on
  • 00:26:28
    going beyond the current chemistries
  • 00:26:30
    this is where we need to really focus
  • 00:26:33
    more embracing yes responsibly of course
  • 00:26:36
    but making sure we lean in and I was
  • 00:26:38
    pleased to see at the Paris AI Summit
  • 00:26:40
    that we all just came from that there
  • 00:26:42
    was a lean in Attitude rather than to
  • 00:26:44
    years ago when people are like oh no
  • 00:26:46
    should we even touch this stuff so small
  • 00:26:48
    teams and let's lean in and let's solve
  • 00:26:50
    the big problems in society now amazing
  • 00:26:53
    tier for that nice job Jack thank you
  • 00:26:58
    Jim what do you want people to take away
  • 00:27:01
    from the work that you're doing what
  • 00:27:03
    should they remember how should they
  • 00:27:05
    utilize the tech that you're building
  • 00:27:08
    yeah so so AI doesn't have to be
  • 00:27:11
    unbelievably expensive unbelievably big
  • 00:27:14
    unbelievably
  • 00:27:15
    proprietary like that's not required
  • 00:27:18
    like the the computational hardware is
  • 00:27:20
    fairly straightforward and we want to
  • 00:27:22
    make that available to lots of people so
  • 00:27:24
    they can use it I I think there is going
  • 00:27:27
    to be a big up level on how we build and
  • 00:27:29
    write software and build machines it it
  • 00:27:31
    shouldn't take two years to build a
  • 00:27:32
    computer right we want to pull that down
  • 00:27:34
    it shouldn't take $10,000 to buy a
  • 00:27:37
    single chip we're going to take that
  • 00:27:38
    down drastically right how much cheaper
  • 00:27:41
    is the systems You're Building compared
  • 00:27:42
    to our Target is 5 to 10x cheaper I'm
  • 00:27:45
    sorry 5 to 10x cheaper 5 to 10x cheaper
  • 00:27:47
    than the current systems and then we
  • 00:27:49
    have a road map to continue to make that
  • 00:27:51
    better and then the other piece is like
  • 00:27:54
    you have to I I like the big swing
  • 00:27:56
    approach on using AI
  • 00:27:58
    but start small right like I'm asking my
  • 00:28:01
    software team to double their
  • 00:28:02
    productivity this year and everybody's
  • 00:28:04
    starting to use the the code generators
  • 00:28:07
    the code helpers we have we're building
  • 00:28:09
    our own tools to go check the quality
  • 00:28:11
    and and just start working on it and get
  • 00:28:13
    used to it because like you're right if
  • 00:28:16
    your system's broken like patching up
  • 00:28:18
    the broken system isn't quite right but
  • 00:28:21
    getting a real feel for it and using it
  • 00:28:23
    and then starting to iterate on how your
  • 00:28:25
    system works is really important and I I
  • 00:28:27
    think everybody body should you know
  • 00:28:29
    dive in and embrace it but we don't have
  • 00:28:31
    to solve world peace first I would like
  • 00:28:33
    to make my code have a few less bugs
  • 00:28:36
    Alexander uh last words from you uh who
  • 00:28:39
    typically comes as a customer to Quantum
  • 00:28:41
    black and what is your value proposition
  • 00:28:44
    that you offer them is it we're going to
  • 00:28:45
    solve understand your problems and solve
  • 00:28:47
    your problems no basically first of all
  • 00:28:50
    and it goes back to a question of what
  • 00:28:51
    needs to be done the customer is the
  • 00:28:53
    chairman of the CEO unless number one in
  • 00:28:57
    or head of the state is really
  • 00:28:59
    interested in this problem and really
  • 00:29:01
    going to invest her his time don't waste
  • 00:29:04
    your time it's not going to work we
  • 00:29:05
    never going to do it right I mean it's
  • 00:29:07
    so important right unless you got Buy in
  • 00:29:09
    from the very top of your organization
  • 00:29:11
    to be prepared to make yourself an AI
  • 00:29:14
    first I me one of the biggest challenges
  • 00:29:15
    a lot of companies have here you're not
  • 00:29:17
    competing against your other typical
  • 00:29:20
    companies you're competing against the
  • 00:29:22
    startup that is AI native from the
  • 00:29:25
    beginning and it starts with your own
  • 00:29:27
    literacy because I think this room and
  • 00:29:29
    we grew up under the Paradigm that
  • 00:29:32
    unless I could explain cannot explain
  • 00:29:33
    you something in two minutes I'm
  • 00:29:35
    probably in competent and it's all right
  • 00:29:37
    but we need to speak the same language
  • 00:29:39
    and the first thing is kind of go and
  • 00:29:41
    study language and while we could claim
  • 00:29:43
    that AI failed in many things what you
  • 00:29:45
    could clearly see the drop of AI during
  • 00:29:48
    summer and you ask yourself why because
  • 00:29:51
    all the kids are out of school or
  • 00:29:52
    university right because AI was the best
  • 00:29:54
    tutor to this world so first of all use
  • 00:29:56
    AI to educate yourself kind of that's
  • 00:29:58
    number one number two why I'm saying
  • 00:30:00
    it's a leadership challenge because what
  • 00:30:02
    you need to do is really to go fully in
  • 00:30:06
    to transform the Enterprise you need to
  • 00:30:08
    get the data right that never is right
  • 00:30:11
    you need to change your architecture
  • 00:30:13
    that essentially means changing politics
  • 00:30:15
    within the organization and we don't
  • 00:30:16
    like to change politics then hopefully
  • 00:30:19
    you need to hire good people but then
  • 00:30:22
    teach rest of the body of the church
  • 00:30:24
    what the hell is it and how to ensure
  • 00:30:26
    you embrace it and then all the sudden
  • 00:30:29
    you have team with human beings and
  • 00:30:30
    certain agents and kind of how do you
  • 00:30:32
    operate it yourself now you multiply the
  • 00:30:34
    likelihood of these things and
  • 00:30:36
    likelihood to success unless you believe
  • 00:30:39
    and go fully in it's risky because you
  • 00:30:42
    put your career you put your company
  • 00:30:43
    future on the line and you cannot go
  • 00:30:46
    small you need to go big to succeed and
  • 00:30:48
    that's what we're trying to do just use
  • 00:30:50
    AI as the way to make the world into the
  • 00:30:52
    better place ladies and Gentlemen let's
  • 00:30:54
    give it up for this incredible panel
  • 00:30:56
    thank you all so much thank you
  • 00:30:59
    okay
  • 00:31:08
    [Music]
Tag
  • AI
  • Generative AI
  • Quantum Technology
  • Stability AI
  • Liquid AI
  • Open Source
  • Enterprise Transformation
  • Education
  • Healthcare
  • Creativity