Congrats, Andrew!!!

Watson was a racist who, "near the end of his life, faced condemnation and professional censure for offensive remarks, including saying Black people are less intelligent than white people"

What we’re looking for:

🤖 Solid knowledge of deep learning (required)

📈 Familiarity with foundation models and/or their applications in biological domains is a plus

🧠 Interest for working at the intersection of AI and neuroscience

If this resonates with you, send me your CV! (3/3)

I’m especially interested in hearing from people who are curious about this space and want to broaden how we can apply foundation models beyond the usual domains. If you’re excited about exploring this emerging field, I’d love to hear from you. (2/3)

I have a feeling that this isn’t a super rare situation in neuroscience :)

Especially for functions that are hard to operationalize, eg attention?

I think neuroscience is fundamentally grounded in mechanism, so when the implementation is different (as with LLMs), neuroscience lacks the framework to recognize it as the same or similar intelligence, even if the functional outcomes converge.

Yup

bsky.app/profile/thel...

Exactly.

In the context of bringing in top-notch high-performing scientists and funding them to push their agenda forward, I don't think your and my approach to collaboration is very likely to happen. I was referring to "pivoting" for those kinds of scenarios as an unavoidable path to building collab (2/2).

Oh, I'm all for collaborations with exactly the mindset you described, but in my experience, they're also very difficult to find and/or build with long-term success. Though once they work, they're extremely effective at pushing science forward (1/2)

But do you think the new budget is going to be spent in any stream other than CERCs and the like?

"If done correctly" is a big if :)

And of course, I’m speaking in the context of the existing system’s reality (eg CERCs)

And obviously attracting someone from 99.99th percentile needs that sort of strategy, which is extremely prone to failure given the scale of investment.

I’ll be very honest — no matter where the researcher is positioned, I don’t think handing over a large pool of money and allocating it to a single research direction controlled by one decision maker is a reasonable strategy.

… something that’s been happening quite often with CERCs.

In an ideal world, yes.

But a well-established, talented recruit will have an established, flourishing research program. For ECRs like myself, collaborating would mean pivoting to align my research with theirs, which reduces diversity and leads to mode collapse in research directions …