The Future of AI in B2B SaaS: Insights from Synthesia and Theory Ventures. Hosted by Jason Lemkin

00:51:02
https://www.youtube.com/watch?v=F_y3Zk9A8t4

Zusammenfassung

TLDRThe conversation discusses the transformative effect of AI on B2B applications, highlighting reduced costs and improved gross margins due to advancements in AI technologies. Speakers note that AI companies are growing faster than traditional SaaS businesses. AI video platforms like Synthesia are cited as examples of how technology can enhance corporate training and communication by streamlining video production. The funding landscape for AI companies has seen significant changes, with increased investments amid concerns about revenue sustainability. The dialogue also addresses the potential for AI to replace lower-tier employees while augmenting the capabilities of remaining staff, and emphasizes the need for businesses to adapt their pricing models in response to these developments.

Mitbringsel

  • 💡 AI is significantly lowering costs for B2B applications, especially in inference.
  • 📈 AI companies are experiencing faster growth rates compared to traditional SaaS firms.
  • 🎥 Synthesia exemplifies how AI video platforms are revolutionizing corporate training.
  • 🤑 Investors are increasingly interested in sustainable revenue models in AI.
  • ⚠️ There are concerns about the longevity of AI revenue streams.
  • 👥 AI has the potential to reduce the number of hires needed in businesses.
  • 🤖 AI agents may replace lower-tier employees and enhance productivity of higher-tier staff.
  • 💼 Pricing strategies for AI services may need to be reevaluated based on delivered value.
  • 🔄 Managing multiple AI tools could lead to cognitive overload for organizations.
  • 📊 The future may see a reallocation of budgets from labor to AI solutions.

Zeitleiste

  • 00:00:00 - 00:05:00

    The discussion opens with the potential for B2B apps to become significantly cheaper due to advancements in AI, particularly in inference costs decreasing, leading to improved gross margins for AI-enabled applications compared to traditional SaaS models. AI companies have been experiencing rapid growth and increased valuations, presenting a significant opportunity for the application layer in business technology.

  • 00:05:00 - 00:10:00

    As part of the digital AI day for 2025, the conversation dives into AI's intersection with B2B, referencing Synthesia's recent funding round and its focus on AI-driven video and training for enterprises. The speakers discuss trends in investment, the speed of AI growth, and major fundraising success stories in the industry.

  • 00:10:00 - 00:15:00

    Synthesia highlights its platform’s evolution, detailing its use of AI to create avatar-based videos that streamline corporate training and communications. The ability to edit and update content easily has shifted the design philosophy from traditional video production to a more agile, collaborative approach that enables non-experts to create high-quality content.

  • 00:15:00 - 00:20:00

    The platform's capabilities position it to cater to a variety of corporate needs, making it essential for communication between employees, customers, and partners. The team emphasizes video and audio consumption trends over traditional reading, showcasing the platform's versatility in corporate settings.

  • 00:20:00 - 00:25:00

    Key features of the Synthesia platform include an intuitive, PowerPoint-like design that enables non-video specialists to create professional content. Traditional video production processes are being streamlined, increasing efficiency and reducing time spent on content creation.

  • 00:25:00 - 00:30:00

    The discussion transitions to the competitive landscape of AI video platforms. Synthesia aims to consolidate the fragmented video creation value chain into a unified solution, allowing businesses to easily produce, translate, and distribute training and marketing materials in various formats.

  • 00:30:00 - 00:35:00

    Victor elaborates on Synthesia's continued evolution, which includes advancements in avatar quality, more engaging content experiences, and integrating analytics to optimize future video creations based on viewer interactions and data.

  • 00:35:00 - 00:40:00

    A debate arises around whether the rapid rise of AI-enabled applications, particularly in areas like customer service and marketing, will lead to significant cost savings or simply a shift in spending priorities within companies that traditionally invest in services. The potential for deflationary pressures in the SaaS market is discussed, particularly concerning the sustainable revenue impact of AI solutions.

  • 00:40:00 - 00:45:00

    The conversation also touches on the evolving landscape of venture capital funding, where AI-driven companies are becoming increasingly sought after. Investors are focusing on high-quality revenue and sustained client engagement rather than mere hype or transient products. Companies exhibiting strong renewals and expansions are gaining more attention in the battle for capital.

  • 00:45:00 - 00:51:02

    Finally, recruiters and tech leaders conclude that with AI's rise, companies might not need to off-shore as heavily as before, given the productivity enhancements possible through well-implemented AI solutions, enabling teams to leverage technology more efficiently than traditional hires alone.

Mehr anzeigen

Mind Map

Video-Fragen und Antworten

  • How is AI impacting B2B application costs?

    AI is reducing costs associated with software applications, particularly inference costs, potentially leading to higher gross margins.

  • What has changed in the funding landscape for AI companies?

    There has been a surge in funding for AI companies, with investors more willing to back them compared to traditional SaaS businesses.

  • How are AI video platforms like Synthesia changing corporate training?

    Synthesia simplifies video production, making it accessible to all employees, improving efficiency in corporate training and communication.

  • Are investors concerned about the sustainability of AI revenue?

    Yes, there is concern about whether the revenue generated from AI is durable or just a result of discretionary spending.

  • What is the significance of durable revenue in the AI sector?

    Durable revenue indicates a sustainable and robust business model, which is crucial for attracting continued investment.

  • Will AI lead to reduced headcount in companies?

    AI may lead to increased productivity, allowing companies to operate with fewer employees, particularly in roles that can be automated.

  • What are the implications of AI agents in B2B environments?

    AI agents may replace some human roles, particularly lower-tier positions, while also augmenting the capabilities of skilled workers.

  • How are companies adjusting their pricing models for AI services?

    Companies might need to rethink pricing strategies, focusing on value delivered rather than simply the cost basis of services.

  • What does the future hold for the integration of AI in corporate structures?

    As AI becomes more integrated, there may be a shift towards software solutions that enhance human efficiency and enable smoother operations.

  • What is the risk associated with a high number of AI tools?

    Managing multiple AI tools can lead to cognitive overload and complicate procurement processes within organizations.

Weitere Video-Zusammenfassungen anzeigen

Erhalten Sie sofortigen Zugang zu kostenlosen YouTube-Videozusammenfassungen, die von AI unterstützt werden!
Untertitel
en
Automatisches Blättern:
  • 00:00:00
    is this going to make B2B apps an order
  • 00:00:02
    of magnitude cheaper what's the impact I
  • 00:00:05
    think the impact is your gross margins
  • 00:00:07
    are so much better than they used to be
  • 00:00:09
    right like the second biggest cost for
  • 00:00:11
    AI enabled SAS applications has been
  • 00:00:13
    inference and when we were looking like
  • 00:00:16
    a typical software companies probably at
  • 00:00:18
    75% gross margin lot AI software
  • 00:00:21
    application companies are like zero to
  • 00:00:23
    30% and now their single biggest cost
  • 00:00:25
    has been reduced by 95% overnight and
  • 00:00:28
    they didn't have to do a thing so
  • 00:00:30
    now they now now they have ai and you
  • 00:00:33
    Jason you and I both know this like the
  • 00:00:35
    growth rates of AI companies are much
  • 00:00:36
    faster than say the classic SAS
  • 00:00:38
    businesses just because there's so them
  • 00:00:40
    and so now you have companies that are
  • 00:00:42
    growing one to seven or some massive
  • 00:00:45
    growth rate before they used to have
  • 00:00:47
    challenging gross margins and now
  • 00:00:48
    they'll probably look the same as SAS
  • 00:00:50
    companies which means you have pretty
  • 00:00:51
    significant multiple and valuation
  • 00:00:53
    expansion yeah so I think it's at the
  • 00:00:56
    application layer this is like an
  • 00:00:57
    unbelievable Christmas present that that
  • 00:01:00
    came on January 22nd thank you thanks
  • 00:01:03
    for joining our digital AI day for
  • 00:01:06
    2025 we did this last year and with the
  • 00:01:11
    goal of really doing a deep dive at the
  • 00:01:13
    intersection of AI and
  • 00:01:15
    B2B and luckily for our kickoff today
  • 00:01:18
    we've got two folks joining us just to
  • 00:01:20
    have a discussion on what's literally
  • 00:01:23
    going on today and I think it's some
  • 00:01:25
    ways it's so I can't even keep up we'll
  • 00:01:27
    talk a little bit about deep I can't
  • 00:01:28
    even keep up after deep seeking
  • 00:01:30
    everything it felt like a year ago was
  • 00:01:31
    insane we've got Victor rabell from Co
  • 00:01:34
    of Synthesia we'll talk we'll first do a
  • 00:01:36
    little bit of a deep dive there
  • 00:01:37
    Synthesia just raised a big round at two
  • 00:01:41
    and a half billion or something like
  • 00:01:43
    that you got to keep up with with open
  • 00:01:45
    Ai and that's like a seed round these
  • 00:01:49
    days but pretty impressive we'll dig in
  • 00:01:50
    a second but Victor has been at the
  • 00:01:52
    intersection of AI and B2B and
  • 00:01:54
    Enterprise doing AI based video and
  • 00:01:57
    training for Enterprises so no one's
  • 00:01:59
    been thinking about the intersection of
  • 00:02:00
    the two probably more and Victor was
  • 00:02:03
    with us in sass Europa in London last
  • 00:02:05
    summer and I'm going to ask him what's
  • 00:02:07
    changed and obviously we all know Tas
  • 00:02:10
    tungus from now Theory ventures in the
  • 00:02:13
    last 14 months he's raised just only 900
  • 00:02:15
    million dollar or something across two
  • 00:02:17
    funds in 14 months or something like
  • 00:02:18
    that is that about right a little less
  • 00:02:20
    than that closer to 700 but okay the
  • 00:02:22
    paces picked up for AI investors it used
  • 00:02:26
    to be that like 60 million was a lot for
  • 00:02:29
    a first-time fund 50 million now guys
  • 00:02:32
    like to now granted not a new investor
  • 00:02:34
    but a new fund can raise like
  • 00:02:36
    approaching a billion dollars in 24
  • 00:02:38
    months to invest in Ai and data honestly
  • 00:02:41
    it is a derivative of what's changing
  • 00:02:42
    isn't it but check sizes are up lead
  • 00:02:45
    check sizes are up 42% in 18 months yeah
  • 00:02:49
    so I want to dig into that but I can't
  • 00:02:50
    think of anyone to talk about right now
  • 00:02:53
    the intersection and I want to go into
  • 00:02:55
    some of the data with Tomas but before
  • 00:02:57
    we get there
  • 00:03:00
    Victor get us up to speed so you've been
  • 00:03:04
    doing give me a deep dive on Synthesia
  • 00:03:09
    and where it wins but if roughly
  • 00:03:11
    speaking you've got hundreds of
  • 00:03:13
    relatively larger customers using your
  • 00:03:15
    AI training here's you yourself on the
  • 00:03:17
    screen right you using your own product
  • 00:03:21
    doing in many cases nuts and bolts B2B
  • 00:03:24
    activities right corporate training
  • 00:03:26
    sales training external on websites
  • 00:03:28
    internal skos which we thought we
  • 00:03:31
    wouldn't need any salespeople with with
  • 00:03:33
    AI it turns out we need more than ever
  • 00:03:36
    but you're at the at the intersection of
  • 00:03:39
    an area that is competitive right
  • 00:03:41
    there's a there's more than one vendor
  • 00:03:43
    doing video and Enterprise you're on one
  • 00:03:45
    side of the other of nine figure in
  • 00:03:47
    Revenue monetizing this so get us up to
  • 00:03:49
    speed on what you do
  • 00:03:51
    today and actually would be interesting
  • 00:03:53
    to me is like with all the change what
  • 00:03:55
    are you most excited about that you're
  • 00:03:56
    going to release this year yeah thanks
  • 00:03:58
    for for having me
  • 00:04:00
    today we're the largest AI video
  • 00:04:02
    platform for the Enterprise right and
  • 00:04:05
    four years ago that all started with
  • 00:04:07
    Avatar so most the builders are familiar
  • 00:04:09
    with probably think of be being like the
  • 00:04:11
    Avatar company when invent this
  • 00:04:13
    technology when instead of recording
  • 00:04:14
    yourself with a camera you simply create
  • 00:04:17
    an avatar of yourself or use one of our
  • 00:04:18
    kind of off the-shelf avatars you type
  • 00:04:20
    out something and boom magic video comes
  • 00:04:22
    out which obviously is a 100 times more
  • 00:04:24
    efficient than having to record yourself
  • 00:04:26
    and giv you all that the benefits you go
  • 00:04:27
    back and edit it after the fact you can
  • 00:04:29
    translate into different languages and a
  • 00:04:30
    whole bunch of other really cool things
  • 00:04:33
    and where we've gotten to today is that
  • 00:04:35
    is still for sure a very important
  • 00:04:37
    headline feature of sesia but really we
  • 00:04:39
    are on AI video platform today and we
  • 00:04:41
    help our customers predominantly in the
  • 00:04:43
    enterprise we also have a free model
  • 00:04:44
    self-service motion but been surpris
  • 00:04:46
    what we're focusing on communicating
  • 00:04:48
    better and that could be with our
  • 00:04:50
    employees could be with customers could
  • 00:04:52
    be with Partners in 2024 people want to
  • 00:04:54
    consume their content in video or audio
  • 00:04:56
    format they do not want to read anymore
  • 00:04:58
    right there's a lot of tech entprise
  • 00:04:59
    it's true what you very successful at
  • 00:05:01
    doing is building um video tool right
  • 00:05:05
    which has Avatar as a component as AI
  • 00:05:07
    boes as a component as an AI screen
  • 00:05:09
    recorder as a component but everything
  • 00:05:11
    is put into this canva PowerPoint style
  • 00:05:14
    flow right and so reason you're growing
  • 00:05:16
    so fast is because what we essentially
  • 00:05:18
    have done is we've enabled any white
  • 00:05:21
    collar office worker who knows how to
  • 00:05:22
    make a PowerPoint to make a video right
  • 00:05:25
    before eia you have to go to the video
  • 00:05:27
    production Department with your idea
  • 00:05:29
    they would just rank it against 10 other
  • 00:05:30
    ideas they'd say no Jason your idea is
  • 00:05:32
    not important enough we'll do what the
  • 00:05:33
    marget team wants to do and you have to
  • 00:05:37
    go through all that stuff right if you
  • 00:05:38
    did get man to get a video through and
  • 00:05:40
    you're doing customer support then one
  • 00:05:42
    month after you've made 20 beautiful
  • 00:05:44
    customer support videos they're all out
  • 00:05:46
    of date because you Chang the product
  • 00:05:47
    name the policies they're all out of
  • 00:05:49
    date their friend everything's out of
  • 00:05:51
    date right and that's a horrible thing
  • 00:05:52
    to change but with this new uh platform
  • 00:05:55
    that we've buil right you can just go
  • 00:05:56
    back and edit it after the fact which is
  • 00:05:57
    incredibly powerful and so I think what
  • 00:05:59
    we've grabed graduated from over the
  • 00:06:00
    last four years right as being the
  • 00:06:01
    training company with avatars now being
  • 00:06:04
    the Enterprise video platform that's
  • 00:06:05
    what we're building and the way we think
  • 00:06:07
    of that is if you look at the video
  • 00:06:09
    value chain no one has ever managed to
  • 00:06:11
    build a really big company about
  • 00:06:13
    Enterprise video and we think the reason
  • 00:06:15
    for that is that the value chain has
  • 00:06:16
    been completely disjointed right you
  • 00:06:18
    have a camera where you shoot some
  • 00:06:19
    content with you go to Adobe to edit it
  • 00:06:22
    you go to frame.io to collaborate on it
  • 00:06:24
    you'd go to Vimeo to then publish that
  • 00:06:25
    content I maybe YouTube depending on
  • 00:06:28
    what your goals are and what we've done
  • 00:06:29
    is we've collapsed all that into one
  • 00:06:30
    platform give you all the tools to make
  • 00:06:32
    the videos and it's so easy that anyone
  • 00:06:34
    can do it it's a full modern
  • 00:06:36
    collaborative platform right it's a
  • 00:06:38
    Content management system where you can
  • 00:06:40
    have teams and organizations and your
  • 00:06:42
    CEO is happy because we have audit love
  • 00:06:44
    all the sort of stuff you need right
  • 00:06:46
    it's collaborative and you can translate
  • 00:06:48
    update the content then we also give you
  • 00:06:49
    a publishing platform which is a
  • 00:06:50
    relatively new product for us it gives
  • 00:06:52
    you a video player we can deliver the
  • 00:06:54
    content to your end users that's
  • 00:06:55
    exciting because when we owe that entire
  • 00:06:57
    video value chain we can year begin to
  • 00:07:00
    do a lot of things you couldn't do if we
  • 00:07:01
    hadn't bundled it for example we can
  • 00:07:03
    begin to use the engagement data of the
  • 00:07:05
    videos to automatically create new
  • 00:07:06
    versions of the video that perform
  • 00:07:07
    better than the previous one we can make
  • 00:07:09
    your videos interactive by directional
  • 00:07:11
    communication which of course is think
  • 00:07:13
    of have ai agents right but you can it's
  • 00:07:16
    not just a video that goes plays from A
  • 00:07:17
    to B it's like an experience it's like a
  • 00:07:20
    small app almost you can right with so
  • 00:07:22
    really what's exciting here is is the
  • 00:07:24
    future of communication which I think
  • 00:07:25
    you look at people's private lives
  • 00:07:27
    everybody listens to podcast and what
  • 00:07:29
    Tik Tok videos and YouTube videos but
  • 00:07:31
    when we go to work did the buying
  • 00:07:32
    process in as a customer I you're
  • 00:07:35
    receiving support somewhere all that
  • 00:07:36
    stuff's T we helping customers and we
  • 00:07:39
    helping Enterprise essentially up L the
  • 00:07:41
    communications which which drives all
  • 00:07:43
    sorts of different business outcomes
  • 00:07:44
    depending on our customers so you're
  • 00:07:46
    saying that folks that have blogged for
  • 00:07:48
    a while like Tomas and I were just it's
  • 00:07:50
    dead like there like we just we should
  • 00:07:52
    just it's it's over it's it's distant
  • 00:07:55
    past time to hang it up there actually
  • 00:07:57
    are many bloggers left okay one quick
  • 00:07:59
    question on this and then then we can
  • 00:08:01
    dig in a little bit in into today and I
  • 00:08:02
    want let's use your round in a second as
  • 00:08:04
    a case study let's use it with toas but
  • 00:08:07
    what is some with the latest models
  • 00:08:09
    latest everything we can talk about deep
  • 00:08:11
    seek for a moment in the next what's
  • 00:08:13
    something you're excited that you can do
  • 00:08:14
    this year in an upcoming release that
  • 00:08:16
    maybe you couldn't even do last year
  • 00:08:17
    what is something literally that's at
  • 00:08:19
    the bleeding edge that you're doing for
  • 00:08:20
    real in the
  • 00:08:21
    Enterprise so we're launching our new a
  • 00:08:24
    models within the next 3 to six month
  • 00:08:27
    and they have a massive upgrade both in
  • 00:08:29
    terms of quality and Fidelity of the
  • 00:08:31
    Avatar so those of you who tried the
  • 00:08:33
    product will probably have experienced
  • 00:08:34
    the avatars are great they're lot better
  • 00:08:36
    the last 12 to 18 months than they
  • 00:08:38
    previous to that but there's still some
  • 00:08:39
    way to go before it's completely photo
  • 00:08:41
    real right we got a pass through that
  • 00:08:43
    barrier and we're also done a you to
  • 00:08:45
    basically promp your avatar into new
  • 00:08:47
    environment for example if you create an
  • 00:08:49
    avatar for yourself today you're stuck
  • 00:08:50
    wherever you would call it your avatar
  • 00:08:52
    in your office for example what you can
  • 00:08:54
    do with this new tech is you can say hey
  • 00:08:55
    put me in the jungle put me in the
  • 00:08:57
    supermarket put me like I'm playing
  • 00:08:59
    tennis or something like that that opes
  • 00:09:00
    up an entirely new creative dimension in
  • 00:09:02
    the product and so those two things
  • 00:09:03
    together have always been one of the
  • 00:09:05
    core mechanisms of of suia right the
  • 00:09:07
    better the quality of the avatars and
  • 00:09:09
    voices the more T we unlock so a lot of
  • 00:09:11
    people laughed that us four years ago
  • 00:09:13
    for going out the training market and
  • 00:09:15
    because it's not a very sexy Market VCS
  • 00:09:17
    hate Learning and Development I get it
  • 00:09:20
    but I get it maybe didn't sound like the
  • 00:09:22
    hottest face right we get it but a
  • 00:09:24
    midsize Market in all fairness right
  • 00:09:26
    even it's midsized right classically I
  • 00:09:29
    think more importantly than that you
  • 00:09:31
    could get to double digit AR just
  • 00:09:34
    serving that market and the thing that
  • 00:09:35
    we I think saw that are with there so
  • 00:09:37
    many graveyard competitors that have
  • 00:09:40
    died around us over the last eight years
  • 00:09:41
    was that we focused on a use case where
  • 00:09:43
    the technology was actually good enough
  • 00:09:45
    not where we could dream it to be one
  • 00:09:47
    day but do doing all sorts of cool
  • 00:09:49
    marketing and sales content it worked
  • 00:09:51
    well for training because the
  • 00:09:52
    alternative to training was a boring PDF
  • 00:09:55
    document and so what we've seen over the
  • 00:09:57
    last couple of years about is that the
  • 00:09:58
    better the Avatar quality comes the
  • 00:09:59
    bigger the tan get so the last 12 to 18
  • 00:10:01
    months we've seen a massive shift
  • 00:10:03
    towards product marketing with customer
  • 00:10:05
    support which collectively those two are
  • 00:10:07
    now actually the biggest uh segments
  • 00:10:08
    that we serve and that's purely because
  • 00:10:10
    the quality of the Avatar have gotten
  • 00:10:11
    better and better we're not making like
  • 00:10:13
    we're not making Super Bowl ads and key
  • 00:10:15
    brand moments but make a lot of tactical
  • 00:10:17
    marketing content right you see the cool
  • 00:10:20
    ad on on meta YouTube whatever you click
  • 00:10:21
    to the website and then once you're on
  • 00:10:23
    the website then you learn more about
  • 00:10:25
    the product with some easier videos and
  • 00:10:27
    we could just move further and further
  • 00:10:28
    up that f the better the quality of
  • 00:10:30
    theat so the further you leave The
  • 00:10:33
    Uncanny Valley the customers can see it
  • 00:10:35
    and you're see more Market Pole
  • 00:10:37
    right yeah cool all right I want to dig
  • 00:10:40
    into that a little bit more but before
  • 00:10:43
    we do just because we agenda changed in
  • 00:10:46
    a week deep seek right I I saw a lot of
  • 00:10:49
    folks talking about this on social media
  • 00:10:51
    and X I don't know if you guys did I
  • 00:10:52
    want to get both of your opinions but
  • 00:10:54
    and listen I'm not an I'm the least
  • 00:10:56
    smart person probably of all the folks
  • 00:10:58
    that are watching
  • 00:10:59
    but I have an AI test and I want to know
  • 00:11:01
    what you got It's called The Sopranos
  • 00:11:03
    test and I've been running this for a
  • 00:11:04
    while okay I'm a basic guy and but I
  • 00:11:07
    enjoy writing in the word and I go in
  • 00:11:09
    and I ask what happens after Victor do
  • 00:11:11
    you watch Sopranos over there in Europe
  • 00:11:13
    have you seen the show probably don't
  • 00:11:14
    know anyhow famous show in the United
  • 00:11:17
    States and it's crazy at the end of the
  • 00:11:19
    now I there is a rule I think after 20
  • 00:11:21
    years there's no spoilers right I think
  • 00:11:23
    that's the rule on Reddit okay so in the
  • 00:11:25
    very end the screen goes black and
  • 00:11:26
    people have debated this for 20 years
  • 00:11:28
    what happens in and so I asked I asked
  • 00:11:31
    chat GPT and I asked and I ask every and
  • 00:11:34
    I asked Claude And I asked deepy and
  • 00:11:36
    deep c one it's the best ending it was
  • 00:11:39
    the best ending that doesn't mean that
  • 00:11:41
    it gives me my best diagnosis for why my
  • 00:11:43
    cholesterol is high or others but that
  • 00:11:45
    was my moment where it was pretty
  • 00:11:47
    cool for this one use case it it seemed
  • 00:11:51
    like instant parodies first toas what
  • 00:11:54
    what's your take in general and is this
  • 00:11:57
    re it's only been a couple days days
  • 00:11:59
    right is this going to make B2B apps an
  • 00:12:02
    order of magnitude cheaper what's the
  • 00:12:04
    impact I think the impact is gross
  • 00:12:07
    margins are so much better than they
  • 00:12:09
    used to be right like the second biggest
  • 00:12:10
    cost for AI enabled SAS applications has
  • 00:12:13
    been inference and when we were looking
  • 00:12:16
    like a typical software companies
  • 00:12:17
    probably at 75% gross margin lot AI
  • 00:12:21
    software application companies are like
  • 00:12:22
    0o to 30% and now their single biggest
  • 00:12:25
    cost has been reduced by 95% overnight
  • 00:12:28
    and they didn't have to do with thing so
  • 00:12:30
    now they now now they have ai and you
  • 00:12:33
    Jason you and I both know this the
  • 00:12:35
    growth rates of AI companies are much
  • 00:12:36
    faster than say the classic SAS
  • 00:12:38
    businesses just because there's so much
  • 00:12:40
    them and so now you have companies that
  • 00:12:42
    are growing one to seven or some massive
  • 00:12:45
    growth rate before they used to have
  • 00:12:47
    challenging gross margins and now
  • 00:12:48
    they'll probably look the same as SAS
  • 00:12:50
    companies which means you have pretty
  • 00:12:51
    significant multiple and valuation
  • 00:12:53
    expansion yeah so I think it's at the
  • 00:12:56
    application layer this is like an
  • 00:12:58
    unbelievable Chris present that that
  • 00:13:00
    came on January 22nd I want to share a
  • 00:13:03
    story but Victor you will this is your
  • 00:13:05
    cogs is your cogs a huge issue and is it
  • 00:13:07
    g to plummet or just from the
  • 00:13:08
    perspective since Asia is it's not as
  • 00:13:10
    big a deal as it sounds for from cogs
  • 00:13:12
    and the rest what's your learnings in 48
  • 00:13:14
    hours we have great so Madness already
  • 00:13:16
    so I wouldn't say it's not something we
  • 00:13:17
    like put a lot of of Fus and emphasis to
  • 00:13:19
    what but obviously to scale up it's
  • 00:13:21
    great I would say though I think I
  • 00:13:24
    totally agree with Thomas it's a massive
  • 00:13:25
    technical breakthrough I think it's
  • 00:13:27
    interes when you look at the market I
  • 00:13:28
    think if anything to me this is just the
  • 00:13:30
    confirmation that humans and markets are
  • 00:13:31
    driven by stories and emotions not
  • 00:13:34
    fundamentals right it's absolutely
  • 00:13:36
    hilarious to see all the world's Tech
  • 00:13:38
    analysts who know nothing about AI just
  • 00:13:39
    panicking because it's like a new model
  • 00:13:41
    out I think people have a tendency to
  • 00:13:43
    overestimate the impact of this in the
  • 00:13:45
    short term and maybe underestimated in
  • 00:13:47
    the long term I think the long-term tail
  • 00:13:49
    effect as Thomas just said for the
  • 00:13:51
    application there is huge I think that's
  • 00:13:53
    really massive but I think in two months
  • 00:13:55
    I think 99.9% of the people who using
  • 00:13:58
    anic open eye to is going to go back to
  • 00:14:00
    using their products not deep seeks
  • 00:14:02
    model and there's a big differentiation
  • 00:14:04
    between those two right I tweeted the
  • 00:14:06
    other day that smart people talk about
  • 00:14:09
    models and benchmarks and customers talk
  • 00:14:11
    about app flows apps and and workflows
  • 00:14:13
    right and there's such a big difference
  • 00:14:15
    of a model that performs better and it's
  • 00:14:17
    interesting technically right they have
  • 00:14:18
    a big again longterm effect in the
  • 00:14:21
    market but but uh but yeah I think this
  • 00:14:24
    kind of the panic that's going on right
  • 00:14:26
    now I think is pretty funny I don't
  • 00:14:28
    think that most people who are using
  • 00:14:30
    open the fry today is going to switch
  • 00:14:32
    over to deep seek magic de in the next
  • 00:14:33
    couple weeks I'll tell you just my and
  • 00:14:36
    again I'm the least smart my but my gut
  • 00:14:39
    just from my little portfolio little s
  • 00:14:40
    of the world is I think cost is less
  • 00:14:43
    important in B2B in some cases then that
  • 00:14:46
    you can use
  • 00:14:47
    more because when I look at a couple
  • 00:14:50
    companies one one I invested in the
  • 00:14:52
    contact center which is all about AI
  • 00:14:54
    agents called gorgeous free Commerce
  • 00:14:55
    they're at almost 100 million okay at
  • 00:14:57
    first the AI was a million bucks a month
  • 00:15:00
    right to deploy these agents now it's
  • 00:15:01
    fine they've optimized it and then I was
  • 00:15:04
    a small investment at the low end of
  • 00:15:06
    video in a company called Opus Pro Opus
  • 00:15:08
    clip that Victor probably knows and when
  • 00:15:10
    I invested I'm like this company's got
  • 00:15:11
    to be burning you're uploading billions
  • 00:15:13
    of videos and you're creating clips for
  • 00:15:15
    free like this business model can't work
  • 00:15:18
    and I look and when I meant oh it's
  • 00:15:20
    actually cash La positive right and it
  • 00:15:22
    was wildly expensive but as soon as they
  • 00:15:24
    made huge commits and optimized what
  • 00:15:26
    they're doing and at the corner Maybe
  • 00:15:29
    even cut a few Corners not literally
  • 00:15:31
    it's a great product but maybe didn't
  • 00:15:32
    have to do everything right they don't
  • 00:15:33
    render all the videos in real time they
  • 00:15:36
    adapted so my thought is a lot of B2B
  • 00:15:38
    folks I know I don't think salesforce's
  • 00:15:40
    gross margins are going to plummet to
  • 00:15:41
    20% because of agent force but we might
  • 00:15:43
    be able to make these products in one
  • 00:15:45
    year like 10 times better whether it's
  • 00:15:47
    the getting rid of the last bit of the
  • 00:15:49
    candy Valley or giving agents out for
  • 00:15:52
    free I you guys probably know this but
  • 00:15:55
    when I talked with Brian hallan the
  • 00:15:57
    other day I didn't realize dares right H
  • 00:15:59
    spot isn't charging for their agents and
  • 00:16:01
    this is a related point I don't know
  • 00:16:03
    what you think toas I think what if
  • 00:16:06
    everything gets better and cheaper we
  • 00:16:07
    just might not be able to justify $2 for
  • 00:16:09
    a resolution it might have to be
  • 00:16:12
    included you think it'll be bundled
  • 00:16:14
    because it's so inexpensive God
  • 00:16:16
    hubspot's not sure they can charge Mark
  • 00:16:18
    benof is but it's early it's early
  • 00:16:20
    customers aren't stupid and I think
  • 00:16:22
    we've all gotten ripped off the last two
  • 00:16:24
    and a half years with endless price
  • 00:16:25
    increases endless price increases and
  • 00:16:29
    think another one now you're crappy if
  • 00:16:31
    your agents mediocre I ain't going to
  • 00:16:32
    pay for it I think it might be a VC
  • 00:16:34
    fantasy just like the fact going to
  • 00:16:36
    Victor's point just like deep seek has
  • 00:16:38
    changed the world in 48 hours it might
  • 00:16:41
    be a VC fantasy that every B2B company
  • 00:16:43
    can charge 50 to 70% more because of
  • 00:16:45
    these additional per resolution charges
  • 00:16:48
    I just don't know but I actually think
  • 00:16:49
    it might be deflationary in SAS for the
  • 00:16:51
    first time it's much more difficult for
  • 00:16:54
    incumbents who are adding on AI
  • 00:16:55
    functionalities that they need to charge
  • 00:16:57
    for to not erod their is essentially
  • 00:16:59
    much easier for start who build a
  • 00:17:01
    product where it is the product I think
  • 00:17:04
    ultimately you don't pay for the models
  • 00:17:05
    right you pay for the product and I
  • 00:17:08
    think it's just for AG and fall for some
  • 00:17:09
    of these big things I think frankly a
  • 00:17:11
    lot of them I don't really not sure they
  • 00:17:13
    figured out exactly where this stuff is
  • 00:17:15
    valuable there's a lot of hey we're
  • 00:17:17
    doing like cool things and everyone
  • 00:17:18
    agrees that it's cool but does it
  • 00:17:19
    actually impact your business whereas I
  • 00:17:21
    think some of those smaller AI native
  • 00:17:23
    companies I put that category that you
  • 00:17:26
    found a very real problem with build
  • 00:17:27
    very real business value
  • 00:17:29
    and then you know it's built into the
  • 00:17:31
    built into our business model already
  • 00:17:33
    right but I everything is co-pilot right
  • 00:17:35
    you really going to pay a lot more for
  • 00:17:36
    copile for your employees like clearly
  • 00:17:38
    that seems to not be the case for a lot
  • 00:17:40
    of big companies because they just can't
  • 00:17:41
    B point value yeah I don't know what do
  • 00:17:45
    you think about I just don't I I'm just
  • 00:17:46
    trying to learn in real time I don't
  • 00:17:47
    know I just don't know in a year if
  • 00:17:49
    everyone will be able to charge another
  • 00:17:51
    50% for Enterprise applications for
  • 00:17:53
    their AI add-ons and agents they may be
  • 00:17:55
    forced to especially if the cost goes on
  • 00:17:59
    another I don't know exactly how the
  • 00:18:01
    economics really work at Deep seek but
  • 00:18:02
    if it's really an order of magnitude
  • 00:18:04
    cheaper right how long how often can you
  • 00:18:06
    charge two bucks for what costs a
  • 00:18:09
    fraction of a cent to deliver yeah I
  • 00:18:11
    maybe but like software has never been
  • 00:18:13
    cost based pricing right twio okay fine
  • 00:18:16
    they they stack a margin on top of the
  • 00:18:17
    SMS cost to the
  • 00:18:19
    underlying stripe okay but like a lot of
  • 00:18:22
    apis are cost based but keep going I'm
  • 00:18:24
    I'm with you I'm with gain site they're
  • 00:18:27
    not marking up AWS store right or
  • 00:18:30
    compute so I hear what you're saying I I
  • 00:18:33
    think but I I do think the point that
  • 00:18:35
    you made at the beginning I completely
  • 00:18:36
    agree with which is if deep seek like
  • 00:18:39
    models with great reasoning 95% cheaper
  • 00:18:42
    yeah software companies will just like
  • 00:18:43
    10x for 100x the use and probably keep
  • 00:18:46
    the margins the same to get to a much
  • 00:18:48
    better user experience as a way of
  • 00:18:49
    differentiation right so I think that
  • 00:18:51
    the ux our expectations as users for
  • 00:18:54
    what these things are going to do are
  • 00:18:55
    going to go way up because they become
  • 00:18:57
    increasingly sophis iated and if you
  • 00:19:00
    play with deep seek I was trying to
  • 00:19:01
    create a flow I took the Deep seek paper
  • 00:19:03
    the academic paper and I put it in deep
  • 00:19:05
    seek and I said create a flowchart for
  • 00:19:06
    me of what they did hard read an
  • 00:19:09
    academic paper it's spent two and a half
  • 00:19:11
    minutes thinking and I was watching it
  • 00:19:12
    at first it said okay like the user said
  • 00:19:15
    create a single chart but the
  • 00:19:16
    experiments actually have three branches
  • 00:19:18
    so now I have to decide do I create one
  • 00:19:20
    single chart or do I create a chart with
  • 00:19:21
    three different branches that like what
  • 00:19:24
    back and forth and now I understand okay
  • 00:19:26
    well next time I asked it to make it
  • 00:19:28
    chart I need to be clear on this point
  • 00:19:30
    just the same way that we learn how to
  • 00:19:31
    use Google we're learning how to use
  • 00:19:32
    these models and this is one of the
  • 00:19:33
    first models that narrates the way that
  • 00:19:35
    it thinks and I think that way we'll
  • 00:19:38
    start to get to more outcomes and then I
  • 00:19:39
    think the other question that's really
  • 00:19:42
    relevant to this conversation is are we
  • 00:19:44
    selling outcomes now in software as
  • 00:19:46
    opposed to selling
  • 00:19:47
    seats yeah and I don't know the answer
  • 00:19:50
    like but the Mantra has been we're
  • 00:19:51
    selling outcomes we're selling outcomes
  • 00:19:53
    we're selling leads not not SDR seeds
  • 00:19:57
    we're selling resolve customer support
  • 00:19:59
    cases or videos or or clicks and if that
  • 00:20:04
    we do get to that place then I think it
  • 00:20:06
    doesn't matter how cheap these things
  • 00:20:07
    become because there's some break even
  • 00:20:10
    for a customer to use the AI rather than
  • 00:20:12
    yes it's funny I I dug up I have a post
  • 00:20:16
    coming out up in my que but I dug up I
  • 00:20:18
    went back through a bunch of those old
  • 00:20:19
    KeyBank SAS studies right and it turns
  • 00:20:22
    out going back to the beginning per seat
  • 00:20:25
    was less than 50% of SAS Revenue when
  • 00:20:27
    you add API price other pricing just
  • 00:20:30
    because we think everyone talks about
  • 00:20:31
    blah blah blah Salesforce it doesn't
  • 00:20:33
    mean it's most of software and so it's
  • 00:20:35
    it was like 47% in like 2020 or 2021
  • 00:20:38
    okay so I'm not saying it won't it
  • 00:20:40
    continue to decline but VC shooting from
  • 00:20:44
    the hip on social media saying the seat
  • 00:20:46
    is dead the seat was never was at best a
  • 00:20:48
    plurality of Revenue and but I wonder I
  • 00:20:53
    have just one investment I've made where
  • 00:20:56
    the customer was playing in kind of a
  • 00:20:58
    hybrid model model right and listen this
  • 00:21:00
    is not like the most profound Insight
  • 00:21:01
    but it crossed a million bucks that they
  • 00:21:03
    were paying very quickly and of course
  • 00:21:04
    what happened the next day hey Victor we
  • 00:21:07
    love but a billion bucks we we need a
  • 00:21:10
    flat R to like no no more of this like
  • 00:21:13
    variable C it destroys the Enterprise no
  • 00:21:15
    enterpr no matter what and maybe you
  • 00:21:16
    guys both have insights Enterprises
  • 00:21:19
    still want um visibility there is the
  • 00:21:21
    ability to move some budget around there
  • 00:21:23
    is ability to take budget a piece but
  • 00:21:26
    everyone wants certainty right everyone
  • 00:21:28
    want everyone everyone wants Snowflake
  • 00:21:30
    as much as they can in their snowflake
  • 00:21:32
    spin they want snowflake here right so I
  • 00:21:34
    don't know if this is as radical as the
  • 00:21:36
    internet wants it to be how do your
  • 00:21:38
    customers want to pay Victor do they
  • 00:21:40
    want to do they want to pay in brand new
  • 00:21:41
    ways or they want to pay in the same way
  • 00:21:42
    their grandfather paid for business
  • 00:21:44
    sofware we still sell seeg predominantly
  • 00:21:46
    I think it depends just a lot of the
  • 00:21:48
    product right I think the way to think
  • 00:21:49
    about sesia right now is Powerpoint 2.0
  • 00:21:52
    right and like PowerPoint is one of
  • 00:21:54
    those things where we have thousands of
  • 00:21:55
    people using this inside big Enterprises
  • 00:21:58
    and and I think that's that's easy to
  • 00:22:01
    understand why we're selling seats I
  • 00:22:03
    think we all want to sell outcom we are
  • 00:22:05
    building kind of more specific workflows
  • 00:22:07
    that solve you know redu deflecting
  • 00:22:09
    customer support calls to the call
  • 00:22:12
    center or increasing in yourself
  • 00:22:14
    prospect thing which would build
  • 00:22:16
    separately but I think it just m I think
  • 00:22:19
    it really just depends a lot on on the
  • 00:22:20
    product if it's I think there's probably
  • 00:22:23
    a lot of companies who over index on
  • 00:22:26
    seeds where it doesn't really make sense
  • 00:22:27
    in a Creator tool I think seats make
  • 00:22:29
    sense right but a whole bunch of other
  • 00:22:30
    types of tools where seats maybe isn't
  • 00:22:33
    like the right is is isn't the right
  • 00:22:36
    mity yeah I don't know I I want to move
  • 00:22:38
    on to fundraising I just my only thought
  • 00:22:40
    is for Founders and others I just think
  • 00:22:43
    just got it you just I would just be a
  • 00:22:45
    little cautious on social media think
  • 00:22:47
    because I don't think most business
  • 00:22:50
    customers want pricing that is radically
  • 00:22:52
    different than their last product right
  • 00:22:55
    if I paid this much if I'm paying 300
  • 00:22:57
    bucks per month for sales course and I'm
  • 00:22:59
    moving to this new version like 200
  • 00:23:02
    bucks seems
  • 00:23:03
    cheap but you want 83 cents per contact
  • 00:23:08
    resolution it's just it's that may add
  • 00:23:11
    friction to the sales process right and
  • 00:23:13
    there are certain ways and if I'm I
  • 00:23:14
    think of everyone that's like Stripes
  • 00:23:16
    prices Stripes that I've seen right and
  • 00:23:20
    we may if you if and I think we may be
  • 00:23:22
    playing spending too many mental
  • 00:23:24
    calories and playing too many games on
  • 00:23:25
    pricing when we should focus more on
  • 00:23:27
    closing customers in a low friction
  • 00:23:29
    manner right I think this is over
  • 00:23:31
    overdone this issue right I think the
  • 00:23:33
    deeper inside here which I I have
  • 00:23:35
    confirmed so many times through the last
  • 00:23:37
    years is just to not over index and
  • 00:23:39
    social media think right of what VC's
  • 00:23:41
    want or like what I want as a Founder
  • 00:23:44
    like just talk to your customers listen
  • 00:23:45
    to what they want help them deliver
  • 00:23:47
    awesome value and the rest will take
  • 00:23:48
    care of itself but it's a bit like you
  • 00:23:50
    know I think with all this like model
  • 00:23:52
    stuff it's that most customers don't
  • 00:23:53
    really care about models right they buy
  • 00:23:55
    products and I think in the tech world
  • 00:23:57
    we love talking about Technologies and
  • 00:23:58
    whatever it's like agents latest like
  • 00:24:00
    hype cycle but when the robber hits the
  • 00:24:02
    road eventually you can't just live on
  • 00:24:04
    like
  • 00:24:05
    bus Grand ideas yeah that that's always
  • 00:24:09
    that's a big lesson for me the last
  • 00:24:10
    eight years for sure okay I want to
  • 00:24:12
    spend just a couple minutes on overall
  • 00:24:14
    VC fundraising because I have both of
  • 00:24:15
    you right Victor's been at this since
  • 00:24:18
    2017 before the buzzwords were cool just
  • 00:24:21
    raised what I think it's a real round
  • 00:24:23
    right 2.3 billion or whatever was that
  • 00:24:25
    sounds like a lot but you're not a 2.3
  • 00:24:27
    million in are right you you have't
  • 00:24:30
    disclosed I won't ask you but you're on
  • 00:24:31
    one side of the other of nine figures in
  • 00:24:33
    Revenue like I think they got a good
  • 00:24:35
    deal great deal great seriously so to I
  • 00:24:38
    stole this chart from eny which stole
  • 00:24:40
    half their data from crunch base which
  • 00:24:42
    probably got it all out of uh deep seek
  • 00:24:45
    and and then uh open AI anyway so we're
  • 00:24:47
    all stealing each other's data but this
  • 00:24:49
    isn't mine but I like this version of
  • 00:24:51
    what we've seen before even though which
  • 00:24:54
    is what happened at the end of last year
  • 00:24:57
    is money went up right as we can see on
  • 00:24:59
    the left which we all saw at least on
  • 00:25:01
    Tech runch right but deals
  • 00:25:03
    didn't deals went down right A lot of it
  • 00:25:06
    was growth focused and half of it was AI
  • 00:25:09
    focused but this says to
  • 00:25:12
    me if you're in The Hot Zone of
  • 00:25:16
    Synthesia and we Victor in a minute you
  • 00:25:18
    can tell us what hot is it's the good
  • 00:25:20
    it's as good as almost as good as 2021
  • 00:25:23
    right maybe it will be better I think
  • 00:25:24
    this year but overall deals are down is
  • 00:25:27
    this what you're seeing and so does this
  • 00:25:28
    mean we're in a real world where the
  • 00:25:29
    folks that are saying fundraising is
  • 00:25:30
    harder are right and the folks that are
  • 00:25:32
    saying it's froth are right is everybody
  • 00:25:35
    right I think everybody's right you look
  • 00:25:38
    at data bricks raising 15 yeah that's as
  • 00:25:41
    big as the Facebook IPO was which was
  • 00:25:44
    the largest technology IPO at the time
  • 00:25:46
    in 2014 or 2015 whenever it was and they
  • 00:25:48
    did that in the private markets right I
  • 00:25:51
    raising five billion so there you have
  • 00:25:53
    20 billion let's say of that 60 billion
  • 00:25:55
    in Q4 going into two companies a third
  • 00:25:58
    of the dollars and so it's really like
  • 00:26:00
    what what you're seeing here is the
  • 00:26:02
    early stage deal counts I think probably
  • 00:26:05
    the same the growth stage deal counts
  • 00:26:06
    are down in a pretty meaningful way
  • 00:26:09
    after 2022 the series B through the D
  • 00:26:11
    Market was effectively dead and it
  • 00:26:14
    starting to come back if your if your
  • 00:26:17
    check sizes are measured in the billions
  • 00:26:18
    now to lead to financing you can't
  • 00:26:21
    invest in many companies so
  • 00:26:24
    for and what's your
  • 00:26:27
    pulse for your a and even this is top 1%
  • 00:26:31
    or even higher but what's your pulse for
  • 00:26:34
    the classic B2B triple triple double
  • 00:26:37
    company which was what everyone wanted
  • 00:26:40
    to fund in 2021 what's your pulse on
  • 00:26:41
    that in 2025 I think it's a it's like a
  • 00:26:45
    lower multiple a I think it's like a
  • 00:26:48
    very reasonably priced investment but
  • 00:26:51
    the deal gets done but they all get done
  • 00:26:52
    all the triple double doubles get done
  • 00:26:54
    no I don't not even maybe on the margin
  • 00:26:57
    depending on the valuation and depending
  • 00:26:59
    on the market I think the hard part is
  • 00:27:00
    those triple doubles in classic software
  • 00:27:03
    there's a lot more competition than
  • 00:27:04
    there once was now all of a sudden you
  • 00:27:05
    look at a company in a particular space
  • 00:27:07
    and the last 10 years we you and I have
  • 00:27:09
    funded a lot of those companies and many
  • 00:27:11
    of our friends have funded those
  • 00:27:12
    companies and so if you're tripling
  • 00:27:15
    doubling without AI It's A Hard Sell and
  • 00:27:19
    then that's the thing with AI like you
  • 00:27:20
    used to have two competitors now you
  • 00:27:21
    have
  • 00:27:22
    400 I can't even keep track like I can't
  • 00:27:25
    even meet with a startup that's free
  • 00:27:27
    seed whatever because there four how
  • 00:27:29
    many other AI legal in-house litigators
  • 00:27:33
    could there be like this was the most
  • 00:27:35
    boring category when I invested legal in
  • 00:27:37
    2015 now there's a thousand startups
  • 00:27:40
    right ER for to be remember that era
  • 00:27:43
    Groupon Groupon came out and then there
  • 00:27:45
    were a thousand Chinese competitors
  • 00:27:46
    within six weeks yeah a few hundred in
  • 00:27:49
    the US it's the Groupon era for AI where
  • 00:27:53
    one company takes off and then there are
  • 00:27:54
    15 to 20 others because you're wrapping
  • 00:27:57
    the API
  • 00:27:58
    you have a lot of capability out of the
  • 00:28:00
    gate and then long-term product
  • 00:28:02
    differentiation probably matters less
  • 00:28:04
    than go to market
  • 00:28:05
    execution yeah the products are
  • 00:28:07
    superficially great now right versus
  • 00:28:09
    they used to all our products in the old
  • 00:28:11
    days the first 18 months were terrible
  • 00:28:13
    like they had one cool thing they did
  • 00:28:15
    one cool thing everything else was
  • 00:28:17
    terrible now these products at least
  • 00:28:18
    superficially like you you talk about AI
  • 00:28:21
    agents how you know people can't get Roi
  • 00:28:23
    from them but superficially all the the
  • 00:28:25
    AIS SDR superficially all the products
  • 00:28:27
    are great like in a quick driveby every
  • 00:28:29
    aiso product's great isn't it the Demo's
  • 00:28:32
    amazing Demo's amazing was amazing was
  • 00:28:36
    amazing all right Victor what did you on
  • 00:28:38
    this slide was this the easiest round to
  • 00:28:40
    race it wasn't the easiest round to race
  • 00:28:42
    I think at the V that we're in now yeah
  • 00:28:46
    it's it's and especially when you have a
  • 00:28:49
    really good business which I think we do
  • 00:28:51
    not just from amazing Vision great star
  • 00:28:53
    but actually numbers customers spend
  • 00:28:56
    their contracts like the more data but
  • 00:28:58
    you have the more you will get Valu in
  • 00:28:59
    those things right I think there's
  • 00:29:01
    there's these like very outlier deals
  • 00:29:04
    where you're like de in AI or something
  • 00:29:05
    and you're like he six Engineers is
  • 00:29:07
    absolutely insane and they're going to
  • 00:29:08
    change the entire world coding like
  • 00:29:09
    boom2 billion valuation you can do that
  • 00:29:12
    maybe if you have a really strong Story
  • 00:29:13
    and there's like a team and whatever but
  • 00:29:15
    I think those are still pretty outlier
  • 00:29:16
    my sense from like also speaking to
  • 00:29:18
    peers is that when you're raising Pro
  • 00:29:20
    rounds now and you know without
  • 00:29:22
    Foundation model compan we're very
  • 00:29:23
    applied it we definitely evaluated like
  • 00:29:25
    a SAS company definitely with some
  • 00:29:27
    premium because I think it's a fair
  • 00:29:28
    price I think it's a good investment I
  • 00:29:30
    think the risk reward in this route is
  • 00:29:31
    the best it's ever been I think we raise
  • 00:29:33
    the series C that was coming out a lot
  • 00:29:36
    of like hype cycle chat gbt had just
  • 00:29:38
    dropped everyone was like freaking out
  • 00:29:40
    and we were growing super fast and we
  • 00:29:42
    were closing a lot of contracts but it
  • 00:29:45
    was still to be determined like can we
  • 00:29:46
    get to million dollar contracts will
  • 00:29:48
    people actually get consistent value out
  • 00:29:50
    of this are they just trying it out is
  • 00:29:51
    it Innovation budgets all of those
  • 00:29:53
    questions were still open but we raised
  • 00:29:55
    this round what I think stood out to the
  • 00:29:57
    market here was that not only are we
  • 00:30:00
    good at closing new logos and closing
  • 00:30:01
    new business but we're very good at the
  • 00:30:03
    renewals and the expansion right and
  • 00:30:05
    ultimately that's where your customers
  • 00:30:07
    vote if your product is just a cool demo
  • 00:30:10
    of which there's a lot right now or is
  • 00:30:12
    it actually something that deres real
  • 00:30:13
    business value and my sense from from
  • 00:30:15
    from spending time with a lot of
  • 00:30:16
    investors right is that a lot of there's
  • 00:30:18
    a bit of fatigue very great grw really
  • 00:30:20
    quickly you close a lot of contracts yes
  • 00:30:22
    but that does necessarily mean that
  • 00:30:23
    translates into lasting value there's
  • 00:30:25
    this I think there's this weird thing in
  • 00:30:27
    AI right now
  • 00:30:28
    it's very fear-driven and people are
  • 00:30:30
    very willing to Pat ways with their
  • 00:30:32
    money both on consumer side but also
  • 00:30:34
    businesses to try out new things right
  • 00:30:35
    because everyone is scared that their
  • 00:30:37
    business is going to completely
  • 00:30:39
    disappear because open their ey is going
  • 00:30:41
    to kill them next month they're afraid
  • 00:30:43
    to lose their job if they don't have an
  • 00:30:44
    AI strategy if they don't appear to be
  • 00:30:46
    an AI expert reality I think most
  • 00:30:48
    Enterprise buyers don't really know what
  • 00:30:50
    they need they don't really understand
  • 00:30:51
    the technologies that they tried lots of
  • 00:30:53
    different things which maybe a sound
  • 00:30:54
    tactic right but it means that there's a
  • 00:30:56
    lot of this demo budget that's Wasing
  • 00:30:58
    around and I think that's one of the
  • 00:31:00
    things that stood out to investors in
  • 00:31:02
    this round that's interesting how
  • 00:31:04
    important durable Revenue was for the
  • 00:31:06
    round right that it's
  • 00:31:07
    durable is that toas that's another
  • 00:31:09
    thing on the VC social media are are
  • 00:31:13
    folks still
  • 00:31:14
    anxious that so much of this revenue is
  • 00:31:17
    discretionary that it's experiment that
  • 00:31:18
    it won't stick that it's non-durable to
  • 00:31:20
    Victor's point or is that fading is a
  • 00:31:22
    concern I still think it's a concern
  • 00:31:24
    it's probably not as much of a concern
  • 00:31:25
    as it was last year last year by fire
  • 00:31:28
    preferences are changing so fat
  • 00:31:29
    underlying architectures were changing
  • 00:31:31
    the models were changing in a really
  • 00:31:32
    meaningful way are we doing search are
  • 00:31:35
    we doing rag different kinds of
  • 00:31:37
    databases we're trying to figure it out
  • 00:31:39
    and as preferences changed you had whole
  • 00:31:41
    categories come and go within the span
  • 00:31:43
    of a year or 18 months I don't think
  • 00:31:46
    we're quite there yet I think where we
  • 00:31:48
    are particularly at the application
  • 00:31:49
    layer is trying to figure out which AI
  • 00:31:51
    agents actually produce value like yeah
  • 00:31:54
    how much am I willing to pay for a
  • 00:31:55
    notaker right Leslie's notaker just Joy
  • 00:31:58
    the waiting room I wonder how much
  • 00:31:59
    paying for it's like five bucks 10 bucks
  • 00:32:01
    how many do we need 100 no tick her a
  • 00:32:03
    friend of mine tweeted he said this is
  • 00:32:05
    he attended a meeting this week where
  • 00:32:06
    there were more robots than there were
  • 00:32:08
    humans and one didn't attend and just
  • 00:32:10
    sent their notaker and I don't know if
  • 00:32:12
    you've done this but like sometimes I
  • 00:32:13
    want to
  • 00:32:14
    listen look at this call look at this
  • 00:32:16
    call there's at least 20 30 note takers
  • 00:32:19
    if you just tries to them but they keep
  • 00:32:21
    coming yeah and then you so you you can
  • 00:32:24
    attend a meeting virtually whatever
  • 00:32:26
    bionically so there yeah so I think the
  • 00:32:29
    question is okay how much are you
  • 00:32:30
    willing to pay for that will Zoom bundle
  • 00:32:31
    it in will Google meet bundle it in and
  • 00:32:33
    so is there price erosion there and so
  • 00:32:35
    maybe that kind of creates turn but I
  • 00:32:37
    don't think there's this like category
  • 00:32:39
    cataclysm that maybe we saw in 24 yeah
  • 00:32:43
    let me tie this to the last point I want
  • 00:32:45
    to give on but this ties to a meta
  • 00:32:47
    question I have so one of the questions
  • 00:32:48
    last year was is this all
  • 00:32:49
    experimentation budget will it fade away
  • 00:32:51
    is it durable to Victor's point right
  • 00:32:53
    Victor had to prove that he has triple
  • 00:32:55
    digit nrr and all the things eventually
  • 00:32:58
    all these companies have to converge to
  • 00:33:00
    real software companies probably IPO and
  • 00:33:02
    everything then where I'm really getting
  • 00:33:04
    confused is on agents I'm not confused
  • 00:33:07
    about the hype and what they can do and
  • 00:33:11
    whether our agents can all talk to our
  • 00:33:12
    agents and our agents that's going to be
  • 00:33:14
    I think it's exciting right I'm just
  • 00:33:16
    wondering how many of these we can pay
  • 00:33:17
    for how many of these can and I just
  • 00:33:19
    looked this morning J mes from HubSpot
  • 00:33:22
    has a little a AI agent he's got a bunch
  • 00:33:25
    of things on his network but he's got a
  • 00:33:26
    little Marketplace which I'm sure in six
  • 00:33:28
    months will be an order of magnitude
  • 00:33:29
    larger right he's got 629 agents right
  • 00:33:32
    they sales Source charging two bucks but
  • 00:33:34
    this was there was one on the bottom
  • 00:33:35
    right is the latest data from zyo and
  • 00:33:38
    they're all fairly consistent which is
  • 00:33:40
    basically over the last year people
  • 00:33:42
    didn't add any net apps they didn't add
  • 00:33:44
    any net apps to their stack right
  • 00:33:46
    whether that means they're actually
  • 00:33:47
    adding some but consolidating on their
  • 00:33:49
    core apps right I think I should have
  • 00:33:53
    known this but Victor I think sasia is a
  • 00:33:55
    swap right it's going to be a swap in
  • 00:33:57
    some cases right you're going to swap
  • 00:33:59
    it's not a total always a new app you
  • 00:34:00
    may be swapping out a training app you
  • 00:34:02
    may be swapping out part of the Adobe
  • 00:34:04
    Empire as you go right it may not but
  • 00:34:07
    how many can we I believe that the all
  • 00:34:09
    the energy in Venture in startups in YC
  • 00:34:12
    in SF is somehow based on this idea
  • 00:34:16
    we're going to pay for three to four
  • 00:34:17
    times more agents and so we're going to
  • 00:34:20
    like triple or quadruple our appstack
  • 00:34:22
    can we is it true can and can we process
  • 00:34:24
    this many paid agents or is this all a
  • 00:34:26
    big VC hype cycle that's going to
  • 00:34:28
    implode because we can't pay for 500 of
  • 00:34:30
    these agents what do you think first
  • 00:34:32
    toas is this does my question make sense
  • 00:34:34
    is where I'm confused so I think I think
  • 00:34:37
    we see we're seeing consolidation across
  • 00:34:39
    the SAS stack where you just have
  • 00:34:41
    bundling everywhere like you look at the
  • 00:34:42
    sales stack massive consolidation across
  • 00:34:45
    yes it's automation so I I I think
  • 00:34:47
    you'll see a reduction there I think
  • 00:34:49
    people will end up buying agents for
  • 00:34:52
    different reasons than they bought
  • 00:34:53
    classic software for example we're
  • 00:34:57
    investor if you're a large Enterprise
  • 00:34:58
    you have 641 apps you probably have some
  • 00:35:01
    pretty significant security apparatus
  • 00:35:03
    you have another 75 security products
  • 00:35:05
    those produce 8,000 alerts less than 1%
  • 00:35:07
    of those alerts are reviewed by humans
  • 00:35:09
    now law that you have to review all of
  • 00:35:12
    them people aren't going to buy
  • 00:35:14
    additional five other pieces of software
  • 00:35:16
    they'll probably buy a 100 agents and
  • 00:35:18
    those agents will do the job of the
  • 00:35:20
    humans they're going to pay for each
  • 00:35:21
    agent 10 or 20 bucks a month and they're
  • 00:35:23
    going to manage them individually the IT
  • 00:35:25
    team orever is going to have a whole
  • 00:35:27
    another set of 300 agents they have to
  • 00:35:29
    manage yeah I think I'm not saying
  • 00:35:32
    you're wrong I'm just trying to learn
  • 00:35:33
    actually I think that's you think about
  • 00:35:35
    like bdr of the future right bdr today
  • 00:35:37
    manages one Gmail inbox with Max 40
  • 00:35:39
    outbound email maybe they manage two or
  • 00:35:41
    three bdr of the future will be managing
  • 00:35:42
    40 or 50 Gmail inboxes and I think if
  • 00:35:46
    you think about there's a deflation in
  • 00:35:48
    the number of seats because now all of a
  • 00:35:49
    sudden bdr acceleration I don't need
  • 00:35:52
    nearly as many bdrs but the value per
  • 00:35:55
    seat goes up in a pretty meaningful way
  • 00:35:57
    and so I think maybe what we see is
  • 00:35:59
    application count stays flat or maybe
  • 00:36:01
    even decreases but the total amount of
  • 00:36:04
    software spending increases because
  • 00:36:05
    capturing some fraction that labor spend
  • 00:36:08
    but are so that that's a I I get that
  • 00:36:10
    and I can't predict the future but
  • 00:36:11
    aren't if you just think about it from a
  • 00:36:13
    oldfashioned Enterprise got to manage X
  • 00:36:16
    vendors gotta man Rogue renewals Rogue
  • 00:36:20
    increases do I really want an explosion
  • 00:36:23
    of five times as many app whether you
  • 00:36:25
    call them agents they're still apps from
  • 00:36:27
    a procure them an IT process aren't they
  • 00:36:29
    yeah still another goddamn thing I got
  • 00:36:31
    to evaluate do a security audit do a
  • 00:36:34
    cost benefit analysis make sure it
  • 00:36:35
    doesn't go Rogue like everyone went
  • 00:36:37
    rogue in 2021 and bought 50 sales apps I
  • 00:36:40
    is I just feel that the cognitive load
  • 00:36:42
    is high the C not for the end user but
  • 00:36:45
    for I for the buyer yeah I I think the
  • 00:36:47
    successful software applications like
  • 00:36:49
    whatever as a salesperson I might have
  • 00:36:51
    500 agents Each of which is managing one
  • 00:36:55
    lead yeah you could but and so what is
  • 00:36:58
    that agent doing it's Gathering the news
  • 00:36:59
    and the social media and the podcast and
  • 00:37:01
    transcribing it and that's pretty easy
  • 00:37:03
    to review I just go through my list
  • 00:37:05
    what's the latest with data breaks
  • 00:37:06
    what's the latest with Snowflake and
  • 00:37:07
    each one of those is an agent so I don't
  • 00:37:09
    think the cognitive overhead is huge if
  • 00:37:11
    the ux is correct but who pays for it
  • 00:37:13
    that I don't think for the I think the
  • 00:37:15
    bionic I call it the mech AE but the
  • 00:37:17
    bionic SDR I believe is happening fast
  • 00:37:20
    right but who's I've never found a
  • 00:37:22
    single AE on all of planet Earth or even
  • 00:37:25
    our galaxy who will put anything on
  • 00:37:26
    their own credit card never found an AE
  • 00:37:28
    so the company has to pay I've never
  • 00:37:30
    found a single sales rep may maybe in MC
  • 00:37:33
    Murdoch station that will pay for their
  • 00:37:34
    own sales but they won't they got some
  • 00:37:36
    so who's gonna pay for their ad agents
  • 00:37:38
    if they're discreet if they're not
  • 00:37:39
    bundled into another app right if
  • 00:37:41
    they're not part of a platform who's
  • 00:37:42
    gonna pay I I think it comes from the
  • 00:37:44
    sales you seen the sales budget I don't
  • 00:37:46
    know 10 years ago what was the sales
  • 00:37:47
    budget parade maybe 250 300 bucks today
  • 00:37:51
    exactly I think yeah and and then 2022
  • 00:37:54
    what do you think it was 750 800 bucks
  • 00:37:56
    Maybe a, yeah it doubled yeah and so
  • 00:37:59
    Salesforce is half of that with Einstein
  • 00:38:01
    right it's 175 base sales heat 500 to
  • 00:38:04
    get with Einstein and so I think you see
  • 00:38:08
    some consolidation compression within
  • 00:38:11
    the existing sales stack and some of
  • 00:38:12
    that budget shifts AI yeah I don't know
  • 00:38:14
    I want to tr I I you think about this
  • 00:38:17
    much more that this is just a meta worry
  • 00:38:19
    for me is that no one's G Enterprises
  • 00:38:23
    are going to be fatigued reviewing 200
  • 00:38:26
    paid agents not if they're part of my
  • 00:38:27
    stack not if they're part of synth Asia
  • 00:38:29
    for for Content creation not if they're
  • 00:38:31
    part of HubSpot but 200 new vendors is
  • 00:38:34
    exhausting what my limited experience
  • 00:38:36
    both with my portfolio as a CEO and my
  • 00:38:39
    brief experience in a big Enterprise is
  • 00:38:41
    there's there's just there's only so
  • 00:38:43
    many vendors you can bring in and there
  • 00:38:44
    are moments of time like 2020 when the
  • 00:38:46
    world ended and there are moments in
  • 00:38:49
    time and there are always
  • 00:38:50
    experimentation budgets if it's isolated
  • 00:38:52
    if it's isolated from risk but when
  • 00:38:53
    there's risk I just worry there's that
  • 00:38:57
    that
  • 00:38:58
    we're not thinking through how we're
  • 00:38:59
    going to get these through the cio's
  • 00:39:02
    budget right the cio's budget and I also
  • 00:39:05
    don't think SBS want to pay for 100
  • 00:39:07
    agents on their own I see snbs being
  • 00:39:09
    cheaper than ever today they're cheaper
  • 00:39:11
    than ever right they the SBS do not want
  • 00:39:14
    200 recurring things on their personal
  • 00:39:15
    credit card but why do we I don't
  • 00:39:18
    understand this agents talk here because
  • 00:39:19
    would clarify is this like just features
  • 00:39:21
    that we're now calling agents or is it
  • 00:39:23
    platforms or I don't think that we'll
  • 00:39:25
    think of we have two agents like
  • 00:39:28
    researching social media on behalf of
  • 00:39:29
    our customers like you'll have a you buy
  • 00:39:31
    a product that does that for maybe it's
  • 00:39:33
    part of your existing stack well maybe
  • 00:39:34
    that's the question I'm asking you too I
  • 00:39:36
    think if I'm on if I'm on agk made up
  • 00:39:40
    yeah I'm there's 629 I'm gonna find I'm
  • 00:39:42
    gonna deploy a 100 of them I don't know
  • 00:39:44
    I just think it's such a weird it's a
  • 00:39:46
    weird way to talk about software and
  • 00:39:48
    features like I understand it's like
  • 00:39:49
    it's you're at the cutting a you're
  • 00:39:51
    supposed to tell us this is what's
  • 00:39:52
    happening the funny d
  • 00:39:58
    I got here by not listening to the bus
  • 00:40:00
    word was just like great product people
  • 00:40:03
    I'm just trying it just RS me out all
  • 00:40:05
    the stuff it's like it's features man
  • 00:40:07
    it's like it's an llm driven feature
  • 00:40:09
    that does something for you why are we
  • 00:40:11
    like trying to an theorize it to be like
  • 00:40:12
    an agent like I understand like it's
  • 00:40:14
    sell to the marketplace and it's like
  • 00:40:15
    it's cool and in some maybe I can
  • 00:40:17
    understand why I'm called an agent I
  • 00:40:19
    think majority of this is just weird I
  • 00:40:21
    think 5 years time like I think everyone
  • 00:40:23
    like why the did we spend 5 years
  • 00:40:25
    talking about algorithms s be like like
  • 00:40:27
    a coworker it's just like software the
  • 00:40:28
    dusting for us and sometimes maybe you
  • 00:40:31
    want to chat with it that's great that's
  • 00:40:32
    a new interface but I I really don't
  • 00:40:34
    understand this like obsessed with
  • 00:40:36
    agents and that said you know we're
  • 00:40:37
    probably launching an air agent strategy
  • 00:40:38
    soon because everybody likes to talk
  • 00:40:40
    about this stuff but nobody want to talk
  • 00:40:42
    about features so I get it from go to
  • 00:40:43
    marketed sales perspective but it
  • 00:40:45
    strikes me as well but it'll be part of
  • 00:40:46
    your platform you have thousands of
  • 00:40:48
    customers so don't that's like Weir
  • 00:40:52
    really no no I don't have the answers
  • 00:40:54
    I'm challenging it second half of 2020
  • 00:40:56
    to 202 and somehow we magically found
  • 00:40:58
    budget for all these new apps and the
  • 00:41:00
    pandemic did it right and if you saw and
  • 00:41:03
    the craziest one in my portfolio was
  • 00:41:04
    talk Des which is call
  • 00:41:06
    center and forget about how AI changed
  • 00:41:09
    in 2020 you couldn't go to the office
  • 00:41:10
    for the call center right so everyone in
  • 00:41:12
    their space in like they they grew five
  • 00:41:15
    EX in six months
  • 00:41:17
    because it changed the rules right the
  • 00:41:20
    it changed the rules and I think there
  • 00:41:22
    are some rule changing happening but I
  • 00:41:24
    don't think in 2026 it can double I just
  • 00:41:27
    think an agent counts as an app
  • 00:41:29
    potentially if it's a vendor let me put
  • 00:41:31
    it maybe I'm I'm rambling too much every
  • 00:41:33
    new vendor in the Enterprises cognitive
  • 00:41:36
    load a new vendor to put into the
  • 00:41:38
    procurement system to get approved to
  • 00:41:40
    track and every Enterprise knows that if
  • 00:41:43
    I bring in one it's not just one year
  • 00:41:45
    for synth Asia now I'm stuck for eight
  • 00:41:47
    years that's how Enterprises think they
  • 00:41:49
    don't think one year they're like okay
  • 00:41:51
    it's going to take me the first year I'm
  • 00:41:52
    learning the second year I'm rolling it
  • 00:41:54
    out more the third year I'm going to buy
  • 00:41:55
    more stuff and they always like we at
  • 00:41:57
    back at Adobe like when you put any
  • 00:41:59
    vendor in the system you doubled the
  • 00:42:00
    budget because you knew it was going to
  • 00:42:01
    grow with add-ons and seats and it was
  • 00:42:03
    and so how many of those can you add
  • 00:42:05
    each year even at Adobe size you just
  • 00:42:07
    run out of vendors you can and I don't
  • 00:42:09
    know if AI change if agents change that
  • 00:42:11
    rule they're GNA change a lot of other
  • 00:42:12
    rules and so I don't know who's gonna
  • 00:42:14
    pay for that I just don't know who's
  • 00:42:15
    gonna pay for the discreet vendors not
  • 00:42:17
    the agents themselves but so many
  • 00:42:18
    vendors so many vendors I do think there
  • 00:42:21
    like there are fundamentally new product
  • 00:42:23
    categories though I would put a want
  • 00:42:24
    like we never going to replace someone
  • 00:42:26
    or take budget from so we probably will
  • 00:42:27
    over time don't come from the training
  • 00:42:29
    budget or the LMS budget if you're for
  • 00:42:31
    an internal use case no we'll we'll
  • 00:42:33
    sometimes take some of the video
  • 00:42:34
    production budget that they'll go that
  • 00:42:36
    with agencies yeah and the key point is
  • 00:42:40
    that people not evaluting us against
  • 00:42:41
    should we switch out this thing with
  • 00:42:42
    some easier right is a n new fake it's
  • 00:42:44
    also why we're growing so fast I think
  • 00:42:46
    there are new product categories like
  • 00:42:48
    that and then there'll be a lot of these
  • 00:42:49
    agents or features or whatever that'll
  • 00:42:51
    exist better within the platforms people
  • 00:42:54
    already use and they'll be add-ons and
  • 00:42:57
    the modules I believe you that you're
  • 00:42:58
    not directly stealing someone's creaky
  • 00:43:01
    old budget but what I my limited
  • 00:43:03
    experience tells me behind the scenes
  • 00:43:05
    you're stealing someone else's budget
  • 00:43:07
    they're cancelling someone that they
  • 00:43:09
    don't quite need that was on the bubble
  • 00:43:11
    and they this is I don't think this has
  • 00:43:12
    changed since I worked at a Fortune 500
  • 00:43:14
    tech company you get together each year
  • 00:43:16
    this is why budget's annually and toas
  • 00:43:18
    you want to add that cool product what
  • 00:43:20
    are you gonna cut this is why so many
  • 00:43:22
    things got cut last maybe some of the
  • 00:43:24
    budget maybe some of the budget does not
  • 00:43:25
    come from sulfur right it com from like
  • 00:43:27
    Services based so far okay a lot of the
  • 00:43:29
    budget we e into is like video
  • 00:43:30
    production agencies right that do Sol
  • 00:43:32
    for our vendors or translation budgets
  • 00:43:35
    or like more service driven things and I
  • 00:43:36
    think especially we talk about agent
  • 00:43:38
    that's that'll probably be a lot of what
  • 00:43:40
    that can Le into so you may add another
  • 00:43:41
    software wend but maybe the budget
  • 00:43:43
    doesn't come other software platform but
  • 00:43:45
    it comes from like service that's a lot
  • 00:43:46
    of the Hope on X is that we can turn
  • 00:43:48
    humans into a software budget Mor a stud
  • 00:43:51
    in Q3 I think and they said % of the AI
  • 00:43:53
    budget was uh cannibalizing existing
  • 00:43:56
    software and 50% % was new 5050 that's
  • 00:43:59
    probably right they got the data yeah
  • 00:44:02
    and so I bet Victor is right like it's
  • 00:44:04
    but translation or other kinds of
  • 00:44:06
    services Consulting customization legal
  • 00:44:09
    like all that stuff like legal
  • 00:44:11
    Outsourcing that's where the software
  • 00:44:12
    budget's coming from it it's eating away
  • 00:44:14
    the labor cost yeah so that's so the
  • 00:44:17
    answer like a lot of these things is yes
  • 00:44:19
    is it's both uh we are and where're and
  • 00:44:22
    Tomas just maybe to wrap it up if so 50%
  • 00:44:25
    is cannibalization maybe Victor doesn't
  • 00:44:26
    see it but they're cutting videographer
  • 00:44:28
    they're cutting something else the 50%
  • 00:44:30
    that's net new budget where's like the
  • 00:44:32
    fort like the GDP is still somewhat
  • 00:44:34
    fixed in the United States what's it
  • 00:44:35
    growing 3% a year even under our new
  • 00:44:37
    regime yeah so for real where's the
  • 00:44:40
    money coming from to to bet where do
  • 00:44:42
    they get where does Morgan Stanley or
  • 00:44:44
    Theory ventur say where are we getting
  • 00:44:45
    this money and how long will it last I
  • 00:44:47
    think it's labor I'm seeing it in our
  • 00:44:48
    portfolio right like I look at R&D spend
  • 00:44:51
    and that like year-over-year R&D growth
  • 00:44:54
    compared to 10 years ago we have one
  • 00:44:55
    company growing really fast as
  • 00:44:57
    engineering headcount youw will be
  • 00:44:58
    flapped yeah because they deploy some AI
  • 00:45:02
    enabled ID and they can Rite a lot more
  • 00:45:03
    software and so instead of hiring
  • 00:45:05
    another say 10 softare engineers at 250
  • 00:45:08
    300K a year it's been milon and spend
  • 00:45:11
    they're starting to spend I don't know
  • 00:45:13
    $100,000 a year on AI software do you
  • 00:45:16
    think that's true from what Victor think
  • 00:45:18
    we can do more with fewer Engineers with
  • 00:45:20
    AI and the new world or is that a bit of
  • 00:45:24
    a bit of a
  • 00:45:25
    myth I think we can but I think I think
  • 00:45:28
    there'll be many different shapes of
  • 00:45:30
    companies I don't think there'll be one
  • 00:45:31
    type of company where you really want to
  • 00:45:32
    go for it I still think if you want to
  • 00:45:35
    build A1 billion doll company you're
  • 00:45:37
    going to need a lot of sales people
  • 00:45:38
    you're going to need a lot of Engineers
  • 00:45:40
    but I think there's no doubt that
  • 00:45:41
    productivity is going through the roof
  • 00:45:43
    it feels like we are seeing a lot of
  • 00:45:44
    these types of companies as as Thomas is
  • 00:45:46
    saying you can and also it feel like the
  • 00:45:48
    mindset has changed a little bit where
  • 00:45:49
    there's a l a reluctance to like hiring
  • 00:45:51
    more people and like just be more
  • 00:45:53
    efficient which I think is super
  • 00:45:55
    awesome but yeah I I think it's never
  • 00:45:58
    black and white right but I think that's
  • 00:45:59
    a definitely new type of company which I
  • 00:46:01
    think is super cool but if you want to
  • 00:46:03
    tackle the Enterprise like we are that's
  • 00:46:06
    still very difficult to do without
  • 00:46:07
    adding significant account yeah you
  • 00:46:08
    might even if we were on VC Twitter the
  • 00:46:11
    way to take you to the extreme is the
  • 00:46:12
    new vity metric is ARR per employee per
  • 00:46:15
    engineer is more interesting than AR per
  • 00:46:17
    employee I can tell you why it's more
  • 00:46:18
    interest per engineer per developer and
  • 00:46:20
    and that's like the sexiest metric I
  • 00:46:22
    think it's one of the coolest we're 10
  • 00:46:24
    people and we're generating 20 million
  • 00:46:25
    in re like those kinds of ratios have
  • 00:46:28
    seen yeah a decade ago and now people
  • 00:46:31
    are striving for and we hear from
  • 00:46:33
    Founders like we we only have this many
  • 00:46:35
    people we don't want to hire people
  • 00:46:36
    we're raising Capital but we're not
  • 00:46:37
    hiring that's unusual that's a very
  • 00:46:40
    different pitch do you think I'm making
  • 00:46:42
    this number up but I think it's about
  • 00:46:44
    right and Victor could probably for an
  • 00:46:46
    case study then we maybe we can break
  • 00:46:48
    that makes that's what I see too Mike
  • 00:46:50
    and listen the world's changing so
  • 00:46:51
    quickly today even the whole deepsea
  • 00:46:53
    thing it feels like everything's faster
  • 00:46:55
    than it was a week ago but it feels to
  • 00:46:57
    me there's still the issue around 200
  • 00:46:58
    300 employees where you still end up
  • 00:47:02
    needing 10 people to do everything and I
  • 00:47:05
    what I feel like this 10 people can get
  • 00:47:06
    to 10 million err thing it just shifts
  • 00:47:08
    the time like instead of me needing but
  • 00:47:11
    it's still at two at at 100 million at
  • 00:47:14
    50 million eror I need so many humans
  • 00:47:16
    and honestly so many mediocre humans you
  • 00:47:18
    start having to hire okay maybe that's
  • 00:47:20
    the wrong term you start off to hiring a
  • 00:47:23
    minus tier and B tier no matter even
  • 00:47:24
    though you can't say it out loud and it
  • 00:47:25
    just takes so people to do some of the
  • 00:47:29
    basic stuff and you need so many sales
  • 00:47:31
    people and so many mediocre renewals
  • 00:47:33
    people and I just see everything still
  • 00:47:34
    gumming up today at that size maybe if
  • 00:47:37
    AI gets rid of that roadblock that'd be
  • 00:47:38
    awesome but I wonder if it Just sh kicks
  • 00:47:40
    the can down the road a year or two yeah
  • 00:47:42
    it's like Salesforce hiring a thousand
  • 00:47:44
    sales people to sell their SDR sales
  • 00:47:46
    people right the agents I think there's
  • 00:47:48
    still lag like I think we'll I think
  • 00:47:49
    it'll be def be more efficient but I
  • 00:47:52
    still think that we're far from you be
  • 00:47:54
    able
  • 00:47:55
    to yeah this would you C like true
  • 00:47:57
    enterprise software I do think that's a
  • 00:47:58
    bit different
  • 00:47:59
    than than like what or like SB but you
  • 00:48:02
    don't need the hu touch but I still
  • 00:48:04
    feels difficult to close it'll be very
  • 00:48:06
    interesting if a company with 20 20
  • 00:48:08
    employees Whatever Gets to close like
  • 00:48:10
    million dollar deals with Enterprise I
  • 00:48:12
    they that would be like a really
  • 00:48:13
    interesting kind of Tipping Point but
  • 00:48:15
    for so far for most of the ones that
  • 00:48:16
    I've seen at least is still a lot of
  • 00:48:19
    sell service credit card kind of smaller
  • 00:48:22
    Sal because it's still just very
  • 00:48:23
    difficult managing a dual desk right you
  • 00:48:25
    just need people for that maybe let me
  • 00:48:26
    just let me ask one last version of this
  • 00:48:27
    to you toas and we could wrap and this
  • 00:48:29
    was great so when we start when you and
  • 00:48:32
    I started toas a lot of VCS would yell
  • 00:48:34
    at Founders and say you've got to
  • 00:48:36
    Offshore to save money you got
  • 00:48:38
    everything told that the magic solution
  • 00:48:39
    when we started maybe you two Victor the
  • 00:48:41
    magic solution with VCS was have 50 or
  • 00:48:44
    60% of your India offshore that would
  • 00:48:45
    magically cut your costs right magically
  • 00:48:47
    because developers are fungible they're
  • 00:48:49
    all the same so just hire a bunch of
  • 00:48:51
    folks in some country you've never
  • 00:48:52
    personally visited and that's my VC
  • 00:48:55
    solution is the new version today the
  • 00:48:57
    version of today is if you're good at AI
  • 00:48:59
    you only need 10 developers it almost
  • 00:49:01
    feels like it's the new version of
  • 00:49:02
    offshore is AI is the new offshore like
  • 00:49:05
    you you got you why do you need more
  • 00:49:06
    than 10 guys
  • 00:49:09
    know that's a great that's a great
  • 00:49:11
    that's tomorrow's blog post for you I
  • 00:49:12
    hope to I think like some people did
  • 00:49:14
    offshoring really well they h a great
  • 00:49:15
    country manager they really spent the
  • 00:49:17
    time to transfer the culture yeah the
  • 00:49:20
    cost efficiencies and then some people
  • 00:49:22
    basically just mailed it in and you got
  • 00:49:25
    what you paid for I think it's the same
  • 00:49:27
    with AI if you change your recruitment
  • 00:49:28
    processes and you look for people who
  • 00:49:30
    really understand how to prompt these
  • 00:49:32
    systems really well and think about
  • 00:49:34
    building with AI I think you can drive a
  • 00:49:36
    ton of efficiency I think if your CEO
  • 00:49:39
    has been told by VC's hey go employ uh
  • 00:49:42
    AI within your engineering team and you
  • 00:49:43
    buy a bunch of people some subscriptions
  • 00:49:45
    yeah cursor is the
  • 00:49:47
    answer it's not you have to change the
  • 00:49:49
    way that you work and change the way the
  • 00:49:51
    teams working that's really
  • 00:49:53
    hard I I also really still think that
  • 00:49:56
    what I see across no matter like what
  • 00:49:58
    task we're talking about where AI really
  • 00:50:00
    shines is still augmenting humans right
  • 00:50:03
    in the case of us we augment people to
  • 00:50:05
    make make videos cursor works so well
  • 00:50:07
    because it puts the human in the driving
  • 00:50:09
    seat and just makes you more efficient I
  • 00:50:12
    think this like pip dream of agent like
  • 00:50:14
    just doing things complet to Aon in my
  • 00:50:16
    view is still still a while away yeah
  • 00:50:20
    we'll see we can break my limited
  • 00:50:21
    experience is that an an stier agent is
  • 00:50:25
    better than a c-tier employ
  • 00:50:28
    this is what across the contact center
  • 00:50:30
    is you usually see that the agents are
  • 00:50:32
    like in the they can't do everything but
  • 00:50:34
    they're in the top 10% of cessat because
  • 00:50:37
    what they do they get done without
  • 00:50:39
    complaining and they don't argue but
  • 00:50:41
    they can't be below I think the whole
  • 00:50:43
    world's going to that agents will
  • 00:50:45
    replace the C tier like very quickly and
  • 00:50:48
    as and deep in the B tier as they
  • 00:50:50
    can and I don't know all right guys this
  • 00:50:53
    was great we ran out of time Victor so
  • 00:50:55
    great to have you back we'll see you in
  • 00:50:56
    London again and Tomas we'll see you at
  • 00:50:58
    the next one my friend thank you thanks
  • 00:51:00
    thanks guys
Tags
  • AI
  • B2B
  • SaaS
  • Funding
  • AI Video
  • Durable Revenue
  • Corporate Training
  • Cost Reduction
  • Software Pricing
  • AI Agents