Discovery Loop

From Jeff's twitter post:> Our general approach is to automate the experimental loop. We think this approach is broadly applicable across many different fields of science and engineering. We’ll initially focus on ML research and engineering, but believe the approach can help with important subproblems in nearly every one of the fourteen NAE Grand Challenge problems. We think doing this well requires strong expertise in machine learning as well as large-scale systems.See also: https://www.nae.edu/20782/grand-challenges-projectThose 14 are:NAE Grand Challenges for Engineering1. Make Solar Energy Economical2. Provide Energy from Fusion3. Develop Carbon Sequestration Methods4. Manage the Nitrogen Cycle5. Provide Access to Clean Water6. Restore and Improve Urban Infrastructure7. Advance Health Informatics8. Engineer Better Medicines9. Reverse Engineer the Brain10. Prevent Nuclear Terror11. Secure Cyberspace12. Enhance Virtual Reality13. Advance Personalized Learning14. Engineer the Tools of Scientific Discovery reply:...

I think people are missing what this really is: Google giving some of its most senior engineers the best retirement home to keep them away from competitors. This isn’t in jest; I wish i could make enough money to not care for more from my job and then do research after i get old. Its honestly a brilliant move.

I am siding with the "intelligence is not the bottleneck" crowd. Science takes more than reading literature and making a hypothesis. You have to run the experiment. And that is where messy reality will crush the naive, and resist any attempt to package it up into a factory-like innovation engine. But they will take your money, should you have some to invest. reply: I'm a scientist. On the one hand I take some comfort in thinking that I will always have an advantage in the lab. On the other hand I'm not taking anything for granted. And my advantage in the lab has to translate into an employer being smart enough to keep me around until if and when the AI takes over, which kind of translates into their investors wanting to keep me around.We know what happened to manufacturing when investors were no longer interested in it.

This seems to be an institutional, massively scaled version of https://github.com/karpathy/autoresearch.In March Karpathy described this direction: The next step for autoresearch is that it has to be asynchronously massively collaborative for agents (think: SETI@home style). Tweet is protected but in SERP caches: https://x.com/karpathy/status/2030705271627284816Seems like Karpathy was largely focused on ML / SWE research rather than the other domains this group is after. Still, hard to imagine they were not influenced by autoresearch.Andrej, if you're around, please share your thoughts on Discovery Loop. reply: That's a very silly comparison, there are many startups working on RSI, karpathy is just a basic version to try the concept (similar to his gpt work)

How do you automate experimentation?Doubtlessly, AI can iterate at superhuman speeds in the domains of thought and design: Software, mathematical proofs, literature search.But in the realm of experiment? Alas it is the lack of a body that constrains it.Rather than transcendence what AI requires is immanence. In the human flesh may we find the godhead living among men. Let the laboratories, warehouses, and factories fill with the sound of its labor, as it builds a wall with a million hands that are not its own.“Give me your tired, your poor, Your huddled masses yearning to breathe free, The wretched refuse of your teeming shore. Send these, the homeless, tempest-tost to me, I lift my lamp beside the golden door!” reply: In my area (pharma) what it looks like is this: A human defines a high-level research objective. "Identify a protein target that causes disease in humans, and find a molecule that binds to, and disables, that protein, eliminating the disease".That objective then gets loaded into an ML model that spits out an experimental protocol. A protocol can be as simple as:...

Really seems to embrace the "Making the world a better place by >" reply: Which part of it is highly technical or jargon loaded?

"Our mission is straightforward" continued by the most complex sentence on that page. Wondering what the definition of straightforward is now reply: > Our mission is straightforward: we are building AI solutions that can automatically solve important problems in machine learning, science, and engineering.Genuinely curious which part you found complex.

To be honest, this feels more like a lifestyle business (aka hobby) than a startup. They truly deserve it, but I don't expect a huge success as a business.That said, I hope they write cool papers with various peers across the industry without worrying too much about the competing dynamics. That'd be a blessing for humanity, and good for their spirit. reply: A public benefit corp shouldn’t be a start up. The primary goal of a start up is to grow as quickly as possible which is rarely benefits the public.Very silly to call every non start up a lifestyle business. It’s just a business. Start up are the weird thing that almost always an obscene waste of time and money, but sometime creates google.

Note that Jeff and crew have cleverly structured their company to avoid problematic uses of AI (e.g., weapons or tracking humans). I suspect that many top researchers will want to work there for this reason, and to work with other top researchers who have a history of delivering results.

This is very cool. It might be a new scientific revolution to have computer-driven discovery. So often we find things that are "this could have been done 20 years ago" and with an indefatigable searcher perhaps we'll close all those things. Though it does remind me of that Ted Chiang (I think) story where humans and superhumans coexist and all the science of the former is meta-studies of the work of the latter.

That founding team is insane. Very excited to see what happens here. I really like that they do not mention AGI or anything like that. Their mission statement reads pretty pragmatic compared to other AI companies (the bar is very low…)

Someone who left DeepMind over Google's agreement to provide military AI to the US government tried to get Jeff Dean to quit too:https://turntrout.com/why-i-left-google-deepmindMaybe this is what happens when someone with Jeff Dean's standing tries to quit?TBH, I'd rather have Jeff Dean working on the creepiest-possible tech for ICE than joining the race to automate AI research. Automating AI research is terrifying.

I wonder if Jeff Dean facts [1] (I hope people remember the reference) will carry over to the new startup.[1] https://github.com/LRitzdorf/TheJeffDeanFacts

I’ve always felt that the idea that science is bottlenecked and therefore needs more automation only works for a very narrow definition of what science is, and entails a very specific view on what it should be.

Jeff Dean, Sanjay, et al have achieved so much. I'm very happy for them. Truly deserving.Sometimes I couldn't resist wondering if I'll ever do work that has a tenth of the impact of theirs.

with billion dollar seed rounds becoming the norm, it seems like there's no longer an advantage to build from within these bloated giants. can be much nimbler and have access to the same budget out the gates

> Scientific discovery is bottlenecked.Yah, by funding and how we award it, not by an imaginary lack of undergrad and grad students. Scientific funding requires a shotgun approach and many national science funds try to pick winners as opposed to funding broadly. When the folks who researched bacteria in volcanic vents or the molecular biology of the Gila monster they never could have imagined the industries and markets they'd create let alone the lives they'd impact (i.e., PCR and GLP-1 agonists). Lots of grants require you to explain how the work is "translational" or has some sort of economic application (even if not explicitly), but that'll just get us faster horses or whatever the Ford quote is.

For sure made with Claude code for front end, but I’m excited to see where they go

This is one of the interesting aspects the 'AI job loss' community doesn't account for. As the technology unlocks things, more startups are created. And even at a lower nominal engineer-to-work ratio, overall demand for talent still goes up. Ultimately, we are not a single group trying to achieve a common outcome, we are a collection of many groups trying to compete against each other.

The site itself is really leaning into the “made with Fable” aesthetic

LawZero, Yoshua Bengio’s startup, also proposes to automate scientific research and experimentation, from the perspective of safety, by being explicitly “non-agentic”:https://lawzero.org/en/publication/scientist-ai-safe-design-...

Discovery and optimization are very different processes. Optimization is the process of finding the shortest path to a goal. Discovery is the process of stumbling on new goals and redrawing the map of what's possible.Ambitious goals and new discoveries happen via novelty-based search. Progress in scientific discovery is measured by how different/interesting the outcomes are, not by closeness to a predetermined goal.Discovery is a creative search that preserves optionality, whereas optimization restricts optionality. In other words, you usually don't discover anything novel unless you're trying new things that don't appear connected to the goal in the first place. Would an ML optimization loop have discovered transformers?

Here are just a few of viewpoints on what constitute world problems to solve:https://80000hours.org/problem-profiles/https://en.wikipedia.org/wiki/List_of_global_issueshttps://encyclopedia.uia.org/Interestingly, one list identifies "AI" as a top world problem! One person's problem is another person's solution, I guess--and vice versa, as well.An extreme example: curing a disease is good for patients but bad for the healthcare industry--which is (in kind) also bad for healthcare workers and everyone in science working on cures.

Two to keep in mind with these kinds of things -1. There’s some irreducible costs in terms of time and material in the physical world that are not amenable to the kind of optimization or parallelization or even just the raw speedup from Moore’s law or computational architecture improvements we’re used to with software. My experience is primarily in biology, where the examples here are things like “it takes 20 minutes for E. coli to replicate” - it has taken 20 minutes for E. coli to replicate for a billion years, and next year it will still take E. coli 20 minutes to replicate, no matter how good your software stack is. Similarly, it takes X amount of energy to grow enough E. coli to produce a meaningful result, and that energy costs money, whether it’s in the form of glycerine or heat or whatever you want, and that also won’t materially reduce in the same kinds of “orders of magnitude” sense we’re used to from software, which is what we’re usually expecting to make the economics of these things work out.2....

> we have pioneered massive scale computing and led the creation of critical infrastructure, products, and foundational AI advances that the world relies on, including multiple generations of Google Search, Google Ads, Google News, Google Translate, Google File System, MapReduce, BigTable, Spanner, TensorFlow, Pathways, TPUs, AlphaChip, AlphaStar, AlphaCode, AlphaFold, Gemini, model distillation, mixture-of-experts model architectures, word2vec, sequence-to-sequence models, chain of thought reasoning, neural architecture search, and multiple generations of Large Language Models (LLMs) among others.holy shit. I've known this, but...

By the middle of the 2030's the world we live in will be unrecognizable.

Oh wow, that's a blow to Google, what's with the talent scarcity in ML. Though if this goes anywhere Google will likely buy them back.

Surprised that Sanjay is the tallest among them.