1. Podcasts
  2. Latent Space: The AI Engineer Podcast
  3. 🔬 Why AlphaFold Didn't Solve Protein Folding — Pushmeet Kohli, Google DeepMind & Sal Candido, Biohub
Latent Space: The AI Engineer Podcast

10 Oct 2026 · 32m · 12 chapters

Episode artwork

🔬 Why AlphaFold Didn't Solve Protein Folding — Pushmeet Kohli, Google DeepMind & Sal Candido, Biohub

In short

Why scaling data/compute didn’t “solve” protein folding, and what’s still missing for function, dynamics, and design; includes lessons from AlphaFold, protein language models, and the need for problem-first, multidisciplinary approaches.

Guests (backgrounds)

Pushmeet Kohli (Google DeepMind; ML researcher with experience across security, computer vision, Bayesian/discriminative deep learning, AI for coding, now science). Sal Candido (Biohub; ML/data-generation perspective; previously at DeepMind with Rich Sutton’s “bitter lesson” framing). Also referenced: John Jumper (AlphaFold co-developer).

Key claims

Don’t be “religious” about only modeling or only data—start from the problem and choose the needed solution space. AlphaFold matched PDB structures but doesn’t capture full dynamics/distributions. Scaling helps only when data quality/coverage and inductive bias are right; metagenomic protein-language-model training can improve performance despite noisy sequences.

Notable examples

Training protein language models on metagenomic sequences (often not full proteins) to design functional proteins; AlphaFold’s uncalibrated confidence/PLDDT; “virtual cell” ambitions blocked by missing cell-by-gene data.

Written by AI. May contain mistakes.

Chapters

0:15

The Bitter Lesson of Data in AI

Discussion on the importance of scaling data and the bitter lesson in AI.

“We are all modelers, so of course it's natural for us to talk about data.”
2:10

Quality vs. Quantity in Data Collection

Exploring the balance between working with available data versus ideal data.

“So when it comes to data collection, how do you think about what types of data modalities are best?”
4:50

Flexibility in Problem Solving

Emphasizing the need for flexibility in approaches to data and modeling.

“Yeah, so I think the, so I was at DeepMind when Rich was with us and working and thinking about this idea of the bitter lesson.”
8:23

The Art and Science of Model Building

Discussion on the balance of artisanal craft and scalable solutions in modeling.

“So what do you think, like, the actionable takeaway is from this?”
12:20

Current State of Protein Structure Prediction

Insights into the ongoing challenges in protein structure prediction despite advancements.

“Yeah, I think, I mean, I very much agree.”
14:01

Understanding Protein Structure Prediction

Explore the complexities of protein structure prediction and current challenges.

“Like, as more and more data comes in, it's not only helping to curate that data and select what is the next batch of information that we actually need.”
14:37

Progress and Misconceptions in Protein Folding

Discuss the progress made since AlphaFold and the misconceptions surrounding it.

“the news will say protein structure prediction has been solved.”
17:39

Challenges in Data and Model Development

Identify the biggest blockers in protein structure prediction and potential solutions.

“Or if there's one thing you could wave a magic wand and say, like, we have more of blank, that would accelerate one of those things, function, dynamics, or design, what would the magic wand be?”
19:16

The Role of Understanding in AI Models

Examine the importance of understanding models versus using them as black boxes.

“but they're not exactly the problem that most people want to solve.”
22:28

AI's Impact on Healthcare and Predictions

Discuss the potential of AI in healthcare and timelines for its clinical implementation.

“So I think, like, one thing that's, like, really salient to me about this is I think the models actually have a lot more information in them than we know.”
Show all 12 chapters
28:00

AI in Drug Discovery: Current Use and Future Potential

Explore how AI is currently integrated into drug discovery and the challenges ahead.

“I mean, the point is AI results in the clinic, it's ill-posed.”
29:44

The Path to Significant Improvements in Drug Development

Discuss the importance of basic understanding and ambitious goals in advancing drug development.

“Yeah, I mean, I'm tempted to like literally put on my bio hat to answer this question.”
A man walking through an airport, asking his phone a question

Ask about this episode

Get the gist, explore an idea or ask about something said.

Claude

On Latent Space: The AI Engineer Podcast, what are the key points of this episode?

Searching your listening

Transcript

Automatic transcript. May contain errors.

0:04Yeah, so great to be here. What an exciting morning. So many cool announcements. I think the future of bioscience is being announced right here. So this is the modeling session. We are all modelers, so of course it's natural for us to talk about data. So to start out with, I think one of the things that I like to think about when it comes to data is, like, is there a scaling? Like, how do you scale data properly? And this has brought me to this question about the bitter lesson, but recasting the frame of data. So the bitter lesson, for those of you who are not AI people, is the statement that methods that scale win eventually, if you can just scale enough, it wins.

0:52And so my question, I'll start with Sal here, is there a better lesson for data? Yeah, I mean, for sure. I think that you obviously need the right data, right, like in order for it to work. So I think like one misconception of scaling laws is that, you know, scaling laws are everywhere and they always exist. I think a lot of the work is actually finding that scaling law, right? Like, so a lot of what we really do is trying to figure out what's the situation where if you put more compute into it, if you put more data into it, you actually get a better result out. And the reason that's a great situation is then once that happens, you can kind of just turn the crank, right?

1:35Like becomes an engineering problem, which is something I like. But that has to do with architecture, but it also very much has to do with data. So if you don't have the data with the right information statistics to solve the problem that you want, which is, you know, I think probably pretty obvious, like if you think about it for a second, like you're not going to get a model that has the capabilities and the understanding of what you want. So you can only really pull the information from the data that you have and then use that to generate, like, to generalize beyond, right? Yeah. So when it comes to data collection, how do you think about what types of data modalities are best?

2:16And maybe a kind of related question is, Do you think that modelers have a tendency to work on problems which the data is available rather than maybe the problem that solves your goal most, like maybe has the most impact for translational medicine? Oh, for sure. So I think that, like, I won't speak for all modelers, but, like, I'm lazy, so I'm going to work with what's available. And, you know, I think there is actually, like, a good side and a bad side to this. And the good side is that I think when you're doing machine learning or conventional machine learning, you're really looking for the most pristine, high-quality data examples that you can find.

2:59But if you're like me, you're actually someone who goes and you're rooting around in the back room looking through people's junk. like what's there? So like a concrete example of this is when we train like when we train a protein language model we train on metagenomic sequences which are not the highest quality data. In fact much of that data I can guarantee you isn't even like a real whole protein and yet that makes the performance of the model go up for designing real proteins that work that make for understanding proteins that we know are things. And so, you know, that can... That's like the positive side of it.

3:40But that can lead you to a negative place where you say, well, let's just scale up the data that we can generate easily. And I think that's not the necessarily way to do it. And I think that's one of the things that's really exciting to me about what we're talking about here today with the VBI. it's really going out and saying what is the data that we need to solve the problem? It's the right resources, but also the right community. I think that one thing that we do at, or we try and do at Biohub anyway, is to work in the open, work with the community and move the whole community forward. And I think that's a really critical piece for this because if you just have people who are building models, we're going to lean towards the data that exists.

4:32If you just have people who generate the data, they're going to lean towards the things that can be generated. But if you can work together as a community every step of the way in an open fashion, then you can actually figure out what is the data that we need, and then you can find that scaling law. Cool. What do you think, Sal? A better lesson for data. Yeah, so I think the, so I was at DeepMind when Rich was with us and working and thinking about this idea of the bitter lesson. And the thing that I took from Rich's original sort of lecture that he actually gave at DeepMind was not about the actual notion of whether data is useful or how should we think about machine learning.

5:24I think my sort of take was that he was talking about something more conceptual. And the conceptual thing that he was talking about is sometimes when we are looking at problems, we think about solutions in a very religious way. I'm a modeler or I'm a data generation person. And I think that is the bitter lesson that if you approach the problem with that mindset, you might not succeed. The problem comes first. And you should be flexible in terms of your solution space. You should try both things. You should try to understand what is the problem that you're trying to solve. And if it requires modeling effort, do the modeling effort.

6:10And if it requires the data, collecting more data, then collect data. But sort of the thing that happened in machine learning in those times is people said, we are machine learning researchers. The data set is there. Here is some training data. Here is some test data. and we will just optimize the model. And that is broken, right? In the sense that if your eventual sort of goal is to solve the problem, then you have to be able to look at both aspects of what goes into the process. And what goes into the process is not just data or modeling. It's also about sort of expertise. So the example that I have is like with AlphaFold, we looked at what was possible with current existing data sets because we did not have the core expertise of now sort of going to, or even the resources, by the way, to sort of say, let's augment the PDB by a significant sort of order of magnitude.

7:24It was just not, I mean, the amount of sort of investment that organizations and scientists across the world had put in to construct that data was invaluable. So there you have to sort of focus on, well, getting the biggest bang out of your buck by investing it in modeling. But in other areas, if you're sort of approaching, say, cell genomics, then we took the same approach and said, well, what can you do with cell by gene? if you're looking at sort of cell genomics. And there it was very clear that we, after a bunch of work, that the data was not there yet to be able to build, to go after that grand ambition of building the virtual cell.

8:06And so it just brings people together to focus on what are the actual challenges if you want to advance science rather than being religiously following advances on data generation or in modeling. Cool. That's, I really like that answer. So what do you think, like, the actionable takeaway is from this? Like, I mean, always define the problem first and then, you know, figure out what solution space to search over. But more broadly, like, how should the community be thinking about this as we go forward for the next, you know, generation of basically trying to solve translational medicine? Yeah, I think the advice that I give to anyone who is working, who are starting in the area is that think of yourself as a multidisciplinary person.

8:57Understand the problem first. Why are you working on this problem? What are we trying to achieve? And then think about what is it that will be needed, whether it's modeling, whether it will be compute scaling, whether it will be sort of data generation. So understanding of that problem sort of is extremely important. And of course, then you have to build sort of your expertise in the house. but, and there are also, it's not as if like you don't have constraints. Maybe there are constraints, like there is only a specific amount of data that you can sort of generate, or there is a specific bottle size that you can afford to sort of train.

9:36Understand those constraints and stop and try to fail fast, and sort of look at what are the approaches that are going to be feasible in the long term in getting you to that intercept, the intercept level, the impact level that you're going after. Okay, cool. So this actually leads me right into my next question, which is when I look at sort of the evolution of modeling, let's say AlphaFold 2 was essentially a work of art, like a bunch of very carefully handcrafted features. There was a lot of thought, very intentional thought done to every part of the solution. And then I think some of Google or Alphabet's work has been kind of still stays in that space.

10:23But it seems like the general consensus is to move to more scalable, more general strategies. Do you think that there's still a lot of artisanal craft solutions or problems that need that? Or are resources better spent on, generally speaking, fixed resources, fixed money can go to compute. It can go to talent. It can go to data. you know, should we be focusing on scale first? So my, again, sort of let's approach it from first principles. And think about when, even so, like when you think about like the arts and crafts of how do you construct a model to be better, I think it was not an accidental thing, right?

11:07We did a lot of experimentation, but there was a vision behind it that all this scientific intuition that came from biophysics and biochemistry, that those interactions, that sort of amino acid sort of residues are not just doing their own thing. They are being influenced by other residues. So let's bake that in. Like, if you can, if you have learned something from the scientific community, use it. Use that information and try to sort of encourage the model and sort of give it that unfair advantage that it has. And it does basically make the model much more data efficient because it doesn't have to replicate everything.

11:49I also have to sort of mention that curating good data is an art in itself, right? It's not as if like you will say, well, put more data. If you replicate the same amount of data, you're not going anywhere. So it's not just about big data. It's about good data. And actually understanding sort of the coverage of what data is necessary for making progress in the problem is, in fact, I would say, a much interesting and much more challenging problem in itself. Yeah. Well, it looks like you have a thought. So what do you think? Yeah, I think, I mean, I very much agree. I think it really depends on, as you're saying, the problem to be solved.

12:34So if you have kind of a smaller amount of data, then certainly having more inductive bias in the model is going to help you. And you actually need that to get the results on smaller scales of data. As you get more and more data, you see that sometimes the model can find things that you didn't necessarily know about. And sometimes that inductive bias, if it wasn't exactly correct, can hold you back. So there is definitely that tipping point as things get better. But I guess also I kind of object to your question too because I think there's a lot of craft actually to the scaling part of things as well.

13:19So, you know, you need very, there's a lot of algorithmic work that actually goes into taking these models and training them on more data, putting more computing and making them bigger, right? Like, and we see that, like, not only is that from, like, an infrastructure perspective and, you know, making the models, like, run inference faster. But, you know, we're actually seeing that, like, things are going beyond transformers to more bespoke architectures that can, well, I mean, I don't think we're in a post-transformer world in any way, shape, or form, but we're modifying those architectures in order to make them more fit for purpose and actually work better, even at internet-scale data.

14:00So, you know, I think that's kind of the project, right? Like, as more and more data comes in, it's not only helping to curate that data and select what is the next batch of information that we actually need. You know, it's more information than data. How do you get that next batch of information that the models need? But then also, like, at every step of the way, at every scale, what is the right architectures to get the most out of that data? Cool. So if you think about the state of protein structure prediction right now, the news will say protein structure prediction has been solved. But then if you talk to my friends, they will all say, we have so much left to do here.

14:53so many things like function, dynamics, and then design are still just wide-open problems. It's been five-ish years since AlphaFold 2 was announced. Progress has been made. But I'm curious, Pushmi, what do you think? Is there another big leap being made here? Are there blockers in solving these problems? Do we not have the right data yet? Do we not have the right algorithms or ideas? or is this like something that we can see evolving in the future and it's just kind of a matter of time? Yeah, I think science, how it sort of operates is basically by isolating something and then making progress step by step.

15:38So when people sort of say the protein folding problem has been solved, like at a conceptual level, yes, there might be sort of, yes, there have been some advances, but I think it's also like when you think about that narrative of proteins being the building blocks. Proteins are not blocks and they don't act as blocks, right? I say proteins are the building blocks all the time but actually I don't believe in it, right? Proteins are extremely complex. They are disordered. their shape might change depending on the context. And like what AlphaFold did, and in fact like John and I, John Jumper and I basically used to sort of discuss this, like what are we trying to solve?

16:30We don't know what is the actual true ground state that proteins take and what is the actual distribution of structure that proteins take. What we are trying to do is basically Someone got a structure, deposited it in the PDB, and we are trying to replicate and get the same structure. That's what we did, right? And it just so happens that it's useful. But that does not mean that we have understood all of protein dynamics. So I think at the top level, it's easier to communicate that we have made progress, but the scientists among us and in this crowd, we know. Right? And I think that's the really important thing for us to understand and have that common ground that a lot of work needs to be done.

17:26And we have made progress, and there's a lot to celebrate, but let's not stop the funding of protein structure prediction and protein dynamics because we are just getting started. So what do you think the biggest blocker, Or if there's one thing you could wave a magic wand and say, like, we have more of blank, that would accelerate one of those things, function, dynamics, or design, what would the magic wand be? What would you wave into existence? Oh, I've basically went then. So I'm a machine learning, my background is like quite eclectic. I started as a security researcher, then went into computer vision, Bayesian theory, and finally got into discriminative machine learning and deep learning and so on.

18:11and then AI for coding, and then finally science. So when you ask me that question, like my computer vision, the computer vision researcher in me was super excited about cryo-EM micrographs. I was like, what is this sort of PDB data? I should be working at the source, right? I should be looking at the cryo-EM micrographs. I don't want those structures. They must be sort of missing out all the data. I should just directly operate at cryo-EM micrographs. Now, getting models that can scale at that level with the right amount of data and can extract all the dynamics and distributional information that is captured there would be an amazing thing.

18:52I tried it, but it requires more work, right? Still work on it, but you believe that this is a route. Yeah. This can give you... Yeah, I think at some point of time, maybe people better than me would take a stab at it and we'll get somewhere. What do you think, Sal? Yeah, I think that it's really interesting because I think these models are, I mean, they're quite useful, but they're not exactly the problem that most people want to solve. They solve a very specific purpose, and then you can also use them to do other, you can use them to design new proteins, right? Like, for example, which is not the thing that you would kind of start with there.

19:33But, like, so they're very useful. but I kind of think of like the models that we build right now as you've got someone who's trying to like understand how like a bicycle works and you're like modeling like a spoke on it, right? And you kind of have to like, those models can get better and better over time but like a lot of what you really need to do is you need to move from models of like spokes to wheels to maybe whole bicycles because like that's actually what people want to understand. Actually, you know, I think to continue the analogy It seems like people want to use the model of the bicycle to design this part for a pickup truck or something like that, right?

20:12But I think that as we put these models more into the particular context which they're in, then we'll actually be able to learn more about these interactions on a broader sense, right? And that's really, I think, where things are going to be going. But I agree with you. I think we should keep trying to work on folding models because I think they're going to keep getting better and better. So with regards to design, there's this famous Feynman quote, that which I cannot create, I do not understand. Now we're in a world where it's really easy to create things without understanding them at all. So I'm curious what your feelings, your thoughts are about how important it is to actually have models which can help humans understand things versus magical black boxes which can effectively one-shot a PicoMole or binder or something like that.

21:17I mean, I think that, like, I mean, we were designing things with magical black boxes long before, like, AI came around, right? Like, so in some sense, I think, you know, that's still useful. But, like, I think that the understanding is, like, really important. And I think that's actually one of the big things I think about a lot with AI because, like, you know, there's so much for these models to learn. and as intelligence is getting cheaper and cheaper and more available, you can deploy it to learn more and more things about what's going on in the world. But then how do you actually pull that knowledge out of the machine, so to speak?

21:58And so it's something that I can understand. Maybe that's just my esoteric curiosity. I think there's so many things to learn. I think lots of scientists really want to understand things and the endpoints are maybe not as important. But then, you know, we are here to solve translational medicine as a problem, right? Yeah. I mean, I think it's like, I think you do have to, like, the more you dig into things, the more that you can, like, find the right way to keep pushing them forward. So I think, like, one thing that's, like, really salient to me about this is I think the models actually have a lot more information in them than we know.

22:36So we've worked a lot on interpretability, for example, for our models. And you find a lot of information in there, which is... I think people know that protein language models learn some notion of structure within their representations. But we find information about functions. We find information about motions. right? And so I think there's a lot in there still to be unlocked, even from the models that we have now, right? And I think that's important for us to understand. How do these models work as we move up the, you know, as we raise the capability? Like, at the end of the day, like, you expect, like, a world model to come out of, like, trying to compress all this information into the model, How does the model do its job?

23:29How does it design a protein? Well, it's compressed all the information from evolution into this model. So in addition to it being able to spit out something which is useful to us, there's certainly something to learn just by looking at what's inside. What do you think, Prushmeet? Design versus understand. So I have a sort of different take on this in the sense that I think when we think about models, I think some level of understanding is necessary. So I'll say what level of understanding is necessary. So AlphaFold actually is not perfect. Right? It's not perfect. It's 90 GGT. AlphaFold2 was 90GDT at that time on that set.

24:26But even if it was 95GDT, but the PLDDT score was completely uncalibrated, who would trust it? It would just magically sort of give good answers, but suddenly tell you, here's the answer, very, very confident, and you will be working on it for the next one year and finding out it was completely wrong. the calibration of the uncertainty measure was extremely important. So in that sense, we do understand, and we actually made a lot of progress in understanding how does AlphaFold2 behave. Now, that's different from how did it work and find the solution. And there I agree with Sal that there's a notion of interpretability.

25:15Why did it work? Why did it give this answer? And I think if you think about it, AlphaFold 2 was at the highest level interpretable because if you look at how, and this is its ability to generalize, and we did not sort of discover it before launching. When we launched AlphaFold 2 and made the weights available, people found out that it was a great disordered protein predictor. It could figure out which elements of the protein are disordered. So it just shows that it generalizes. Now, interpretability asks the question, interpretable by whom? If you're saying interpretable by a human rational system with the cognitive and computational limitations of the human brain, then no, AlphaFold 2 is not interpretable.

26:13But if you are asking, is AlphaFold 2 sort of interpretable to a much larger and more sophisticated model as to how it works, maybe it is. We just don't get it, right? But I think as users of AlphaFold 2, we do need to understand what is it able to do and what it is not able to do. So understanding the behavioral characterization and the strengths and limitations of the model. And I think that is extremely important. And we can't sort of just be using these models without having that behavioral characterization because otherwise it will sort of, rather than being helpful, harm us, right? So your take is that interpretability, strictly speaking, isn't necessary, but a way of ensuring the model is trustworthy so humans can make actionable decisions is the thing that people should be focusing on.

27:04Exactly. And I also think that interpretability is also in the eyes of the beholder. Like, who is interpreting it? If it is a human scientist who is trying to interpret how the model is sort of going about and sort of how will it make this prediction, that's a different question from another much larger sort of LLM. If you give some of these large frontier models of the future access to the activation layers and say, okay, tell me, can you predict what AlphaFold will do? Maybe they will be able to predict, and they will come up with a theory of how actually AlphaFold 2 was interpreting and producing these results.

27:43Cool. Okay, that's the two-minute warning. So, one last question for both of you, a quick one. So, there's a real chance that AI will make dramatic improvements in human health in the next, you know, immediate future. I like quantitative predictions. best guess how long until we start seeing AI results in the clinic? Vishmi, you want to start? I mean, the point is AI results in the clinic, it's ill-posed. It's not a... AI is being used today already in every part of the drug discovery process. So from that perspective, it's already there. But if you are sort of asking me the question, when will we see that 10x acceleration or 100x acceleration in the timelines, then the other sort of question is, what are we accelerating?

28:43Is it sort of doing lead optimization? Is it the target discovery? Is it sort of the preclinical or the tox sort of work? So I think over the next sort of few years and which is why this effort that we announced today is extremely important. We need to tackle some of these hard challenges of biology. And only then will we be able to get these true unlocks of acceleration that dramatically transforms how drug discovery will happen. So AI will be there in the clinic and will be impacting sort of things that go into the clinic all the time. But like the actual sort of larger acceleration, that will be unlocked only with a better understanding of the biological models that this effort is trying to sort of create.

29:42All right, thanks. So real quick question, if we're almost out of time, but you. Yeah, I mean, I'm tempted to like literally put on my bio hat to answer this question. But, you know, I think it's, like, it's really important to think about this, you know, as you were saying. Like, what does it mean to, like, push the field forward? Like, the thing that we really want to see is those, like, outcomes affected. And, like, so when will it be that, like, a drug that was made entirely by AI, by an AI or something? I don't know. It's hard to know that. But, like, I do think we're going to see rapid progress here, like, very, very quickly.

30:22Because all these tools are being used and stuff like that. And going back to your point about the basic, we need basic research, we need basic understanding. One thing I learned a long time ago in my career, actually back at Google, is sometimes it's easier to approach a problem by looking at what is it going to take to make a 10x improvement rather than a 10 % improvement. And that's not because that's necessarily an easier path. It's because it allows you to take a broader view and see some solutions that you haven't been approaching. And what is actually the path to do it? And you would go back and look from first principles.

31:07How are we going to do this? How are we going to really push this? So I think both the approaches, the 10 % and the 10x are valuable, and we should do both. I'm just very happy that at Biohub, we have a very beautiful and lofty mission statement to cure all disease. And if you want to do that, you really need to go and figure out what is the 10x approach. And so I think it's a really great opportunity to be able to do that. Awesome. Thank you both. Thank you for being in the literal hot seat. Very nice.

From the publisher

From the Bitter Lesson of AI scaling to the unsolved mysteries of protein folding, Google DeepMind’s Pushmeet Kohli and Biohub’s Sal Candido are rethinking what it takes to build AI that truly understands biology. In this special panel moderated by Bra...

More from Latent Space: The AI Engineer Podcast

All episodes