The scaling laws have been remarkably robust.
There's a lot we don't know yet in terms of the vulnerability of these systems.
If you don't need to buy the Galacticos, why do you have, like an Andrew Tulloch, a Daniel Gross, an Alex Wang and the Galacticos assembling?
If I gave you $10 billion, what would you spend it on first?
Knowing what you know, what do you not let your children do?
This is 20VC with me, Harry Stebbings, and today we have one of the leading minds in AI, Joelle Pinault, on the show.
Joelle is the chief scientist at Cohere, where she leads research on advancing large language models and practical AI systems.
Before joining Cohere, she was VP of AI research at Meta, where she founded and led Meta's AI Montreal lab.
Joelle is also a professor at McGill University and renowned for her pioneering work in reinforcement learning, robotics and responsible AI development.
But before we dive into the show today, are you drowning in AI tools?
Chat, GPT for writing, Notion for docs, Gmail for email, Slack for comms, and you're constantly copy pasting between them all, losing context and losing time.
This is the AI productivity tax, and it's killing your output.
At 20VC, we're all about speed of execution, and Superhuman is the AI productivity suite that gives you superpowers everywhere you work.
With the intelligence of Grammarly Mail and Coda built in, you can get things done faster and collaborate seamlessly.
Finally, AI that works where you work, however you work.
Superhuman gets you from day one, with zero learning curve, and it's personalized to sound like you at your best, not like everyone else using generic AI.
Get AI that works where you work.
Unlock your superhuman potential.
Learn more at superhuman.com forward slash podcast.
That's superhuman.com slash podcast.
And once you're moving faster with Superhuman, make sure you're moving safely with Vanta.
Customer trust can make or break your business.
And the more your business grows, the more complex your security and compliance tools get.
It can turn into chaos, and chaos isn't a security strategy.
That's where Vanta comes in.
Think of Vanta as your always-on AI-powered security expert who scales with you.
Vanta automates compliance, continuously monitors your controls and gives you a single source of truth for compliance and risk.
So, whether you're a fast-growing startup like Cursa or an enterprise like Snowflake, Vanta fits in easily into your existing workflows so you can keep growing a company your customers can trust.
My listeners can get $1,000 off Vanta by going to vanta.com forward slash 20VC.
That's V-A-N-T-A dot com forward slash 20VC, 20VC for $1,000 off Vanta.
You have now arrived at your destination.
Joel, it is so great to have you in the studio.
I've heard many great things from Nick, Aidan, Shrep.
So thank you so much for joining me.
Thank you, happy to be here.
Now.
You spent over six years at Meta and I want to start there because it's a very transformative time and place.
What are the biggest takeaways for you from that time and how did that shape your mindset to how you think today?
Well, I was there from 2017 to 2025 and you have to see just how much AI changed over that period of time.
What we were really focused on is fundamental AI research.
One thing that I've learned is just sometimes how long it takes to prove out a hypothesis.
We feel like AI is moving at the speed of lightning.
But in fact there's some things that it just takes a few years to mature to get the right optimizer, the right compute, the right data, for that to really make a difference.
I look at where we are today and everyone kind of goes, it's here, it's here, it's here.
And then you actually look at what a lot of the leaders have been saying recently, where it's like actually Andre was saying it's not the year of the agents, it's the decade of agents.
Sam's kind of pulling back too.
Have we got over our skis and we're actually kind of all pulling back, realizing that time is the factor we need to rely on?
Well, I'll give you an example.
You know, I've been in research for a couple of decades now.
I've been working on reinforcement learning for over 20 years.
And suddenly everyone's talking about reinforcement learning.
You know, since the advent of reasoning models, agents and so on.
So, you know, sometimes you have to be a little patient with these ideas.
And the right algorithmic tweak, the right context, the right problem domains just opens up the magic.
I was listening to Andre yesterday and he said in this show that reinforcement learning is terrible.
Yeah.
Less terrible than 20 years ago.
Have we over-invested in RL-based methods at the expense of maybe more scalable alternatives?
Oh, I'm still super bullish on RL in that, like the concept itself is so fundamental.
You know, this idea of training through a system of rewards, of indicating what's valuable and what's not valuable through numerical values like that is so fundamental.
It's not going away.
Now you know where we're maybe getting a little bit ahead is thinking that just RL out of the box is going to give us AGI.
That part, a lot less so.
You know, if you look at the curve of progress, RL is terribly inefficient.
And so the amount of signal you need to get in order to really shape the behavior of a model is far from where we are today.
And so we'll need to figure out how to really deal with this learning efficiency problem.
You're probably thinking, what did I get myself in for?
And I don't blame you.
I ask questions that I think everyone else thinks, but I'm not afraid to say I don't know.
Why is RL so inefficient?
There's a few reasons.
You're going to get me on like a deep topic.
There's a few reasons.
One is the fact that RL is about sequential decision making.
So you know, think about.
You're starting.
At a point you need to figure out what you're going to do next.
And you might pick the right side of the branch or the wrong side of the branch and then like, the road keeps on splitting.
So every time you make a mistake it sort of compounds through the length of the series of actions you're making.
So that means like the amount of error you can make can be very, very large and to get it right is quite difficult.
Sometimes people compare it to like a needle in a haystack is like finding the right solution in RL.
So there's that part.
The other part that's hard is is the fact that to train the system, to train the models, you have to essentially take actions to learn.
You can't learn from static data.
You can learn some things from static data, but actually to get the right policy you need to test it out.
And so that means you need a simulator.
You need to get the synthetic data.
All of that can be really expensive also.
And so we have difficulty getting like just a variety of environments in simulation to test RL.
When we look at the cost curve for RL, you said you've been working on it for 20 years.
Have we seen that dramatically come down?
Will we see it continue to dramatically come down?
Or is it a case of it is just a fundamentally expensive method of training?
It's coming down, especially in domains where we have good reward functions.
So the place where most people started hearing about RL is around the AlphaGo time.
You know the game of Go, which was sort of one of the goals for AI.
Many people thought we were at that time.
We were still a decade away from being able to have machines play Go at the level of humans.
And out comes a team from DeepMind, you know, goes off, plays against the world champion and shows that RL can basically do it.
And so I would say, in cases where we clearly know what's the goal, we can write down precisely the reward function.
We're good.
We can make a ton of progress.
So that's why you're seeing progress in mathematics very well-defined reasoning tasks games, these kinds of things.
Arel, to shape the behavior of models, to get them to be social creatures that we have no idea how to do.
I mean, I don't know if you have children, but like shaping their behaviors, you know the number of times you can repeat the same thing and still they do something else.
And so there's something there.
You don't know how to write that out mathematically.
And that's where I think we're still in for some hard work.
So when we look at the training versus the inference market today, we've had so much weight on training so far and it's been incredibly costly and expensive.
And then I hear everyone say, well, actually, inference is 95% of the market.
And that's where it's all going.
And that's where Nvidia will make most of their money.
How do we think about the cost curve applied to training versus inference and where it sits today?
I think there's a lot of different variants and, If you'll allow, maybe I'll pivot to where I'm going with Cohere now.
I joined Cohere less a month ago, super exciting company.
I think one of the things that Cohere is doing is actually to develop AI models that run on-premise.
So that means enterprise bring it in, they run it locally.
So the company has to worry about the training of the models.
Obviously, we want world-class models for the needs of enterprise.
Doesn't have to worry about the inference.
Doesn't have to worry about the inference cost.
The client's customers have to figure out what's the right way for them to digest the AI.
That means like there's a lot of motivation to have very efficient models so that they can run really efficiently on-premise.
So, you know, we get caught up into one paradigm, but there are other paradigms as well.
If they're the ones paying for the inference, is there not less incentive to make them efficient?
Because, sod it, you're not the one paying for it.
If you're the one paying for the inference costs God, I want it to be as efficient as possible, because it's my dollar going to that.
But if it's IBM's dollar, yeah, I would love it to be efficient, but we're not paying for it.
We're still in the early days of AI adoption and enterprise.
So what's good for the client is good for us.
Totally get you.
What's the biggest challenge about capital efficient AI today?
I know that sounds strange.
When you look at the economics, so to speak, what's the biggest challenge?
There's a lot of challenge today, I think, in terms of the economics of AI.
I think one of the biggest challenge, the fact that it's very hard to have predictability.
Everyone wants to know when are we going to hit the breakthrough?
Everyone wants to know how many GPUs do I actually need?
Everyone wants to know, like, what's the return I can expect?
There's just a lot of uncertainty built into the system.
A lot of that is because there's a lot we don't know about this technology.
And so that means we have to take in quite a bit of risk.
When you're building out, whether you're building out your data center, whether you're building out your workforce, whether you're trying to figure out you know how much data to curate.
And so that makes it difficult for a lot of people.
People want answers.
And this is a world where we don't have that level of predictability compared to other industries.
Does progression happen in kind of a linear fashion or does it happen in step functions like AlphaGo, like a deep seek which, depending on kind of what you believe, suggests a lot of efficiency in terms of model improvement?
Is it step function or is it linear?
I tend to decompose different ingredients that lead to progress.
You know people often talk about, like the algorithms, the data, the compute.
I think in general, compute and data have a more linear effect on progress.
You build more compute, you run bigger models.
You can typically get better performance.
You feed in more data.
It's not just quantity you need to worry about quality and diversity as well but Roughly it's more linear-ish with respect to the data.
The algorithms are the ones that have the nonlinear effect.
And so you can explore lots of ideas and then something like the transformer comes along and just changes the paradigm.
And it's not just your transformer, you know.
On the optimization side, suddenly we hit upon Adam, which is a technique to do the optimization of your model.
Changes in the paradigm.
Reasoning.
Suddenly we start thinking about how to put that in the loop reasoning.
And it changed the paradigm.
So those ideas tend to have a nonlinear effect.
The challenge with these algorithmic ideas though, is that actually, It may take a long time to prove themselves out.
So like the paper can be sitting out there.
There's thousands of papers coming out.
The idea is sitting out there and we may not think to try it with the right data, at the right scale, with the right combination of hyper-parameters.
And so you don't notice that effect for a while.
So it's hard to predict and it's not linear, more on the algorithmic side than I think on, whether it's data compute, even talent or other things.
With respect to Google, I mean Transformers, obviously birthed in Google and started as papers for many a couple of years.
You mentioned that you compute algorithms, data.
If we just kind of go through them to understand, everyone suggests that it's weird.
There's two different worlds.
Scaling laws exist.
Just throw more compute at it.
When you look at data center investment, when you look at all desirability of compute data, And then again you have GPT-5 seemingly focusing on efficiency and other signals.
Do scaling laws play out from here?
And if so, for how long?
The scaling laws have been remarkably robust.
They don't play exactly as we expect, but still, they've been remarkably robust.
Lots of people have bet against scaling laws in the past.
And I would say overall, you know, we've seen a pretty robust effect.
They don't work alone.
We also need these algorithmic innovations.
But most of the time, you know, I wouldn't bet against it.
On the algorithm side, is that the hardest to innovate on?
You could think about well, you can buy more compute.
Might be hard, but you can buy more compute and data.
There are different ways, whether it's synthetic or human.
Is algorithms the hardest to innovate on? it's certainly the most creative work to be done.
And you know the space of ideas is so wide that I would say it's hardest, in the sense that, like you, can move in.
You know I'm a researcher at heart.
You can move in so many different directions and picking the right one.
You don't know until you get there, whether it was the right one or not.
It's a little bit like reinforcement learning.
So in that sense, I think it's, It's the most interesting one.
It's the most frustrating one.
And it's the most difficult one, certainly from an investor's point of view, to know where to put your chips.
Speaking of kind of knowing where to put your chips, and moving from purely a research lens with Matter to now also building product.
Is there ever this inherent conflict between intellectually interesting research with the need to productize and monetize?
And how do you think about that?
I mean, one of the reasons I'm really excited to be joining Cohere actually is because we're at a stage where AI is really starting to be useful.
Maybe not as useful as people think it is, but we are there.
And by working on AI that's going into enterprise, I feel we're going to get started such an interesting signal of what works and what doesn't work.
You know we keep on talking about, you know agi and ai for the masses and so on, but actually, like when you need to sell ai to a business, you get a real signal of what works, what doesn't work.
And that's what i'm most curious to see and you know we've been using these academic benchmarks for many years.
You get some signal, but it's not the same as as getting this to do productive work.
I'm curious to learn out of that.
You know we're going to get new types of data.
We're going to get, I think, a lot of insights that are then going to drive the research ideas.
I think that's the other thing to think through when you have a large space of ideas to explore.
Getting that feedback signal from the real world is super useful to guide you through that search of ideas.
I just had a great chat to David Khan at Sequoia, who said that he thinks a good barometer for utility value within enterprises is like.
Does it have the ability to replace the work of your bottom 5 in any category?
He says, like, we overestimate a lot.
Can it replace the bottom 5% in any function?
And if it can, that's a very meaningful improvement.
Do you think that's a good barometer?
And how would you advise an enterprise on whether something's useful or not as a yardstick?
You know, I prefer, in terms of a barometer of productivity, something a little bit different, which is to say, can most of your employees do 10x the amount of work with AI versus on their own?
That to me is actually a better barometer.
I think human and AI have very complementary abilities.
So to just like, flat out replace a portion of your workforce is actually pretty unrealistic.
Some may try and some may be slowing down their hiring, but I actually think Respectfully, I think 10x-ing your work feels more unreal.
Is that not a bigger ask?
I'm almost more intimidated by 10x-ing my work.
Oh, I don't think that's unrealistic at all.
Wow.
In a timeline that is the next couple of years?
Yes.
I'm sorry, how does that actually shape out then?
I think you have to identify very concretely the types of work that you are delivering.
But I think we're starting to see, like you know, Hollywood quality productions being made in a matter of hours.
We're seeing.
You know.
To take a super concrete case like machine translation, if humans are doing the machine translation compared to machines doing it.
You go from hours to seconds on long form text.
Multi-page documents.
And so for a lot of work, It's not like AI can do all of the work.
Humans still need to ask the right question.
They need to verify the information.
They need to shape the tasks.
But once the task is well-defined, the product are clear, like all the design considerations are fed into the prompt.
You press the button and you've got an answer in seconds for something that used to take sometimes weeks and months.
I'm just kind of trying to reevaluate a belief that I had for the last few months, which is like I'm a venture investor for a sin.
For all of us to make money, we need to see the transition from kind of human labor budgets to AI spend.
And it's with that transition where we obviously see the town massively increase and we make a lot of money.
But when I hear you say that, I suddenly question that assumption.
It's like the barometer for whether we make money, because you're suggesting that actually we don't replace the human labor budget.
It just makes us 10x more efficient.
Is that correct?
Yes.
You know, there's a lot of nuance to all of that.
And some work will be harder to get that same level of efficiency gain, whereas other work you'll see 100x in terms of efficiency gain.
But I do think that for a lot of the work that's happening right now, that's absolutely feasible.
Where do you think, I'm sorry, where is the efficiency gains most tangible?
It goes back a little bit to this notion of what are the tasks that we can specify well.
In any case, where we can be very precise about what a great result looks like, we'll be able to make that task automatic much more easily than tasks that are much more nuanced and have a lot of complexity.
So it's ambiguity.
Ambiguity in the specification of the task is what's hard for our machines.
How have you seen enterprise reaction to this?
There's fear from workers sometimes, there's excitement from leaders, there's apathy sometimes.
How have you seen and measured enterprise response?
A lot of the workforce can be reasonably fearful about job displacement.
There's a lot of also individuals who have, you know, a bit of an instinctive reaction to change.
And change can be hard for a lot of people.
And we're seeing a lot of change in a very short time span.
And so I think there's also a generational effect.
I think for some generations there's that change is is more, more jarring, i think, for the younger generations have teenagers, young adults, at home.
For them yeah, it's just native.
They just, you know, kind of you know gonna grow up with that technology in a different way than than some of the older generations.
It's interesting Sam, speaking of children there at home and how they engage with it.
Sam Wattman said that young people engage with it as an OS to the world and AI is that companion for them, and then older people use it as a next-gen Google.
Do you agree with that, and do you see that in your work?
I see a lot of people using it as a tool more than as a companion.
People have this Swiss knife in their work life all of a sudden.
That can be super helpful, but that's really most of what I see.
What are enterprises biggest challenges with AI adoption at scale?
You know, one of the challenges is to make sure that the AI comes in and can be integrated in their workflows, their processes, their information.
And so the challenge is to deploy in a way that allows them to exploit all of the information systems that they already have.
And some of them have accumulated these over decades.
That's, you know, some of the work that remains to be done.
So it's integration with existing systems and data flows, right?
And of course, that's something we see a lot of cohere because we do on-premise deployment.
One of the things we focused on the most is data confidentiality and security, so that enterprise can exploit all of that information.
So that's top of mind for us, but it's also a huge opportunity.
I would say there's a big interest in that.
But making sure to get that compatibility, I think, is a challenge.
In many cases, For people, change is hardest for people.
And so you have to get them curious about using the technology.
Many people feel they have to get it right the first time.
And I really think like a spirit of exploration and curiosity is much better suited to the face of maturity of the technology that we have today.
We don't have all the answers of how it should be used.
That's going to come from people on the field.
Security is a topic that we quite often glaze over, especially when investing in kind of application layer AI tools.
What does no one know about AI security that people should know?
With respect to AI security, I think there's a new front that's opening up with the development of agents.
And frankly, there's a lot we don't know yet in terms of the vulnerability of these systems.
With LLMs, we're starting to get a better understanding.
We've had quite a bit of red teaming exercise and jailbreaking and so on.
And so people have identified different risk vectors prompt injections, things like that, which are vectors for malicious actors to interfere with a system.
With AI agents, we haven't seen that.
And one of the features of computer security in general is often you know, it's a bit of a cat and mouse game, quite frankly like there's a lot of ingenuity in terms of breaking into systems, and then you need a lot of ingenuity in terms of building defenses, and so we just have to stay very active in that sense.
What are the potential vulnerabilities though in an agent world?
In terms of agents, you know, we worry a lot about hallucinations in LLMs.
The parallel in agents is impersonation.
So agents that come along and are essentially impersonating entities which they don't legitimately represent.
And, in doing so, taking actions on the behalf of these entities where they don't legitimately represent, whether it's infiltrating, you know, banking systems and so on and so forth.
I do think we have to be quite lucid about this develop standards towards it, develop ways to test for that in a very rigorous way.
There's ways to reduce that risk drastically.
You run your agent completely cut off from the web, you're reducing your risk exposure significantly, but then you lose access to some information.
So, depending on your use case, depending on what you actually need, there's different solutions that may be appropriate.
That's a really hard one because then verification becomes the most important thing.
But then it's like, who's the arbiter of verification?
Is it governments?
Is it companies?
How does one think about that?
Who says you're a valid agent versus an invalid agent?
Governments can be good for defining standards on which we all agree.
Companies are much better at building the solutions at scale and deploying them.
Do you think governments are good at setting the standards when you look at AI and where we're at?
And then when you look at the sophistication levels of government programs or decision makers with respect they're just a little bit behind.
Do you think they are actually equipped?
I don't think you should look at where government are in terms of necessarily AI regulation.
AI as a field is so incredibly young and fast moving and by nature.
And there's some good in this.
Governments are moving a little bit more cautiously and usually need to benefit from our knowledge to make good policies.
And so I do think you can look at other fields in terms of regulation.
You know you look at aviation.
The security record for aviation today compared to where we were 50 years ago is just incredible.
And governments have played a role in defining that in terms of standards and in terms of what are the norms, and so on.
So I'm quite hopeful.
I'm an optimist about this.
Maybe it's my Canadian side that governments can play a useful role in many cases.
You know, clear standards actually means reducing uncertainty for a lot of companies in this space.
But we shouldn't expect that to be ahead of the technology.
I think that would be the wrong order of things.
In some sense, we need to develop that technology with enough of a creative space.
And we need to learn fast and then develop the right guardrails for that technology from the real earnings we have.
We mentioned their kind of governments and their role.
When I had Nick on the show he was saying actually the benefits of not being an American company, given some geopolitical challenges sometimes.
I'm just intrigued.
Do you think we will have these kind of sovereign models for each geo?
You know, we have Mr Al in France, Coherence obviously in Canada, founded in Canada, but then you've got global kind of HQs.
Do you think we will have these sovereign models and regionalized winners?
I do think it's healthy that there are models that are getting built in different places around the world, not just in the US and China right now.
I think this is healthy in terms of diversity of thoughts.
I think it's healthy in terms of having a greater amount of people with access to technology.
I do think for Cohere, the vision isn't to be a Canadian company.
The vision is to be a global AI company, and I think Yes.
You know, we have a headquarter in Toronto.
We have teams that are distributed around the world.
We have a great team here in London, as well as in the US and France and other places.
And so having that ability to deploy models that operate across the world, I think it's going to be absolutely an important part of the strategy for Cohere.
I think there's a great opportunity.
What it gives us to be headquartered in Canada is like a sensitivity to the fact that it's not always a one size fits all solution.
You know, I go back to the research we've done.
We've done leading work in terms of multilingual model.
And it turns out it matters.
You go to Japan, you go to Korea, and they do want models that work well in their language.
People in the workforce are still operating in the language of the country.
So having a company that is attuned to that, that values that internationalization of model, is actually important on the global market.
On the team building side, obviously, Canada has great talent.
You mentioned, obviously, some in London as well.
What have been your biggest lessons?
Observations on team building in this talent frenzy that we're in also.
How do you analyze that?
One of the things that's important when you're building a team for AI.
I do think you need people who have vision, who have a sense of what can we create, just because we're in a space where there's so much innovation that is still needed.
So you need an ingredient of vision.
That can be one, two, three people who bring that ingredient of vision.
You need people who have amazing execution muscle.
They don't care that it's their idea.
They care that if the team agrees on an idea, they are just going to push this and get it done.
They're going to build a system.
They're going to run the experiments.
They just have that technical rigor to execute.
And then you need people who, kind of like, keep the team together, who have like the sense of like who needs what to operate well and who are that social glue.
Humans are still social beings and that social glue in a team matters a lot.
Where I've seen it fail is to have just sort of one type of person inside the team.
So I'm a big believer in building teams with diverse complementary talents.
So you can't just buy the Galacticos?
I don't think you need to.
I think you really have to be thoughtful about putting people in a group.
The other thing that helps a lot is for the team to have focus.
If it goes in all sorts of different direction, you'll lose that power that you get from people working together.
So having a lot of clarity, what's the North Star?
What's the goal?
Where are we going?
Even if over time that needs to change.
But that level of clarity is required for everyone to be working in the same direction.
If you don't need to buy the Galacticos, why do you have like an Andrew Tulloch, Daniel Gross and Alex Wang and the Galacticos assembling?
Is that wrong?
You do need a few of these like uber talents in the team.
There's a relatively short number of people who just... understand this technology very deeply.
You do need some of this talent, and if you can afford it, you should get some of that talent.
But you don't need all of your team.
You need like a team with complementary skill as well.
Does that create a good team?
Like if I gave you, you know, 10 billion to go build a team and you could buy a couple of these luxury star players.
I feel like it's top Trump's cards for like sports teams, but you can buy a couple.
Does that create a good team when one is a 3 billion person and then the rest are just average 50 million people?
Yeah, I wouldn't say no if someone offers me the opportunity to hire.
There's definitely some really talented people in the field and they deserve to be fairly compensated.
This technology is going to make a lot of people very rich and have major effects in terms of society.
And so, you know, we should be rewarding the talent.
But I'd be very thoughtful about what are the teams that I put together and how do they work together, rather than just like higher roster of superstars without being thoughtful how they're going to work together.
So you it's so funny.
So, because of the impact that you can have in these teams, actually the multi billion dollar price tags that you see can even be justified.
Time will tell.
I don't think it's necessarily needed to go at that scale.
But time will tell.
If I gave you $10 billion, what would you spend it on first?
One of the things you need is a balance between talent and compute.
If you have too much talent and not enough compute, you're wasting your time.
Usually like an equilibrium between those two pieces.
I think we often underestimate the importance of data and data is getting more and more expensive.
And so I would certainly spend a good chunk of it on data as well.
So many things to unpack there.
Do you feel you have sufficient compute today?
I think we are reasonably well resourced in terms of compute and building the models that we want to build.
Yeah.
So access is not a massive problem?
No.
Why is data becoming more expansive?
Data comes in different forms.
On the one hand, the days of like having data labelers who can say this is a cat and this is a dog are somewhat over, like the easy task the AI can do,
So we're getting in a space where we need more specialized tasks.
So, you know, imagine you're building AI for enterprise.
There's a particular business logic you need to make sure that you're catching the errors.
You're going to need someone with like deeper understanding of the tools.
So that's more expensive. talent to come in and actually prepare the data.
There's also a lot of data that's synthetic data.
When you're building agents, you need to build environments.
And to build environments.
You need some pretty creative folks who are going to build you synthetic simulators.
We've seen this on the robot side for many years, people building robot simulators.
Now you're building AI for enterprise, so you need to think of how are you going to simulate these work processes in a reasonably realistic way that the AI can train on that?
And so that generation of environments and benchmarks and dynamic domains can be pretty expensive too.
When you look at the expanse of data and then you said oh cat dog, lamppost got these caps.
You know, click the ones, I get them wrong.
I legitimately get them wrong.
I'm like, Jesus.
They're getting harder.
They're getting so hard.
They are.
It's not just you.
The other day I called up my CFO.
I'm like, I failed the Revolut.
I'm so sorry.
I'll try again in half an hour.
Please let my agent answer that one for me.
It was embarrassing.
But the question that I have is you know, when you look at McCaw, when you look at Surge, when you look at Turing, how do you evaluate that market which is providing a lot of that talent?
Is that an ongoing, enduring market or is that just hey?
For the next three to five years, we'll need it in the training phase of these models.
But I don't know what it looks like beyond that.
I don't think it's a phase, in the sense that I do think this partnership we'll call it between humans and machines, where human provide guidance to machine.
We are in this for a long time.
What will change is the nature of the information that the AI provides, versus the information that the humans must provide as a complement.
Some of these firms may not be around in five years, but this notion of having humans guide the behavior, guide and train the behavior of AI system,
It's super interesting.
I kind of as an investor and one of them, see all of them converge around like needing to do three things now.
They used to just kind of be talent acquisition.
Oh, we'll get you these people.
And now they're like we'll get you these people and we'll get you high quality data that you can really use.
And now it's like, oh shit, we need this third pillar, which is we'll also help you implement that data into your models, do training and help you with benchmarking and proving that it's actually viable.
And now they need all three.
Are you seeing that third one where it's like implementation of that data as well?
They don't just hand it over the fence.
There's definitely some of that that's happening.
I think for me, the even bigger trend we're seeing is the moving from just labeling data to crafting environments to produce new tasks.
You said about synthetic data and that also being a very important segment to consider.
Do you get model degradation when you get this kind of reinforcing loop of models learning on synthetic data, which creates more data for synthetic and it actually degrades or does it improve?
It really depends how you're generating your synthetic data.
So in some domains, if you think like images languages, like LLMs talking to each other at some point, you definitely get the degradation.
And that degradation is due to essentially like a loss of diversity of your data.
So you know you can make an analogy.
You know you take a bunch of people, put them on an island and let them reproduce.
You know, at some point the genetic diversity is going to keep shrinking.
And so you get a reasonably similar phenomenon with models, because you're not injecting diversity into the data.
So there are domains where lack of diversity means you get a collapse of distribution.
There's other domains where you don't need diversity.
If you think of like playing chess, playing Go, these kinds of games, we know exactly how to generate board configurations.
And so we can generate tons of synthetic data not endless because it's a closed world, but still tons of synthetic data and through that, learn for a long time.
Then there's domains that are sort of in between.
If I think of coding, we can generate synthetic code.
You take normal code and we know how to inject diversity into the code.
Like I can take a couple of repositories, mix and match, apply an LLM to transform it.
And so there's a way to generate synthetic data.
The language is predictable enough and there's enough structure that I also know how to inject diversity, so that you don't get that collapse.
So the hope is that, especially in these domains, we can use a lot more synthetic data and do it without suffering from the degradation of performance.
Do you worry that we are creating a world with just much worse code?
A lot of people are concerned by the quality of code that's being outputted and actually how we're just relying on it pretty haphazardly.
Do you worry about that?
Let me make an analogy in terms of the quality of generation.
You ask about code generation, but let me take you back to 2015 and image generation.
I don't know if you have it in your mind, but the quality of the images that were generated.
We had image generation models in 2015.
They were really bad.
The resolution was bad.
The composition was bad and so on.
And from 2015 to about 2022 or so, we saw huge progress in terms of the quality of the image generation.
So you think of code generation, like right now, we're in the phase we were for image 10 years ago.
Yes, there's a lot of bad code that's getting generated.
There's a lot of code that will get thrown away.
But wait another 10 years.
And I think the quality of the code that's produced is going to be excellent.
What will the developer world look like in 10 years when that is the case?
If I carry my analogy further, I don't know if it's a reassuring scenario, because if we look at where we are today in terms of image generation, there's just like the volume of image getting generated is huge.
What matters now is sort of, you know, picking the quality out of the volume.
And so, if I fast forward 10 years on code generation, when we have the ability to generate a ton of code to do a ton of different things, we're going to need some space selection mechanism to decide what code we actually want when there's actually value.
And so that's going to come, there's still going to be some sort of editorial design choice.
Someone needs to decide like, of all the code we can generate, what's the code we want to generate?
What do we need to be running in terms of our digital world?
So it's like a chief curation artist.
Yes, within the curation, that curation doesn't go away.
Curation, verification this is work that doesn't go away.
Does the structure of teams fundamentally change then?
It's funny kind of playing that back to you and then also playing back to what you said earlier about the human.
If that is the case, there's not much of a partnership, is there, between human and AI?
It's a chief curation person sitting on top of a huge amount of artificially created code.
Well, that's your 10x productivity improvement there.
It is, but you're ticking that box, but it removes the human element there.
You still need people with intent.
That's one thing that you know.
You need to decide what you want to build, and what purpose does it serve?
And so that intent is still there, that role of critique is still there.
So the team composition does change significantly.
Once you suddenly have designers who in their hand have amazing tools to go directly from the ideas in their head to the digital world, maybe eventually to the physical world, that equation definitely changes.
Do you think prompts and the way that we interact today with prompts and with chat largely is the enduring interface for human engagement with AI?
Yeah.
It's awfully limited and prompts can mean a few different things, but the idea of like typing in a box, that to me is very limited.
And we're going to break out of that box already.
We're seeing a lot of cases where voice is a lot more natural as an interface.
I do expect we'll see, you know gesture, eye gaze these kinds of much more multimodal ways to interact with the AI, rather than just stick in that prompt box.
But language is incredibly powerful.
So if you think of prompt as being more language, as a way to express ideas and communicate with a machine, that's a powerful paradigm.
As humans, so much of our communication is based on language.
I don't think we're going to move away from that.
It encodes information.
You know, language, words are symbols that encode so much information so efficiently.
And so I don't think we're close to getting away from that.
When you think about what you did, believe that you now have changed your mind on what's most prescient.
Oh, I'm a scientist that is happy to be proven wrong anytime as long as there's new evidence.
I'm genuinely curious to know.
Other scientists are much more like holding on to very, very strong conviction.
I have weak conviction but very strong respect for the scientific method and rigor, experimental rigor, you know theoretical rigor as well.
I used to be quite skeptical that neural networks were necessarily the ultimate solution to machine learning.
I'd seen enough cycles of neural networks kind of peaking and then being less useful.
And I used to think every time you change the scale of the data, you know you go from hundreds of examples to thousands, thousands to hundreds of thousands, to millions of examples.
Every time you change the size paradigm.
That neural networks were the first thing we tried, because they're a universal function approximator, and then something else comes out that was better.
And that was true for the previous generations.
You know, some of You may remember SVMs as like being better than the neural networks in the early 2000s.
They seem to be quite wrong on this one, like neural nets seem to be here to stay.
And the ability to do back propagation and gradient descent and all that seems to be a really powerful way to learn.
What does everyone else believe quite strongly that you think they are quite wrong on?
I don't have a lot of patience as a scientist for people who are predicting sort of the extremist scenarios, whether it's the catastrophic risks of AI or whether it's the you know winner takes all you know, AI becomes our overlord kind of scenario.
I don't have a lot of patience for that.
I wouldn't say it's necessarily widespread, but I just you lack patience, scientific rigor to analyze these kinds of scenario.
I'm much more pragmatic, grounded.
I'm pro-innovation.
I'm excited to see where AI is going and the problems it can solve.
But I'm not so interested in just going around and, you know, making up science fiction scenarios.
You've been on the most incredible you said there about image generation in 2015 and how much it's improved.
We're seeing this kind of unbelievable capital supply go into the space in a way that we haven't seen obviously, for many, many years.
Is it a good bubble where we are getting incredible improvements and it's fundamentally advancing technology.
Or is it a bad bubble where costs are becoming too exorbitant?
Teams are too impossible, to build computers too difficult.
Is it a good bubble or a bad bubble? about it as a bubble with bigger variance.
It's like you know the upswing is going to be bigger and you know there's going to be big downswings as well.
And so there's a lot of variance into the system right now.
As long as people have a tolerance to risk,
Then I think AI is a great investment and we should continue to be supporting.
You know, risk taking, new enterprise, new ideas.
There's a ton of exciting new startups being created.
We should continue to support them.
You just have to be tolerant to risk.
I've had some people on the show suggest that evals are, to put it delicately, bullshit.
And that they don't actually mean anything anymore.
And like, you know, humanity's last test.
Like, what does that really even mean?
And we have these new tests that come up.
What is this?
And leaderboards, what is this?
Is that fair?
Or do you think they actually serve a very effective utility to the ecosystem?
I do think they are really good indicators.
So I think you do need to take evaluation seriously in terms of knowledge, but you shouldn't take them seriously in terms of the ultimate goals.
So you know evaluation and there's lots of different benchmarks and so on.
You have to decide like what type of models or model are you building, what's the characteristics of your system, and then think of evaluations as like unit test for the performance of your system and software engineers will know what that is right.
Like you run through that evaluation and that gives you like, a signal of how the system is doing in a particular dimension.
But as we're building systems that are more and more general, do very specific tasks, you don't optimize for these.
We build AI systems that go into enterprise.
None of our clients ask about like.
Are you able to win the Math Olympiad with this model?
That's not what they care about.
They care about bringing value to their business.
Now we're curious to know how well we do on math problems, because it can be predictive of behavior on other things.
But you don't obsess over specific benchmarks.
You kind of look at the ROI in terms of what you're trying to build.
We mentioned there about kind of access for enterprises.
Enterprises have money and that's a great luxury in a lot of cases.
Research institutes, universities often don't.
With the kind of bubble-like tendencies, people with money are able to afford to compute the talent.
Are we seeing this kind of lack of access or democratization for great institutions that are educational, maybe who now can't afford to compete in this new world?
Certainly, a lot of universities have a lot less resources than companies today.
That's not completely new.
When I joined Meta in 2017.
One of the reasons I did that is because I could already see the disparity in terms of access to compute.
And I was really curious to see how you could do research with a lot more compute.
But there's still amazing research that's being done in universities today.
You go to the major international conferences NeurIPS, ICML and others and often the best paper awards are actually won by researchers out of universities.
There's a lot of good ideas that you need to test out at small scale.
And in a university, you have a lot more freedom to pick pretty risky ideas at a small scale.
But still, you know, no one's asking you to test to justify your research in ways that often happens in companies.
So I think they play different roles in the ecosystem.
And what's actually especially good is talent flows between them.
University students come in, do internships, take jobs at companies.
We've also seen a movement of people coming out of these large companies, going back to university teaching, sharing with the next generation what they've learned.
How important is it to have seen success and how valuable that makes you?
When you look at people like Amira, raising 2 billion at 10 billion, it's like well, no one's seen the success that she's seen with OpenAI, so it's valid.
To help me out as an investor.
Is it that valid to place that much of a premium on access to seeing it at that level that we are?
Or is that slightly overpricing it?
I think in many cases you know when it comes to deciding where to invest very early on, when you don't have tangible information.
I mean, you look at people's track record and there's a part of that that may be like.
You know, what have they learned in terms of the core recipe right?
But the other thing is also the achievement of having put together amazing teams who are building world class models.
And there's a lot of subtlety to that.
And so I think, you know, both of these ingredients are important to consider.
If you were investing today and you were joining my team, which category would you most like to invest in?
Be it security, be it generative AI, compliance, you name it.
Yeah.
There's a lot of verticals whether healthcare, scientific discovery that I think have incredible promise, where we're going to see real tangible progress within five years, that are going to change completely the face of what we can do.
That's probably where I'd push.
That's very exciting on the healthcare front in particular.
When you think about that timeline as well.
I'd love to do a quick fire round with you, if that's okay.
So I'll say a short statement.
What would you most like to do, but because of technical or financial limitations you're not able to.
I'm super keen to figure out how we build societies of AI agents.
We're doing it implicitly, but how do we look at populations of AI agents interacting together and having like a sandbox for doing that?
Maybe something I'll get to do.
Is it lack of time, resources, something else?
There's just like a ton of different things to do, but keen to see what happens there.
When you think about that ecosystem of agents, you have children.
Yes.
And AI changes our relationship with other humans and friendship and social.
How does AI impact social friendship connection?
It definitely does.
And, you know, there's a sense that we spend a lot of our time in the digital world.
And for some folks, you know I look two of my children.
They spend a lot of time in the digital world playing online games with their friends.
It's still very social.
It's, there must be some AI, there's the digital platform, but it's still a very social experience.
Others have more individual experience.
There's definitely a shift of this time we spent towards that platform where we go look for that social element.
Knowing what you know, what do you not let your children do?
Ah, eat too much sugar.
Totally.
So that's that like physical diet.
Completely agree with that.
Yeah.
Is there a technical diet?
I spend some time discussing like settings.
I mean, like you get an Instagram account.
Great.
You can have an Instagram account.
But like, what are the settings on that account?
Making sure they understand.
I mean, they'll go and change them if they want.
We're going to discuss settings.
I know that was not a popular one.
Did they listen?
The thing with children is you don't know till later.
Do you limit screen time?
I spent a lot of energy, especially in their younger years, limiting screen times.
My kids did not have a cell phone until they were 14, 15.
Did you see adolescence?
I have not.
Okay.
Watch it.
Basically a little boy goes up to his bedroom and gets lost down kind of rabbit holes of TikTok and Reddit and it does not turn out well.
I've heard about it.
I just haven't had time to sit down and watch it.
Do you worry about the loneliness pandemic and then also just the mental health crisis that we have?
I do worry a lot in general about making sure that people are healthy mentally and We have to be careful about taking shortcuts and saying you know, because suddenly you know we have certain platforms, we have AI and so on.
That is causing that mental illness.
You know, there are a number of people who are suffering and they deserve to have good answers to the situation.
And we deserve to find real solutions to that.
And there's a lot of people looking for shortcuts and short answers.
But I think more research into that is definitely warranted.
What's your biggest lesson from working with Zuck?
He is incredibly deep into understanding the work.
Like he does not coast.
You know, when he started getting into AI, just the depth of the question that he'd ask.
He just gets really interested in the topic and goes super deep.
And that then just informs everything he does after.
So you can have the most amazing team but, like as a leader, you need to go deep and understand the work.
Did you see him change?
As anyone gets more knowledgeable about a topic, you get more decisive.
There's a phase where you're really learning and trying to understand, and there's a phase where you understand a lot of things and then you make your decisions faster.
So certainly that shift happened.
What one AI buzzword would you ban if you had a magic wand?
Existential risk.
Why?
Because it just makes people afraid.
And it's not out of fear that we make our best work and we take good decisions.
Do you find the cost of talent exorbitant?
Talent is costly.
Talented people deserve to be paid well.
And again, you know, someone coming in just because of money rarely is going to be the right person, but you do need to compensate people fairly.
Final one.
What are you most excited for?
You don't like the existential risk.
I don't like the doomsday planning.
When you think about the positivity that can come.
What are you most excited for when you look forward to the next three to five years?
I do think some of the work in terms of AI for a scientific discovery is going to be pretty fascinating to see.
Just in terms of the doors, it's going to open up the ability to explore combinatorial space of solutions.
I'm curious about that.
And then I'm super curious to see how can we actually make our models more efficient?
There's larger and larger and larger models.
No one wants to run these models.
I spent a lot of my career building open source models.
And I'll give you one example.
You know we were in the frenzy of large language models and I pulled the stats on.
You know, most downloaded models of last month.
We had a model like Roberta from 2019 small language model was getting 20 million downloads a month.
People want efficient models that they can use, that they can run.
So I'm also super keen to see what we're going to be able to do at the scale that runs on like one or two GPUs.
Final, final one.
You said there about kind of open.
We seem to like be reverting to a closed world now.
Is that the world with which we should predict and plan on?
That's a deep mistake.
I mean, I will continue to believe that, especially for research, the ideas need to circulate.
And this thought that you can just like close this down is absolutely false.
I mean, people are circulated.
Do you not think we are moving into that world?
Everyone seems to be closing systems, closing access.
There are definitely a number of places, people that are closing down access.
I don't think that is going to be effective.
Ideas will circulate.
And I also think it's a mistake from a point of view of fostering innovation.
This has been such a joy.
I've learned so much from this conversation.
Thank you so much for putting up with my very basic questions.
But I've loved having you on the show.
My pleasure.
Thank you.
But before we leave you today, are you drowning in AI tools?
Chat, GPT for writing, Notion for docs, Gmail for email, Slack for comms, and you're constantly copy pasting between them all, losing context and losing time.
This is the AI productivity tax and it's killing your output.
At 20VC, we're all about speed of execution, and Superhuman is the AI productivity suite that gives you superpowers everywhere you work.
With the intelligence of Grammarly Mail and Coda built in, you can get things done faster and collaborate seamlessly.
Finally, AI that works where you work, however you work.
Superhuman gets you from day one, with zero learning curve, and it's personalized to sound like you at your best, not like everyone else using generic AI.
Get AI that works where you work.
Unlock your superhuman potential.
Learn more at superhuman.com forward slash podcast.
That's superhuman.com slash podcast.
And once you're moving faster with Superhuman, make sure you're moving safely with Vanta.
Customer trust can make or break your business.
And the more your business grows, the more complex your security and compliance tools get.
It can turn into chaos and chaos isn't a security strategy.
That's where Vanta comes in.
Think of Vanta as you're always on AI powered security expert who scales with you.
Vanta automates compliance, continuously monitors your controls and gives you a single source of truth for compliance and risk.
So, whether you're a fast-growing startup like Cursor or an enterprise like Snowflake, Vanta fits easily into your existing workflows so you can keep growing a company your customers can trust.
My listeners can get $1,000 off Vanta by going to vanta.com forward slash 20VC.
That's V-A-N-T-A dot com forward slash 20VC, 20VC for $1,000 off Vanta.
