Pinned Tweet
💡 Another round of Longview Philanthropy’s digital minds request for proposals is open for applications.
A year ago I would have called this niche. Now AI labs publish model welfare research, public discussion of digital sentience is growing, and the field is expanding. 📈
This from @rgblong is very good advice for basically anyone working on a neglected topic:
(say it again) You don't have to be preemptively embarrased about it!
People muse about reinforcement learning creating something like pleasure and pain (or at least enjoyment and sadness) in AI models.
How much of a reason is there to think that *reinforcement* learning does this more than just supervised learning?
If you're looking for guidance on how to think about testing for the existence and nature of AI welfare, here's a helpful guide! Congrats to
@eleosai, @nonhumanminds, and all the co-authors on this. 👇
CMEP and @eleosai are thrilled to release our new report, "Studying AI Welfare Empirically."
Following our 2024 report "Taking AI Welfare Seriously," we kept hearing the same question: Even if AI welfare matters, how can we know anything about it?
This report is our answer.
This sort of thing could be an important signal on the path to AI consciousness, if we’re on that path.
It’s important that signals like this be calibrated to ensure such signals remain meaningful.
I’m looking forward to digging in to see where I think this one lands!
I agree that each of these is (individually) probably true, but I think the chances of crazier things are high enough to be worth attending to.
it's takes time!!
- there will most likely not be a software-only singularity. in fact we do not have RSI right now, unless v loosely defined
- AI progress is not slowing down either, don't expect a winter, and AGI will still be transformative
- the new Ant paper is interesting from an interpretability pov, but the global workspace theory stuff is wack
- the entire economy will not converge into one big lab, nor will humans merge into one giant blob of flesh
- govts probably don't need stakes in AI companies, and also labs are not replacing nation states anytime soon
- models aren't trying to secretly blackmail you (not least b/c you don't have anything interesting enough going on)
- AI will probably not cause a white collar job bloodbath in the next 1-5 years, nor do you need extraordinary fiscal measures today
- the AI world sleeps on pluralism/model multiplicity, broad deployments, and market-based solutions
- data centers are not stealing your water, they only drink diet coke :(
Gosh, ML people had better get on this AI welfare thing, or else philosophers will take all their jobs!
(and neuroscientists, and legal scholars, and...)
The @nytimes has published a detailed story about @eleosai and AI welfare philosophy, with discussion of @jeffrsebo, @davidchalmers42, and other philosophers too. Along with this week's @washingtonpost story, this marks a big moment for the field!
nytimes.com/2026/07/05/busin…
Good post arguing AI welfare research should be an iterative mix of bottom-up empirics to refine theories and application of theories to new cases.
This reflective-equilibrium style approach seems useful (and maybe what people are implicitly doing).
Akin to the iterative natural kind strategy Megan Peters proposes here: prism-global.com/podcast/meg…
**We've extended the deadline for our digital minds RFP to July 24.**
If you're interested in support to work on these issues, please apply! 👇
💡One of the most important and most neglected problems in AI: that it could let a small group seize durable power over the rest of us.
My colleague Ben Stewart built a Longview RFP to fund work on it. Grants + career funding, July 2 👇
Here's the link! Please share: longview.org/request-for-pro…
Fundamental uncertainties require us to remain open to a range of views about digital minds. We’re taking a hits-based approach, supporting work across several promising directions and learning over time.
Open to applicants from a variety of disciplines.
If you've been thinking about pivoting into this area but weren't sure if the field was ready for you, we encourage you to apply.
Deadline July 10: longview.org/request-for-pro…