# Democratizing Self-Accelerating AI for Enterprise R&D

**Podcast:** AI + a16z
**Published:** 2026-06-24

## Transcript

With an AI technology that is helping with AI research, what is the role of everyone else in the company?
This is a very disruptive technology.
With disruptive technologies, you need to rethink a lot of pieces to kind of enable them to actually grow and flourish.
And some of these pieces are how you build a company around it.
If the business model of the company is, I train a big model and charge people for using it, how is this company incentivized to...
share this technology with everyone else.
We've been at the Frontier Labs for a long time and we know how much work it is to do something.
And we've been able to do it in like maybe 10 times less people and less resources.
The jump from Sonnet 3.5, 4.0, 4.5 has been materially better in terms of like, where does it get stuck?
Where does it need oversight?
Even the tiniest reduction in oversight can lead to like large amounts of token spent and just effective outcomes being generated.
What you guys are doing is almost the next level of cheat code.
When the internet gets faster, you don't recursively get faster internet.
Where do you think it all ends up?
It's a wide spectrum of outcomes.
One of the oldest ideas in AI is also one of the most ambitious.
The idea that intelligent systems could help improve themselves.
For decades, that possibility remained largely theoretical.
But as AI systems become more capable at coding, research, reasoning, and engineering, the question is becoming increasingly practical.
What would it mean to build AI systems that contribute to their own development?
And if that becomes possible, where should that capability be directed?
Matt Bornstein speaks with Mirandil co-founders Benham Neshabor and Harsh Mehta about self-accelerating AI, scientific discovery.
and why they believe the most important applications of AI may be in advancing science itself.
Benham, Harsh, welcome to the A16Z podcast.
It's awesome to have you here.
Benham and Harsh were most recently research scientists at Anthropic, worked together before at Google at Blue Shift Labs, which Benham, you are a founding member of.
Thrilled to have you here.
Tell us a little bit about Mirandil, what you're aiming to accomplish, and tell us where the company came from.
A lot of reasons behind building the company comes back from when...
Scaling law was happening at OpenAI and back then I was at BlueShift team and that was the moment when I realized we are on a verge of a revolution and then everything is about to change.
And for me, the most important thing was what are the technologies that would accelerate all areas of science and what are the main bottlenecks for making that happen.
And since then, I've been kind of trying to close, remain close to that path and remain on the shortest path.
and known harsh since then and kind of trying to be on that path.
And more recently, as we've been in Entropic, we've talked a lot about what does it mean to kind of remain on this shortest path.
And we felt like the main labs are starting to diverge a little bit from what does it take to accelerate all the areas of science that led to studying Merndl.
We think the self-excelerating AI is a technology that is disruptive.
And at the same time, it's important for...
accelerating science and technology, and we want to focus on building that and making it available.
So science is notoriously hard, right?
I mean, science is science.
It's an experimental discipline where you have to run experiments in the real world to understand the laws of physics or biology or so forth.
Can you guys describe a little bit what do you think is the shortest path for AI to sort of accelerate science and why?
So from the last five years, when the models kind of like started kind of like getting better at some of the primitives of conducting science, like high school math and college math, and then a little bit of coding, and then now the coding models are really good.
Computer use just started to really...
And is this all in scope when you say science, by the way?
I just want to make sure.
Are we talking about strictly like physical, you know, natural sciences, or is it a broader set of things?
So ultimately, what we want to build is like AI systems, which in a very broad sense can conduct AI research and engineering itself.
And if it has...
this capability in a very broad sense.
It has the right primitives and the capabilities which are needed to conduct any science which is in the realm of digital world or any science which is in the physical world but then has a digital component as well.
And so what do you think is sort of the path and maybe contrast to what the major labs are doing because people may be familiar with that at some point.
Yeah, so when you think about what is a scientist or what does a scientist or an engineer do, the main skill they have is get deep.
very deep in a domain and build expertise and they build a very sharp expertise that accumulates over time.
And as you kind of become expert in a domain, your expertise is becoming sharper and sharper in something that has a tiny bit of volume in the whole set of possible things you could be working on.
And getting to that point, like being able to get there is a capability that is needed.
to advance all areas of science?
What does it take to have an AI system that is able to be directed at a direction and make improvements faster and faster and gets eventually to the edge of that area and eventually make progress?
And what that view means is that you have to work on what does it take to direct a system toward like a particular angle and make fast improvements.
And that's the self-accelerating AI technology.
So this in principle is very disruptive.
Because this technology allows, you know, like a model to kind of keep making improvements and keep getting better at that direction.
And as you know, like when you think about the business model, you can build around this, etc.
It's just a very different world where you want to make this available to everyone as opposed to make AI available to everyone.
Yeah, so explain this term self-accelerating AI, unlike a huge sci-fi.
fan.
So I've always hoped that, you know, the machines would start improving themselves and take over the universe.
But that doesn't seem to be happening, you know, and I'm not sure if it really is possible or makes sense.
So explain what you think of, like, how is this actually going to happen?
And so what does self-accelerating AI really mean?
I think I have many different definitions and we have a specific version of it, which we are targeting.
If you look at the history of our field, in some sense, AlphaGo, where we can create like all these self-play loops, is already self-accelerating AI.
Another version of it is if we can put AI in an unknown domain or an environment, can it learn by itself and improve its own capabilities and context?
That is another version of self-explanatory.
And this is sort of like the ArcGIS benchmark, for instance, where it sort of dropped into an environment and has to understand the rules before optimizing them.
Correct.
And more...
Like inside like enterprises or general knowledge work, another example could be like a system which ends inside a company.
If an employee kind of like enters a company, then they have to gain all this context to function well.
And can an AI system do something like that?
Our version of self-excelerating AI is can these systems do the work of an AI researcher or an AI engineer?
Like everything from the low-level kernels to like works to conducting research, it's often a very high throughput.
And we think that that's kind of like a sharp version of self-excelating AI will lead to a broad version of self-excelating AI, which kind of ultimately has the capability to conduct kind of like advanced science in a more broader way.
And so it sounds like part of what you're saying is that scientists kind of have to become AI people in a way, and they're not right going to like do it on their own, but like the tool that you're building for them can help them actually sort of become AI people.
Yeah, so one way we like to think about this is if you think about each areas, each of these important problems that exist in science and technology, and what does it take to solve these problems?
These are grand challenges that they're not human level problem.
These are superhuman level problems, things that we don't think people can solve.
And we don't think it's enough to build an AI scientist that would do the job of a scientist.
These are superhuman problems.
So what does it take to solve these problems?
Each of these problems need a periodic lab, which is a bunch of experts who understand the domain and are excited about kind of attacking this problem directly.
And then a whole big AI effort just to push AI for that particular problem.
And the thing that's difficult to do for a periodic lab is assemble the best frontier AI team and kind of iterate over the AI solution that's best for them.
And what we want to do is take that part and minimize it, make it small so that you can.
Maybe today, if you need 200 people to have a good frontier lab and 200 of the best people, which is almost impossible to do, we want to reduce it to 10, 2, and eventually 1.
And eventually you wouldn't need to hire on the AI side so that you can move much faster when it comes to solving a problem.
Because each of these problems have a lot of other constraints to solve, like physical elements, like getting access to data and all the other autonics that exist.
So we just want to remove the AI bottleneck.
And this has been sort of a longstanding goal in the machine learning community.
You know, I remember even 10 years ago, Google had a product.
It was sort of like a neural architecture search thing, but they sort of claimed, like, machine learning is now solved, right?
You just run the computer and it finds the right architecture for you.
What's changed now, or what's the state of the industry on self-acceleration now?
I think that's a great example.
And in some sense, it kind of like boils down to the intelligence of the agent of the system, which is conducting this search.
And as you move the intelligence dial higher and higher, you are much more compute efficient in getting the results that you want.
And that has been happening from that particular point to kind of like where we are today.
And it's also happening in a very broad way.
It's not just architecture search that you can do, but much more large swaths of software that you can improve on its own.
And it's a very expressive space of things that you can do right now.
Can you explain this point about general versus narrow?
And I'm guessing you don't want to reveal all the secrets of the models you're training, but I think it's a really important question.
Like, will general language models or general coding models, like, be able to run these sorts of AI experiments or is something more focused necessary?
So I think the importance about kind of like, you know, the general kind of like, you know, sense of and the taste of like these AI models is like, if you make like if you kind of like turn down the intelligence dial and like make them focused on like, you know, something very specific, then they would kind of like, you know, still make very like, you know, dumb mistakes, like, you know, not wire things in the code properly or like, you know, they wouldn't have the right kind of mindset of an expert in that particular field.
The jump from like, you know, Sonnet 3.5, 4.5, that has been kind of like materially better in terms of like how long can we run an AI system on a particular problem or where does it get stuck?
Where does it need oversight?
And every reduction and like, you know, even the tiniest reduction in oversight can lead to like large amounts of token spent and like, you know, just effective outcomes being generated.
And so you were actually at Anthropic when a bunch of these models released that you just mentioned.
And I know you were starting to work on some of these.
ideas then.
Can you just explain a little bit of the history behind this project you're working on now?
Yeah, so Benham and I have been excited about this grand challenge that he mentioned about accelerating science with AI.
And this was five years ago.
And we have been at it for a while.
One of these overnight breakthroughs that takes five years to develop.
Yeah, exactly.
We had to start with pretty bad models, models which are not good at even high school math or autocomplete coding and stuff like that.
So we first built the math specialized models at Gemini and worked on the reasoning models.
And then some of the primitives kind of started to come together when we joined Anthropic.
And that felt like the right moment where we can actually be ambitious enough to directly attack the problem of AI being able to conduct its own engineering and research.
And the biggest jump is kind of like, you know, like, you know, Dario calls it like a smooth exponential, like, you know, and scale has been very helpful, both in terms of the models, the underlying models being better, but also conducting the inference time kind of like, you know.
research and engineering at a very high throughput rate.
I'll embarrass you a little bit, Harsh.
You told me when you started working on this at Anthropic, nobody else wanted to work on it with you.
Is that true?
I think they had good reasons.
Meaning the models didn't work yet?
Yeah.
So like, you know, when we joined, it was like...
It feels a long time ago, but like, you know, the state of the art model was Sonnet 3.7.
And we are, I think I was too excited in some sense to kind of like jump in and attack this problem.
And one of the reasons why I joined Anthropic was like, you know, Dario had this essay, like Machines of Loving Grace, which was very inspiring.
And it really felt like the place where this kind of ambitious work can be done.
That's why I felt the conviction to start there.
It is funny.
It's almost easier to use model version numbers as time reference points now, rather than years and months.
I can actually understand what you're indicating when you say it's not a 3.7 versus it's like 20, 22.
I'm like, hmm, what's going on?
Yeah, no, and I think a lot of people feel that way.
I guess...
What gave you guys the sort of impetus to go out and, you know, start Mirandil, you know, having been in that position and, you know, having done some interesting work in Anthropic?
I think, so I was, you know, like, co-leading the science team in Anthropic.
Harsh was kind of initiated that led the automated pre-training project in Anthropic.
And we were both, like, talking about, like, what does it take to...
kind of apply this to science and accelerate science.
And one of the, there are a few observations that kind of made us decide to, you know, like take this challenge.
One is, this is a very disruptive technology.
It, with disruptive technologies, you need to rethink a lot of pieces to kind of enable them to actually grow and flourish.
And some of these pieces are how you build a company around it.
You can't let, you know, it's very hard to have a disruptive technology show up in an existing company and flourish because a lot of elements have to be redefined.
Like with an AI technology that is helping with AI research, what is the role of everyone else in the company and how do they rethink?
And it's very hard if the culture is not built for it, it's going to be harder.
The incentive of the company, the business model of the company, if the business model of the company is I train a big model and charge people for using it, how is this company incentivized to share this technology with everyone else?
Because that's directly letting everyone else train a model which reduces their dependency to the company.
So these are like so fundamental that it doesn't matter who runs the company, but you can clearly see that it's in...
contrast to what you want to happen.
And those observations and seeing that these are like practical limitations would lead us to kind of start a company that is really rethinking all pieces for this technology and making it available to accelerate science.
So just because you mentioned it, I have to ask about the Fable launch, right?
Where it's, you know, we've guard railed certain important safety you know, areas like bioweapons and, you know, et cetera, and AI research, right?
Like one of these things seems a lot less dangerous than the other things.
Do you think that, you know, like access to sort of cutting edge AI research assistance is going to be sort of curtailed or like inaccessible to most people if, you know, you guys don't succeed or, you know, companies like Mirandil don't succeed?
Yeah, I think...
I think there are good reasons for why you want to be careful about the technology, and it's very powerful.
Can you actually sketch out the pro case?
So for me, as sort of a non-expert, to see, it's like, okay, I understand why bioweapons are dangerous, why drug synthesis is dangerous.
There's a bunch of things that actually should be filtered out of pools of information generally.
The AI training one stood out to me.
It's like, well, this actually seems like a positive good.
But since you guys were sort of on the inside of it, maybe sketch out the case for why that does make sense.
So there are a few cases.
One is if you think about, let's say, speaking of entropic, if you're concerned about other states developing and competing with U.S.
then you would be concerned about them using your technology to move faster.
And I think that has been like a concern from Antropic.
And the other concern is like, let's say for bioweapons, like if you're concerned about bioweapons, then you'd be worried about someone using this technology to build a model that would help them with bioweapons.
So it's very powerful.
It can be applied to anything.
So there's, you know, like some merit in being worried about this.
At the same time, the same way as AI came to existence and it was a powerful technology and you have to have considerations around it, but you can't limit the whole technology because it's very useful.
It's true for self-accelerating.
It's powerful.
It's very enabling.
And what it really requires is rethinking how you should do safety, how you should...
Think about Godrails, when you build a company around it, you would be focused on how to get this right, as opposed to when you're trying to do a lot of things, then you would be looking at this and I'm like, okay, this is too much trouble for me.
And also, it's very hard to separate incentives from reasons.
And it's just hard to separate it because you're also deeply incentivized to...
to keep your distance with others, et cetera.
And it doesn't matter who runs the company.
It's just like when the incentives are set this way, it's hard to fix it.
No, that's fair.
And suffice it, I guess I'm inferring from what you're saying that both the incentives and sort of the value structure of Mirandil are designed to enable access to these sorts of technologies rather than curtail it.
Yeah, there's a...
kind of like, you know, very large benefit that we can get from offering these, these, this technology.
And there's also harm.
And some companies would prefer to kind of like, you know, throw a very large hammer that, okay, we block access to everything.
But because of our focus and our kind of like, you know, the intent to advance science, and we want to take like a sharper approach where, where we.
put the energy of like, you know, making sure that in every use case that we're offering this technology, we can make sure that it's kind of like, you know, being used in the most positive sum and safe case.
That makes a lot of sense.
Do you want to talk a little bit about the products and models that you plan to launch?
Again, you know, don't need to reveal anything super secret because I know you haven't launched yet, but it's a great chance just to, you know, tell people they should look forward to.
Yeah, I think in a...
In a broad sense, models and product look very similar to existing state-of-the-art models, but sharper in the specific capabilities which are needed to conduct engineering inside an AI lab or anyone who is really trying to train or serve their models.
So the hope would be that it would be really good at kind of like, you know, everything from low-level kernels to, like, you know, libraries like PyTorchacks, RL frameworks, prefending frameworks.
And the product will kind of, like, amplify these capabilities to make it very, very accessible for anyone who is doing this kind of work and, like, specialize towards that.
I want to add maybe one other point related to kind of...
our approach and how we are thinking about incentives, et cetera.
So the way we are thinking about the product and how we are forming relationship with the rest of the world is how can we enable businesses to start owning more pieces to have their own infra, their own AI.
And I think what we are noticing today is that as a result of creating more dependency to these big AI labs.
They're gradually losing control and losing bigger parts of the business and then it becoming weaker and weaker.
What we want to do is kind of reverse this so that every business, every lab would have their own AI optimized for their own workflow and with their own data.
with their own infra, and this would allow them to have better margins, more control, and kind of this is like a much more enabling future.
And we think it's going to be just a matter of time.
I think already businesses have started being worried about this, and it's going to be just a matter of time before they want to lean into this angle more.
Yeah, I mean, that I think is a really interesting question.
It's almost the most important question.
right now, like, if you think about major technical breakthroughs that have, like, massive economic impact, you know, one characteristic of that is that everybody can use them, right?
Everybody can build on them and build with them, not just consume them.
You know, you have some monopoly industries like power generation where the CapEx is so high that everybody can, but, you know, it's like consuming power is, like, not very limited, right?
This is like...
really low level infrastructure versus AI is not, right?
AI is like application level all the way sort of down.
So this is the thing that we think about a lot on our team.
It's like, how do we get to a point where all the major companies, startups, you know, big banks, healthcare companies, hardware companies, et cetera, can actually build with this?
And, you know, I guess the hope is that something like what you guys are doing sort of can help with that.
Yeah, I think it's interesting because, you know, like let's say with Cloud Code, coding went from just being a thing related to programmers to like something that everyone can do.
And now everyone is enabled to do a lot more.
And the future we see is same happening to AI where everyone can go from their wish to have AI do something for them to seeing that happen for like their own AI.
It's just a matter of having the compute and resources and not matter of like expertise, et cetera.
So businesses, if they want to kind of take an angle, they have, you know, like their own data, they have their own processes.
This should give them naturally a lot of advantage.
And the thing that they're missing is putting everything together to build the moats.
And, you know, we think AI is going to be that enabling thing that would kind of give the control back to them.
Is it true that all researchers at Miradil have to...
submit their traces and chats to model training.
You don't have to answer that if you don't want to.
All I can say is that when you start a company with a certain goal, you optimize the entire company end to end for what you're going for.
Maybe sort of on a related note, like are there any moments that you guys have seen either in the last few months at Mirandil or just in general that kind of gave you the...
chills that like, oh my God, like self-accelerating is actually working.
I think the kind of like the rate at which we have made progress, technical progress, I think that surprised me as well.
I think we've been able to do a lot with a lot fewer resources and people.
And like we've been at the Frontier Labs for a long time and we know how much.
work it is to do something.
And we've been able to do it in kind of like, like maybe 10 times less people and like, you know, less resources.
That was surprising to me.
I also, maybe a surprise for me is when I talk to candidates and they get surprised that, you know, like, how do you think you can compete with big places with, you know, like only having, you know, like, I don't know, 20 people.
And I'm like, Looks like you're not a believer in this.
Have you been using?
More should say that.
We have this thing now that can help us.
You know, also start, right?
Like, you know, in many ways, startups are just a more efficient way to allocate, you know, resources and people, right?
You know, compared to sort of diminishing returns at big companies and things like that.
I guess part of what I'm getting at is like, we have this term self-accelerating.
You know.
Do you believe in the kind of like real like self-acceleration sort of recursive thing or like to you is this a tool that is just going to like bring prosperity, you know, and like these great capabilities to everybody?
I think a lot of people have a sci-fi view of it, which is like a model kind of making changes, etc.
I think one thing that maybe we have been kind of thinking slightly differently is building AI systems as opposed to like thinking about one AI model.
This is where like...
you know, ecosystem of models and ways of working with each other.
And I think we think for foreseeable future, maybe, you know, like, I don't know, maybe a few years, this ecosystem also includes human.
And then you're thinking about this entire system as one intelligent being.
And you're asking this question of how can this system improve itself?
And so that's like a gradual view of like, how do we get there?
But also systems are much stronger than individual models.
Like if you think about, you know, like best AI researcher in the field, what can that one person accomplish versus an entire company that's built?
But it's also not obvious how to build an entire company from individual brilliant people.
It's actually a hard thing to do to join like 20 people and then make a company out of it.
Poor AIs are going to have to figure out org charts.
Yeah, so it's actually not a straightforward problem how to create a system that is like fast moving and makes improvement.
So the way you think about it is like there's a system of...
AIs that are not necessarily the same model.
It could be a specialized, could be a bunch of them are generally smalls and then some people.
And then this system wants to kind of keep getting better, which getting better means this system kind of develops the next AIs that would be part of the next system and, you know, kind of keep making improvement.
And that's like a very realistic view of how, you know, it would evolve into at some point becoming fully autonomous.
So you're saying in the future an agent will leave its 10 million agents swarm and talk to, you know, the 1,000 agents swarm and say, how do you guys get any work done with all the 1,000 agents?
No, I think that's, yeah, no, I think it's a very interesting point.
You're almost saying that, because one issue I've always had with this is somebody has to prompt these things.
They're not self-directed, but you're sort of saying as the level of abstraction goes up, there's almost like a standing prompt.
You only need one prompt anymore, which is make improvements to yourself, and this sufficiently large and complex system can kind of operate continually.
It's almost like prime directive rather than prompting at that point.
Yeah, exactly.
And if you look at kind of like how the models have been developing, like it requires less and less oversight.
And I think this trend is just going to continue.
And at the end, the ultimate prompt is like, you know, to achieve goals.
And the system should be able to kind of like learn on its own, figure out all the problems, ask for directions when needed, just as humans would, and achieve those goals.
I mean, related to the systems, I wanted to kind of throw this out as like an interesting thought experiment is, you know, when you think about systems, there's this other angle of scaling up, which, you know, we have so far scaled up models in terms of size and compute, et cetera.
Now we are facing this another axis, which is a scaling up systems.
And these systems today, they're like systems composed of people and agents.
So X-axis is number of agents, like some of them are people and some of them are models.
And, you know, they have different properties.
But the question is really, how do you get a favorable scaling of a system?
Like companies scale and their productivity go down with scale.
So people don't have a favorable scaling.
So you go from, your company size goes 10x and your productivity is like 1.2.
So that's not great.
And today, agents are also not that amazing in terms of like being able to actually scale them in a way that you can see that productivity of the entire system grows.
So the companies of the future where you would want to be able to kind of throw compute at problems and kind of scale up agents and see that you, let's say when you double number of agents, you get to the same goal two weeks faster.
That's the question.
Like how do you save time?
The real thing that the world is fighting for is how to get to a point faster and all the competition between companies.
And a lot of these companies are willing to pay 10x more compute to just get to something like one month faster than others.
And I think you have to solve the scaling problem.
And one interesting thing that we know about scaling laws is that if you want to get scaling laws right, you have to start small and get the scaling right before you scale up.
Otherwise, things are not going to work.
So that's, I think of ourselves as like the experiment.
You're like running an experiment at a small scale, thinking about like, how can we get favorable scaling and then scale up the system?
I mean, it's really interesting.
Is it a communication?
Like what aspects of the system do you think need to be solved in order for this to work better?
Oversight, a lot of oversight issues.
I think a lot of resource allocation problems.
Like if you have the same compute, but you increase the number of researchers, you can't, you know, how do you decide which ideas would get prioritized, etc.
So a lot of technical...
This is what really happens in human researchers.
Yeah, also with AIs.
Yeah, yeah, yeah.
So there's a lot of technical problems to solve there that are very interesting.
So now we're talking interagent politics.
It's like, why did that agent get 1,000 views?
At the end of the day, with agents and with people, incentives are very important.
And this is orthogonal to the performance of the agent itself.
Because I know you've worked a lot on long-running agents and, you know, like Harshi was saying, sort of narrow slices of problems.
But these things you're talking about are actually orthogonal to its individual agent performance.
Yeah.
Well, one thing I want to highlight is such a very interesting and impactful time that we're living in.
Like the histories of science and, for example, physics, a whole bunch of really, really impactful, work happened in 1920s and 30s.
And then there was a little bit of a blank space.
And I think we're going through a similar thing in intelligence.
And it's going to be improving at a very fast rate.
And it's just a really interesting time to live in.
Do you think this sort of self-acceleration is already starting?
to happen?
Meaning when you look at like the latest releases from the big labs, you know, is the slope getting steeper because of this?
Absolutely.
I think it has started kind of like, you know, in some sense and small ways when the coding models kind of like, you know, became like truly useful and it has picked up by kind of like every generation of kind of like new models and it improves everyone's productivity by itself, but that's also kind of like improves the time to the next breakthrough, next model.
This is sort of a unique aspect of AI, isn't it?
That it's like you have these sort of cheat codes.
It's like coding is kind of a cheat code because a lot of the system is built on code.
What you guys are doing is almost the next level of cheat code because like the actual experiments are sort of, you know, which you don't see with like, like when the internet gets faster.
You don't recursively get faster internet.
So it's sort of an interesting thing.
Where do you think it all ends up?
It's a wide spectrum of outcomes.
But our preferred outcome is just generally prosperity, specifically scientific prosperity, in the sense that a lot of the science is just...
unknown at the moment.
And a lot of the engineering kind of like, you know, large engineering projects are also just bottlenecked by people being able to design things well and just the intelligence being the bottleneck.
And we hope that like some of the progress can translate into just us knowing a lot, lot more about ourselves and the universe.
Your question is like maybe also one of the other reasons where it just kind of pushed me to.
kind of start a company is I think the current picture that's being painted about the impact of AI is not that positive, like automating people's job away.
And, you know, like that doesn't seem like an exciting future.
And so, you know, I'm thinking you're trying to, in a way, focus on building a different one that has more positive outcome for everyone.
And we think accelerating science is pure good for humanity.
So that's where you could decide where to put, like you're building something powerful, you could decide where to put it at work.
And I think we should feel like we are in charge and we can kind of build the future we want.
I think these problems that, you know, like solving Alzheimer's disease, being able to predict it much in advance and solving it, these are like super challenging problems.
There are so many things that have to be solved on the path of it.
it would show everyone that AI, it should have been used this way from the beginning, but somehow we are kind of losing focus on like, these are the important problems that have to be solved and smart people should be working on these problems.
So we want to kind of change, at least we can choose what we work on and we want to kind of be focused on creating this powerful technology and directed at these problems that are longstanding.
And it would still take a lot of resources.
It would create jobs because, you know, like you have to solve all other aspects of it that are not intelligence related.
Bottlenecks are going to move away from intelligence.
So it's interesting to kind of direct it at these problems that are just not about day to days, but solving problems that help everyone.
And is it fair to say that...
continuing to scale up pre-training is probably not sufficient to solve these kinds of problems you're talking about.
Because you're almost talking about generating net new knowledge, which I guess is hard no matter how much pre-training you do.
Yeah, in some sense, I think of pre-training, post-training, and all the paradigms that we have of using compute effectively as tools in my toolbox.
And at certain points in time, like, you know, some tool is effective.
And like, you know, we go for the fixed amount of compute that we have.
We go for the best tool.
But what you're trying to build is like, is not even just like adding compute, but kind of the system of putting like expert people into like the system as a whole.
And this is like part of what Amox is, I think you're saying.
Yeah.
So, you know, like for, let's say, you know, do this minor experiment of like, the way things are going.
You know, our best hope is that we're going to have models that are going to be as good as our best scientists or maybe become better than the best scientists.
But how is that going to solve Alzheimer's disease for us?
Like, these are problems that are, you know, we don't even know if it's possible.
We don't even know what are the limits that exist.
So we really need to move at much higher speed.
And if you're not thinking that way, if you're just thinking about slightly higher...
building an AI scientist, et cetera.
This is not going to solve these problems.
Like Alzheimer's disease, for example, it has so much structure in terms of data that you have to use, et cetera, that you cannot even see how in 10 years existing models would be able to kind of move things much faster.
And with the current speed, we are not going to get there anytime soon.
So we are impatient about solving those problems.
Awesome.
Well, this was awesome, guys.
Thank you so much for working on this and for ever coming on the podcast.
Thanks for listening to this episode of the A16Z podcast.
If you liked this episode, be sure to like, comment, subscribe, leave us a rating or review and share it with your friends and family.
For more episodes, go to YouTube, Apple Podcasts and Spotify.
Follow us on X at A16Z and subscribe to our sub stack at A16Z.substack.com.
Thanks again for listening and I'll see you in the next episode.
As a reminder, the content here is for informational purposes only.
Please note that A16Z and its affiliates may also maintain investments in the companies discussed in this podcast.
For more details, including a link to our investments, please see a16z.com forward slash disclosures.