Skip to content
a16za16z

Databricks CEO: Stop Scaring People About AI

Databricks co-founder and CEO Ali Ghodsi joins a16z General Partners Martin Casado and Sarah Wang for a conversation about AI risk, recursive self-improvement, cybersecurity, and what’s actually holding back enterprise adoption. Ali argues that today’s models are already capable enough to automate far more work than most companies are using them for. The bigger problem is context: models haven’t been in every meeting, don’t understand how decisions actually get made, and lack the institutional knowledge that experienced employees accumulate over years. He explains why building an organizational “ontology” could help close that gap and what Databricks has learned from doing it internally. They also debate the current conversation around pacing frontier AI, what would constitute meaningful recursive self-improvement, and why Ali distinguishes speculative superintelligence risk from the much more immediate challenge of AI-powered cyberattacks. They close with how enterprises are managing exploding AI usage and costs, the shift toward multiple models and harnesses, and why agents are beginning to reshape infrastructure itself. Timestamps: 00:00 - Intro 00:48 - Pacing the Frontier: Where Ali Lands in the AI Debate 16:01 - The Black Box Test: Is RSI Actually Happening? 20:17 - Why We Haven't Seen the AI Cyber Apocalypse Yet 31:27 - The 4D Chess Problem: Doom Talk vs IPO Allocations 33:33 - Industry Self-Policing vs Federal Involvement 41:49 - The Enterprise Use Cases Surprising Even Ali 44:36 - How Enterprises Actually Operationalize AI in the Next 12 Months 45:24 - Defining Ontology (Beyond the Palantir Version) 50:55 - The Finance Anecdote: Real AI Value in the Boardroom Resources: Follow Ali Ghodsi on X: https://x.com/alighodsi Follow Sarah Wang on X: https://x.com/sarahdingwang Follow Martin Casado on X: https://x.com/martin_casado Stay Updated: If you enjoyed this episode, be sure to like, subscribe, and share with your friends! Find a16z on X: https://twitter.com/a16z Find a16z on LinkedIn: https://www.linkedin.com/company/a16z Listen to the a16z Show on Spotify: https://open.spotify.com/show/5bC65RDvs3oxnLyqqvkUYX Listen to the a16z Show on Apple Podcasts: https://podcasts.apple.com/us/podcast/a16z-podcast/id842818711 Follow our host: https://x.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see http://a16z.com/disclosures.

Ali GhodsiguestMartin CasadohostSarah Wanghost
Sep 18, 20261h 6mWatch on YouTube ↗

EVERY SPOKEN WORD

  1. 0:000:48

    Intro

    1. AG

      As a business leader, there's a tragedy of the commons. If you wanna stop, if you wanna go slower, why don't you go slower? Like, I'm competing, I wanna win.

    2. MC

      There's almost two camps. There's one camp which believes that this actually is an engineering problem, and there's others which actually believe you have to slow it down.

    3. AG

      Humans don't respond fast enough to the attacks that are happening. You need to automate all of those, and most organizations are actually not close to doing that. Is RSI and recursive self-improvement that the labs are doing leading us there? That's the big question.

    4. SW

      Something Elon said, "This is some elaborate 4D chess," because on the one hand you're saying all of humanity will die.

    5. AG

      Mm-hmm.

    6. SW

      On the other hand, you're saying, "Hey, what do you want for your IPO allocation?"

    7. MC

      For the first time ever, a company at scale last week said that they're moving from the frontier models to GLM. Do you think that that's a trend or do you think that's just like a one-off anecdote?

    8. AG

      Um...

  2. 0:4816:01

    Pacing the Frontier: Where Ali Lands in the AI Debate

    1. SW

      Thank you for being here, Ali.

    2. AG

      Super excited.

    3. SW

      So we obviously wanna get to Databricks, but there is a broader conversation going on right now about AI and Dario's weighed in, Jakob's weighed in, Elon's weighed in, but we wanna hear what Ali Ghodsi thinks. In terms of, you know, if you called the topic broadly speaking, pacing the frontier, et cetera, um, what is your strongest agreement with what's being out there? Where do you disagree? And maybe where is there nuance that's not being captured?

    4. AG

      Yeah. Ha- happy to cover it, and me and Martin argue a lot so, um-

    5. MC

      [laughs] So

    6. AG

      ... I'm sure that's not gonna take long.

    7. MC

      We'll try and re-

    8. AG

      Yeah

    9. MC

      ... we'll try and rein it in this time.

    10. AG

      Yeah. Try to stay calm. But, uh, well, I, I do think first and foremost that there, uh, maybe we agree on this, that, um, uh, leaders have responsibility to not freak people out unnecessarily unless there's really, really good reason. And I think, you know, there's always different people in society that are at different places, you know, in their mind space. So, you know, talking about these kind of existential risks and, you know, uh, scenarios where all of humanity is gonna be wiped out, I think, uh, um, is ir- irresponsible. Uh, like it can tip a lot of people over and it can cause a lot of-

    11. MC

      Uh, un- un-

    12. AG

      ... like mental health issues.

    13. MC

      Unless you have something that's gonna wipe people out.

    14. AG

      Yeah. As I said, yeah. If, if there is a actual reason for it, then, you know, that's a different story. But I think that right now the existential risk is close to zero. Um, so why freak everybody out? It's not actually needed. Uh, there are risks. We'll get into it. That's probably where we disagree. Um, but first and foremost, I think that leaders should not freak everyone out. And I mean, you know, if there's like technical nuances in how we're doing AI research and so on, well, researchers can discuss that.

    15. SW

      Yeah.

    16. AG

      You don't need to every time go on TV and, or blast on Twitter to millions of people that, "Hey, you know, I think there's like this percentage, 10% risk that all humanity's gonna be wiped out." I don't think that's like helpful for a lot of people. Actually, I think it causes a lot of harm for a lot of folks who, who get stressed out and actually are not in the nuances of all of this stuff and what it means. So that I don't think we should do. Uh, I don't think it's u- fruitful. It doesn't really help anyone.

    17. MC

      I, I mean, I think this is very, very true for the general public.

    18. AG

      Yeah.

    19. MC

      Like my sister-

    20. AG

      Mm-hmm

    21. MC

      ... who's great, who's a school teacher in rural Arizona, on Sunday texted me and [chuckles] said, and said, "Martin, should I prepare the cabin for..." You know, she's kind of a prepper anyways, "But should I prepare the cabin for the AI apocalypse? You know, I've got water set up, like when are you showing up?" I'm like, "Hold on." [laughs]

    22. AG

      Yeah.

    23. MC

      Like we're not there yet. So clearly this has kind of spilled over to the populace, which I agree-

    24. AG

      Yeah

    25. MC

      ... is unnecessary and has blowback. I think there's a second one which is, um, I don't know if you saw like walking in here, I was checking X and-

    26. AG

      Mm-hmm

    27. MC

      ... Elizabeth Warren just talked about, um, pausing all of AI development. That of course is on the coattails of Bernie, who is also working with Bannon, [chuckles] like-

    28. AG

      Mm-hmm

    29. MC

      ... Steve Bannon, like a-

    30. AG

      Yeah

  3. 16:0120:17

    The Black Box Test: Is RSI Actually Happening?

    1. MC

      So I'm, I'm gonna have like another like litmus test here, is like I actually love your four criteria. I was like literally just waiting to argue with it, but I, I actually think it's very [chuckles] like this is very good.

    2. SW

      [laughs]

    3. MC

      Uh, so I'm gonna... Let me, let me give you like, um, like a black box one. When, like when you're dealing with these dy-dynamic adaptive systems, like what are you gonna believe? Are you gonna believe like the numbers or your lying eyes, right? So I think you kind of have to go to the numbers on these ones. So like what are the numbers to look at? I really think [chuckles] you should just basically... And maybe going public is the right way to do it. Like, like if, if these companies continue to grow, reduce the number of people and the number, amount of money that goes into them, um, then I would say something is, is definitely happening here. Like I do think that like you can actually black box this and take a look, but none of those indicate... Like they're hiring like crazy. They're-

    4. AG

      But that's not fair

    5. MC

      ... burning money like crazy.

    6. AG

      That's not fair, 'cause, you know, companies are not necessarily efficient, right? So like what if you have, like... I mean, OpenAI itself was doing like a million different activities. A very small-

    7. MC

      Well-

    8. AG

      ... team of like 10 people were doing LLMs, and the LLM stuff was useful. Twitter there was a lot of people, now it's much less people at Twitter-

    9. MC

      I just have another... I agree. Just another litmus test. We have kind of two litmus test. We have your litmus test, which I think is great, but then you would actually have to have a way to instrument it.

    10. AG

      Yeah.

    11. MC

      And then we should have the black box litmus test. Like, I mean, listen, if, if Anthropic in two weeks is, you know, 12 people and they continue to grow and they're putting out models at an increasing rate, I think we should probably take notice of that.

    12. AG

      That's sufficient criteria, but it's not necessary condition, right?

    13. MC

      Totally agree. Yeah.

    14. AG

      But I'm just saying that, you know, there could be that, you know, and really the right way to do this then to look at, okay, the pre-training and the post-training that's been doing that's really necessary.

    15. MC

      Yeah.

    16. AG

      'Cause they have so much resources that they might be doing a lot of other stuff they don't need to do.

    17. MC

      Fair enough.

    18. AG

      But they're doing it just, and they can just hire-

    19. MC

      Right

    20. AG

      ... the people 'cause they have infinite money and infinite s-

    21. MC

      Right.

    22. AG

      So really the people that are training the next model, is that team tiny, tiny, and it's actually getting reduced, and they're doing less and less work, and just AI is doing it-

    23. MC

      Yeah

    24. AG

      ... and the post-training, and then they're all just using less GPUs? That's not the case.

    25. MC

      The, the, we've, we've had this argument many times as an industry before. I remember when like, like we learned how to really cluster computers.

    26. AG

      Yeah.

    27. MC

      Because like the mainframe was actually kind of limited by things like memory coherence. Remember that?

    28. AG

      Yeah.

    29. MC

      Like you can only make it so big.

    30. AG

      Yep.

  4. 20:1731:27

    Why We Haven't Seen the AI Cyber Apocalypse Yet

    1. MC

      w-why do you think we just haven't seen very much then? Like I-- Again, again, I'm, I'm much older-

    2. AG

      Yeah

    3. MC

      ... than you. I remember very well. So when the [laughs] when the-- Like literally when the internet, when the, when the internet came out, by this point, we had literally taken out 10%, we'd disabled hospitals, we'd taken out critical infrastructure, we'd caused tens of billions of dollars in economic damages from worms. Like, all of that had already happened. And to your point-

    4. AG

      Yeah

    5. MC

      ... we had much less build-out, you know, less of the, the economy was on it. And so, you know, AI has so many people that want to find risks and threats. We're running so fast, so much money has been poured into it, and we haven't seen anything commensurate with the early days of worms. What is that disconnect?

    6. AG

      Yep. Um, look, so I do remember that, those days. Um-

    7. MC

      But it, it is the same time.

    8. AG

      Yeah. Uh, so look, I would just say that, uh, I, I am sleeping well at night, and I don't think there's existential risk right now. Um, I do think there's a lot of infrastructure that needs to be secured.

    9. MC

      Yeah.

    10. AG

      Uh, we have, uh, a product in the market, in, in the detection market, Lakewatch, that helps you do detections. And the space is-- just there is moving so fast, uh, because, you know, you used to have these SOC teams, security operations center people, that would, you know, look at, uh, what intrusions are happening, how are we being attacked, and so on. And now the humans just can't keep up.

    11. MC

      Yeah.

    12. AG

      So this is the whole space, security cyberspace, is being transitioned into fully automated, using agents for detection on the other side. If we don't do that-- I mean, now we're rushing. We are rushing, the industry is rushing-

    13. MC

      Yeah

    14. AG

      ... to do that super, super fast. If we don't do that-

    15. MC

      Yeah

    16. AG

      ... uh, I do think you will start seeing those kind of things, like sites going down, you know, um, you know, whole systems that stop working for a while. And there will be consequences, not existential, uh, but economic damage and, you know, people getting hurt and so on could happen. Uh, so we just have to race very, very fast to do, to do all of those things. Um, it's just you don't have the... The hu-humans don't respond fast enough to the attacks that are happening, so you just, you need to automate all of those. And most organizations are actually not close to doing that. The banks are doing it. Some of the people that are super security conscious are doing it. But most of the industry today is running with old school security operation centers and people that are waking up every day, and there's like hundreds of emails of detections that have fired. Many of them are just false positives, so you don't need to-- you can ignore them. But some of them are not. They just don't have time to go through those. And you need to identify that. Uh, you need to have threat hunting that's automated, where you're actually attacking your own systems automatically with agents and so on. It hasn't happened. So I do think, like if, if we just say, "Hey, this is just like the internet in the early days," uh, you know, bad things are gonna happen. So there is a race going on.

    17. MC

      You know, I was actually very surprised, um, earlier this morning, I was on a conver- Like, I feel like you and I are like pretty close, so we talk periodically. I feel like I, I know a fair bit about Databricks. Uh, I was on a call this morning where a f- a founder was basically like, "Yeah, listen, like, you know, we're doing all of this like observability agent threat detection, and we're [chuckles] using Databrick." I didn't even know that you, you had this offering, like quite frankly. So like, I mean, this is j-just from an education standpoint, like how, how, how extensive have you gotten in like the agent AI observability security safety-

    18. AG

      Yeah

    19. MC

      ... thing?

    20. AG

      Yeah, I mean, we gave a talk this year at RSA, actually with Ben Horowitz. But, uh, the issue is that data and AI is blending with cyber. These two markets are collapsing-

    21. MC

      I don't know. I think we know. So at least-

    22. AG

      Yeah. And the reason they're collapsing is that it used to be like, okay, we have like data and AI, the kind of stuff Databricks and these kind of companies used to do, which is like, okay, you have a bunch of data and you run AI and machine learning, and that just lives separately. And then you have the cyber world. Cyber world is, you know, we want to detect if something bad... If like bad people are trying to hack us, if bad people are doing things, we need to detect that. Okay? But now on the data and AI side, we have agents running. Internally in the company, people are having agents running, and the agents are also like doing things with other people's agents, and they're producing a lot of data.

    23. MC

      Yeah.

    24. AG

      Uh, logs, trails, you know, fingerprints that are being left. Uh, and so, you know, then now you have internally these agents that are doing that. So these, these worlds start merging more and more, which is like, okay, well, we-- all the data that's being produced needs to be analyzed. And the scale at which you need to do that is just like many, many orders of magnitude more than just one or two years ago. So things have changed dramatically. Like 2018, '19, the time it would take from, uh, you know, a CVE, uh, vulnerability being sort of, uh, published until you see it actually weaponized, uh, in the industry would be like two, three years. That went down to, you know, 2022 significantly, but it was still like eight, nine months.

    25. MC

      Yeah, yeah.

    26. AG

      So that's kind of fine. You have eight, nine months from a vulnerability to... That was 2022.

    27. MC

      Yeah, yeah.

    28. AG

      Now, if you look at this curve from 2022 until now, now it's down to, like, basically hours.

    29. MC

      Yeah, yeah.

    30. AG

      So it's, like, down to, like, basically no time. Like, things get immediately weaponized. So, uh, so you need to just do it in a automated, with the data and AI sort of platform approach. So these markets, I'm gonna argue, are g- just gonna collapse, actually.

  5. 31:2733:33

    The 4D Chess Problem: Doom Talk vs IPO Allocations

    1. AG

      other.

    2. SW

      So, so I have to ask, Ali, do you think, um... Something Elon also said, I think it was on the All In, um, you know, summit. He was like, "This is some elaborate 4D chess," because on the one hand you're saying all of humanity will die.

    3. AG

      Mm-hmm.

    4. SW

      On the other hand, you're saying, "Hey, what do you want for your L, you know, IPO allocation?" [chuckles]

    5. AG

      [chuckles]

    6. SW

      Right? And so, I mean, that is probably a more cynical view, but like how do you, how do you reconcile that? I mean, the, the dissonance I think gets a lot of people. Like, how do you think that gets reconciled?

    7. AG

      Look, I think all of these things get mixed. Like, I think there are people that are freaked out. And I do think that there are people that are saying like, "Hey, if there was regulation that would pace us," sorry to use the word, "uh, that would be good for us." [chuckles] Right? That, that would be good for us. But I also think that people have vested interests.

    8. SW

      Yeah.

    9. AG

      Right? These things like, you know... And usually people figure out a way to always get all of these things to align in their, you know, harmonically in their head.

    10. SW

      Mm.

    11. AG

      Uh, so yeah. Do I think that there has been a tendency in the past of, in general, using also marketing stunts, uh, by saying, you know, "Oh my God, this latest model is so good that I trained. It's like unbelievable. It's like almost scaring me." And then how-

    12. SW

      Mm-hmm

    13. AG

      ... the whole world like kinda starts focusing on it. Yeah, there's been that, that kind of marketing going on.

    14. SW

      Yeah.

    15. AG

      Uh, but at the same time also, as I said, the time from CVE to actually-

    16. SW

      Right

    17. AG

      ... weaponize exploit has like been going down from years down to like minutes now, just in like three, four years. Uh, so it's real. The cyber attacks are real, and this is... But, but there's also a great marketing ploy, uh, to, you know, whenever you train a new model, uh, make lots of noise around how much of a, you know, crazy risk it is to the world. It, it helps you, right? Uh, so you know-

    18. MC

      Maybe, and, and-

    19. AG

      ... maybe they're not in contradiction, these things.

    20. MC

      So s- I mean, you, you and I are networking folks. And there's-

    21. AG

      Yeah

    22. MC

      ... a long history of, um, uh, forming third parties to help arbitrate things, right? Like IETF or-

    23. AG

      Yes

    24. MC

      ... you know, IEEE or, you know, even like ICANN, you know?

    25. AG

      I see where this is going.

    26. MC

      No, no, no, no, no. So no, my question to you is like, so I think it's actually this is a very sensible proposal that they actually have.

    27. AG

      Mm-hmm.

    28. MC

      I actually agree with you. You probably want to make sure it's independent, which is not clear right now. Whatever.

    29. AG

      Yeah. And there's gonna be a lot of arguments of who you put there and everybody's gonna disagree.

    30. MC

      Right,

  6. 33:3341:49

    Industry Self-Policing vs Federal Involvement

    1. MC

      right, right, right. But you said, you know, "Why do we have judges?" So that's, like, actually, like the state stepping in is actually quite a different thing than basically industry self-policing.

    2. AG

      Mm-hmm.

    3. MC

      So like at what point in time do you think it makes sense to actually consider federal involvement? Or do you think like now is the time to actually consider actual federal involvement as opposed to like more industry self-policing?

    4. AG

      Well, I mean-

    5. MC

      These are just a very different approach

    6. AG

      ... they are different, but they kind of bleed into each other like, you know, like for instance FINRA, um, you know, is not like a completely independent, uh, self... It is, but you know, it's, uh, linked to the government. So like I think these things are kind of will bleed over. I think it's hard for it-

    7. MC

      Do, do you think that they evolve into... Historically they've-

    8. AG

      Yeah, I mean-

    9. MC

      ... the industry self-polices and then it evolves into regulation

    10. AG

      ... look, if, if they are saying there is existential risk, which they're saying, you know, and they're saying, "Come police us and regulate us," I think it's very hard for regulators to say, "No, we're not gonna do that." So far they've said that-

    11. MC

      [chuckles]

    12. AG

      ... but I think that's not gonna last very long. Uh, you know, I think-

    13. MC

      It's so funny. Wasn't it David Sacks was like, "I've never had a CEO ask us to regulate them." [chuckles] And my favorite thing is like-

    14. AG

      And the CEO's are, "I've never had a regulator that says no to that." [laughs]

    15. MC

      Say no. And like, and so... I mean, the, the reality is like the actual like metapolitical machinery is actually in motion already, right? I mean-

    16. AG

      Yeah

    17. MC

      ... like everyone has a talking point. Obama has came out, it is a major issue. Like do you think that there's a reality that it's too late, this will be a major issue in the midterms, and we're actually going to like heavy-handed federal regulation, and this is all gonna be paused, you know, and goes into the, Anthropic goes into the DOE, and we're, we're past that point? Or do you think we can actually end up with like a sensible self-policing regulation?

    18. AG

      I mean, I think we should-

    19. MC

      Because if you listen to headlines, you cannot

    20. AG

      ... we should strive towards doing the right thing.

    21. MC

      Yeah, of course.

    22. AG

      I think there's still some degrees of freedom of how things evolve and there's still time. And yeah, you're right that largely you have these companies where we're pumping in so many billions of dollars.

    23. MC

      Yeah.

    24. AG

      And the way reinforcement learning works is that you, you know, give it the reward function that's verifiable, like we're gonna solve this math problem or this kind of, you know, this narrow area of programming and so on. And we pour in so much money into that, you can get quite good results in that narrow kind of... That doesn't mean that you're getting that superintelligence.

    25. MC

      No, but you can even, you can even-

    26. SW

      It's probably-

    27. MC

      You can even trick yourself into thinking that like less inputs are giving you a better outcome-

    28. AG

      Mm-hmm

    29. MC

      ... just because you're running so many experiences and you've thought about it so much, right?

    30. AG

      Mm-hmm.

  7. 41:4944:36

    The Enterprise Use Cases Surprising Even Ali

    1. AG

      Why are we-

    2. SW

      So, so how do they get there? Like what are some of the use cases you've seen to date-

    3. AG

      Mm-hmm

    4. SW

      ... that have maybe surprised you to the upside?

    5. AG

      Yeah, I mean, f- first of all, there's like so much, uh, wor-worry about, you know, existential risk and so on. So I think a lot of people just don't know what a, you know, cool use case is where people are actually doing interesting things. Uh, we have a lot of use cases that are, I mean, just fascinating. Um, one that I like is, uh, Crisis Text Line. So-

    6. MC

      Wow

    7. AG

      ... you know-

    8. SW

      Interesting

    9. AG

      ... they actually use large language models with us to, uh, detect if teenagers wanna do self-harm or suicide.

    10. MC

      That's awesome.

    11. SW

      Oh, wow, yeah.

    12. AG

      You know?

    13. MC

      That's cool.

    14. AG

      That's an awesome use case.

    15. MC

      Super cool.

    16. AG

      And it can actually... So it actually saves, uh, saves lives. Uh, so that's a great company, and that's, you know, that organization is doing amazing work.

    17. MC

      Wow.

    18. AG

      Um, another one that's kind of interesting is the Omnipod, which, uh, is, uh, for diabetes patients. They can put the Omnipod, and it uses AI to l- really learn, uh, your insulin release and your glucose levels-

    19. SW

      Oh, wow

    20. AG

      ... and actually exactly release... You know, I don't know if you remember, people used to like-

    21. SW

      Yeah, totally

    22. AG

      ... stick themselves, right? Um, but this now happens automatically, and it's like, you know, self-learned, uh, AI for your body. Uh-

    23. SW

      Wow

    24. AG

      ... you know? It's a cool use case. Uh, Zipline is another one. They're doing awesome.

    25. MC

      Oh, yeah.

    26. SW

      Oh, yeah.

    27. MC

      That's like-

    28. AG

      You know, but when they started, it was like these drones that had, you know... They were completely automated, all AI driven, everything from the, you know- Battery optimization to the routes and everything, and they were delivering food in, you know, areas of need

    29. MC

      Like blood, yeah, blood, blood to refugees.

    30. AG

      Blood to refugees.

  8. 44:3645:24

    How Enterprises Actually Operationalize AI in the Next 12 Months

    1. SW

      [laughing] Yeah, exactly. You're totally right. And so-

    2. AG

      Yeah

    3. SW

      ...how do they-- W- uh, let's say if you map out the next 12 months-

    4. AG

      Yeah

    5. SW

      ...how do the enterprises actually get value? You, you, you dropped the word context, but like-

    6. AG

      Yeah

    7. SW

      ...how, how do they operationalize it?

    8. AG

      I, uh, it's h- actually harder, uh, than most people believe. But, uh, you know, st- first and foremost, we have to make sure that we have digitized everything that's happening in an organization. That actually, you cannot actually just, you know, uh, have a magic wand and make that happen. So, you know, every meeting has to be transcribed.

    9. SW

      Mm.

    10. AG

      You know? Uh, so you, you have to be able to get all the context of all the meetings and everything that's happening. All the digital content has to be fed to the AI. So you have to build... We, we call it an ontology. We build that. But first and foremost, you have to collect that. That itself is a problem in many organizations, because legal teams will say, "Don't record every call. Don't record every thing." So you have to do that in a way where,

  9. 45:2450:55

    Defining Ontology (Beyond the Palantir Version)

    1. AG

      uh, it's-

    2. SW

      Can you define ontology for everyone? 'Cause I know Palantir says the word a lot.

    3. AG

      Yeah.

    4. SW

      But it's not like they own the word ont-

    5. AG

      Yeah

    6. SW

      ...like what does that mean? And for the people listening, like how should they think about it?

    7. AG

      Yeah. I mean, it's, the, you know, ontology just means that in an organization, uh, the relationship, uh, between all the abstract concepts of all the goals and all the departments and all the people and all the projects that are going on, what do they exactly mean, and what's the relationship between them, the people, the resources, and what that company does? Uh, so it's the difference between a person who is a new employee in the company and just started today-

    8. SW

      Mm-hmm

    9. AG

      ...and a person that has worked there five years. You know, let's say they're equally skilled.

    10. SW

      Perfect.

    11. AG

      They have the same educational background. They're equally smart and hardworking and all of that. But one, it's, you know, her first day today at work. The other one, she's been there five years.

    12. SW

      Yep.

    13. AG

      Uh, what's the difference between these two people? The one has an ontology of how that organization works, who the people are, how you get stuff done. Don't look at the org chart.

    14. SW

      Mm.

    15. AG

      That's... Don't go ask that person.

    16. SW

      Yeah.

    17. AG

      He will not get anything done. You go ask this person, you know, he'll get it done for you. And, and, and that's not how it works. You don't need to file that paperwork here. And, you know, and this is, th-this project, this is what's going on. This is essential. So there's just a lot of ingrained knowledge that's sitting-

    18. SW

      Yeah

    19. AG

      ...in everybody's heads-

    20. SW

      Mm-hmm

    21. AG

      ...who knows how an organization works. That's why it's, people say, in startup land, they say, "Hey, if you lose most of your people, that company can't recover from it." It doesn't-

    22. SW

      Mm-hmm.

    23. AG

      You can't just replenish and hire new people.

    24. SW

      Right.

    25. AG

      Like, the people are so essential. How do we get that context, that's the ontology, and give it to the AI? Part of that is we just have to have, you know, the recording and all of that. But the second part is how do you actually distill it down into a graph?

    26. SW

      Mm.

    27. AG

      Actually, a digital graph, uh, that you can then feed to the AI. So the way a lot of the, uh, agents work today, like a Claude Coder, any, any of them, Codex or Pi or, you know, OpenCode, or you can go through the whole slew of them. You know, they have this loop, agentic loop. It can reason, but then it goes and checks every resource one at a time. So it'll go to this MCP server for your question and try to see, is the answer here? Is there another one? It synthesizes it and gives you an answer. But it's kind of slow. I, I liken this to if Google would've built Google Search this way 25 years ago, we would've said, "Okay, we're gonna get 10 blue links. We search for our key terms here." But instead of giving you 10 blue links, it would've gone to one website, summarized with an LLM what it does, found a few h- hyperlinks, jumped in parallel to a few of them, read a few websites, done that for 10 minutes, and then given you, like, its best 10 blue links it would find. Well, that would be very expensive. Costs a lot of money to do that every time, go on the web. Two, it would've, uh, taken a long time, and then you gotta wait to 10 minutes. And three, the quality would be bad, 'cause you're actually only looking at a very small subset of everything that exists out there, right?

    28. SW

      Mm-hmm.

    29. AG

      Uh, so how do they do it? They have an index, right? You never leave Google servers. You search for-- It hits the index. The reverse index immediately gets you the 10 blue links within, you know, less than 100 milliseconds. Um, we need to do the same thing for the AI. So the ontology is that we need to compute that index offline all the time. Um, so it's almost like the PageRank algorithm that Google had invented back in the day, but it's more complicated. 'Cause Google was just looking at a web where everybody can go on the web. Here, there are permissions.

    30. MC

      And the, and the links existed and-

  10. 50:551:06:38

    The Finance Anecdote: Real AI Value in the Boardroom

    1. AG

      see everybody go to their phone.

    2. SW

      Can you share that finance, uh, like the finance anecdote you, you mentioned once in a board meeting?

    3. AG

      Yeah. It's, uh, yeah, sure.

    4. SW

      [laughs]

    5. AG

      Internal board meeting. Uh, yeah, so, um-

    6. SW

      Only in kosher. [laughs]

    7. AG

      Yeah, exactly. No, it's, it's, uh... Uh, it, it actually needed for, uh, for one of our presentations. I need to know how many customers do we have in Fortune 500-

    8. SW

      Uh

    9. AG

      ... that use. What's our penetration of Fortune 500? And I asked one of the people in sales ops, 'cause I thought she would have it. And she texted me back and said, "Oh, sorry, I can't log into Genie right now, I'm on a flight."

    10. SW

      [laughs]

    11. AG

      And I said, "Well, if, if you're just gonna log into Genie, I can do that myself."

    12. SW

      [laughs]

    13. AG

      Like, I don't... I asked you 'cause I thought you had, like, something authoritative that I don't have access to.

    14. SW

      [laughs]

    15. AG

      So then I, I was kind of a little bit angry, so I texted the CFO instead, Dave. And so I texted Dave, and I said, "Hey, do you know what our Fortune 500 penetration is?" And he just copy-pasted a screenshot of Genie back to me.

    16. SW

      [laughs]

    17. AG

      That's it. So he also has that.

    18. SW

      I love this so much.

    19. AG

      So then I said, "Does anyone do anything novel here? Are they... Is everybody just going to Genie and asking the ontology, you know, for questions?" Um, so-

    20. SW

      It's like, "Let me Genie that for you instead of let me Google it." [laughs]

    21. AG

      Yeah. That's what everybody's doing. Now we just say, we say, "Hey, can someone just Genie this?"

    22. SW

      [laughs]

    23. AG

      Like, you know, can you just get it from the ontology? Uh-

    24. SW

      Awesome

    25. AG

      ... so I do think it's a game changer. Um, but it's not just you press a button and you have an ontology in an organization. And I think Palantir actually has done a great job of going to organizations and getting a lot of that tacit knowledge written down and getting it into the organizations. We automatically then take that and build the graph, and then we feed that graph into the agents-

    26. SW

      Mm

    27. AG

      ... so that we can answer the question and answer it in a way that business leaders would like to see it, which is in graphs, you know, analytical way, and a way where you can interrogate that question, um, and, you know, continue asking questions and getting answers, uh, to those, so that you can make decisions. And then dissem- disseminating that information in the organ- organization.

    28. SW

      Yeah. It's, it's pretty amazing. Um, you, you sort of bookmarked the, uh, developers are obviously using AI, questionable value. Um, I wanna follow up with you on that because, um, I feel like you guys were one of the earliest-

    29. AG

      Mm-hmm

    30. SW

      ... um, and I, I say, I don't wanna use the word token maxing 'cause it has such a negative connotation, but I think in terms of applauding people who can use-

Episode duration: 1:06:52

Install uListen for AI-powered chat & search across the full episode — Get Full Transcript

Transcript of episode GzEtpAKYRvE

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.