Gavin Crooks

Physics-native unconventional computing.

[Cascade Quantum @CascadeQuantum]

Hello, many worlds.

We’re Cascade Quantum.

Our goal is to make it effortless to create elegant quantum applications. Here's what we're building:

#QuantumComputing #QuantumMachineLearning #QuantumAI #LifeSciences #DeepTech

[Gavin Crooks @gavincrooks]

Everything's Ising model

[Surya Ganguli @SuryaGanguli]

Our new work on “Physics of Agents” https://t.co/qKo3O77kZ9 lead by Batu El and Jinhee Paeng in collab w/ @james_y_zou

The outcome of many interacting agents seems hard to reason about. Yet we were able to study the opinion dynamics of 10,000 different LLM agent communities as they communicated with each other to solve both objective and subjective questions.

Remarkably, we could account for their opinion dynamics through a simple Ising model that involved minimizing an energy function corresponding to social pressure to conform. Our dynamics could explain the build up of consensus, polarization, and societal correction of initially incorrect majorities.

Lots more to do on statistical mechanics of interacting agent dynamics!

[Surya Ganguli @SuryaGanguli]

Our new work on “Physics of Agents” https://t.co/qKo3O77kZ9 lead by Batu El and Jinhee Paeng in collab w/ @james_y_zou

The outcome of many interacting agents seems hard to reason about. Yet we were able to study the opinion dynamics of 10,000 different LLM agent communities as they communicated with each other to solve both objective and subjective questions.

Remarkably, we could account for their opinion dynamics through a simple Ising model that involved minimizing an energy function corresponding to social pressure to conform. Our dynamics could explain the build up of consensus, polarization, and societal correction of initially incorrect majorities.

Lots more to do on statistical mechanics of interacting agent dynamics!

[Gavin Crooks @gavincrooks]

@ito_sosuke Ah ha, I (and Claude) missed that paper! Looks very relevant.

[Theofanis Karaletsos @Tkaraletsos]

Our perspective on “Fifteen challenges for generative AI applications to cell biology” is now out in Cell, led by @califano_lab .
It was a pleasure to brainstorm through this with a great group of people.

What we tried to do is list a few key challenges worth working on we see ahead for AI in biology, with optimism and enthusiasm for the road ahead.

https://t.co/vuJSMRKy5W

[Gavin Crooks @gavincrooks]

@BeaverCattle Its a problem. It take me longer to verify the results than it takes claude to generate them. But Ive now made Claude verify all the math in lean.

[Gavin Crooks @gavincrooks]

Physics is going to be as cooked/cooking as math. I fed Claude an open problem in stochastic thermodynamics of the kind I'd suggest to a mathematically inclined grad student. And over a few days of back and forth, it did months of work and closed the whole problem class.

[Gavin Crooks @gavincrooks]

@ChavarriaAlvaro @skdh The AI revolution is coming for everything I suspect.

[Gavin Crooks @gavincrooks]

@bladeofshanghai In knew it wasn’t a solved problem because its in my field of expertise. Plus I got Claude to trawl through the math and physics literature.

[Gavin Crooks @gavincrooks]

@MiguelArratia The Book of Nature is written in the language of mathematics.

[Gavin Crooks @gavincrooks]

@yanyanmomoko I’ve now gotten Claude to prove all the math in Lean, so there’s a good chance its correct and rigorous.

[Gavin Crooks @gavincrooks]

@FoleyTtfoley I gave Claude a handful of initial references, but it tracked down a bunch of others on its own. My role was mostly uploading pdfs, since @arxiv is still holding back the progress of science by rate limiting Claude.

[Gavin Crooks @gavincrooks]

https://t.co/VYYeAie4YZ

[Gavin Crooks @gavincrooks]

@FoleyTtfoley Absolutely nothing special for this project. Claude Max plan, mostly over the app on my iphone.

[Jason Locasale @LocasaleLab]

Any experienced scientist who has not already done so should try working through a difficult research problem using about $100 worth of compute, with AI acting as a postdoc or graduate student.

The experience, provided you know how to use the AI, will change your understanding of what is now possible.

In experimental science, the experiments can now even be run through cloud or remote laboratories at a fraction of the cost of operating a university lab.

[Gavin Crooks @gavincrooks]

@TorstenAkesson @skdh Indeed. It could be that Claude et al speed runs the math/theory/data parts of science, but hits a wall
at the experimental connection to reality.

[Gavin Crooks @gavincrooks]

@benjfdup It might just be a question of knowing the right questions to ask, which Claude doesn’t seem that great at. Yet.

[Gavin Crooks @gavincrooks]

@IvanZupic It is the dawn of a great new era of science. But how we do science (the grants, the mentoring, the publishing, the credit….) is in for a shake up.

[Gavin Crooks @gavincrooks]

@yaroslavvb That’s great idea, I shall set up something similar forthwith.

[Gavin Crooks @gavincrooks]

@JivanShah139154 We must reclaim twitter.

[Gavin Crooks @gavincrooks]

@MattLeighton5 Ah, I forgot those! I’ll put up a proper pdf presently.

[Gavin Crooks @gavincrooks]

@Brackto Quntum gravity is an interesting case, since progress seems to have stalled due to a dearth of experimental constraints. Can AI break the impasse?