Profile avatar
ntraft.bsky.social
PhD student at the Vermont Complex Systems Institute. Interested in ML, evolution, self-organization, & collective intelligence. http://ntraft.com https://t.co/ja3fdtRLdM
151 posts 342 followers 232 following
Regular Contributor
Active Commenter
Posts 21 Comments 29

PC message for CVPR. 2 remarks: - 22% is lower than usual. I wonder what happened with the general mood of the reviewers. - 19 desk rejects for irresponsible reviewers. I'm 100% behind the PCs on that one. I feel bad for the students who worked hard, but it had to be done.

submitted 15 hours ago • 1 comment

Everyone claims to be "data driven" nowadays but rarely put their money where their mouth is. I experienced this in self-driving car startups. We were excited to build a robot car so we focused on the product. But in retrospect, years later, I wish we had started with a data science super-platform.

submitted 3 days ago • 1 comment

I heard you like nostalgia? smbc-comics.com/comic/fads

submitted 3 days ago • 36 comments

If you write well enough, 90 years later people you've never met will try to make the thing inside your head real.

submitted 6 days ago • 4 comments

We are truly living in the worst timeline.

submitted 4 days ago • 0 comments

Rings very true! "You don’t actually have to invest in tech that works, or that even makes a profit, you simply have to have a narrative that is compelling enough to float those valuations." Though some things e.g. the PC really did revolutionize things. But the system can't tell the difference.

submitted 7 days ago • 0 comments

The methods of McCarthyism are alive and well, perhaps even MORE extreme than ever before. Our society has apparently learned absolutely nothing and has no qualms with doing it all over again.

submitted 7 days ago • 0 comments

I'm very happy to announce a strategic partnership between @wikimediafoundation.org enterprise and Pleias for open, ethical and trustworthy AI innovation. enterprise.wikimedia.com/blog/pleias-...

submitted 8 days ago • 1 comment

I feel like I'm going insane every time I see a media story about DOGE that doesn't also mention the fact that the Government Accountability Office exists and performs the exact function that DOGE pretends to, only in a credibly constitutional, non-partisan, and conflict-of-interest-free manner

submitted 10 days ago • 188 comments

Our latest work uses multi-agent reinforcement learning to model corporate investment in climate change mitigation as a social dilemma. We create a new benchmark, and show that corporations are greedily motivated to pollute without mitigating their emissions, but if all companies defect...

submitted 14 days ago • 2 comments

That process for making a decision? THAT'S the free will! That's just you. Deciding.

submitted 11 days ago • 2 comments

Building a computer is like LEGOs for adults, but instead of having a fun Bionicle at the end of it, you have a portal to Hell that constantly streams unrelenting horrors into your consciousness.

submitted 12 days ago • 0 comments

Likely the original source for the idea of the NIH indirect cost cut... But the crazy part is that the report doesn't actually recommend a universal cap; it simply recommends that the govt refuse to pay more than private donors, so that the rates become equalized! www.wired.com/story/nih-in...

submitted 12 days ago • 0 comments

Nice perspective on the Paris AI summit, and the concerning trend "towards equating AI progress with national capabilities." Looking at AI as a nationalistic competition is honestly very sad and would be a really limiting way to frame any new technology. These are not nuclear bombs.

submitted 12 days ago • 0 comments

UVM is now a Carnegie R1 institution! This achievement reflects decades of dedication to cutting-edge research, innovation, and excellence. With over $260M in funding (FY24) and groundbreaking projects, we're on the course for an even more ambitious future! #UVMresearch

submitted 13 days ago • 0 comments

So excited about AI tools that make us better: collaborative assistants, tools for scientific investigation, decision support, markets that incentivize good outcomes, social media algorithms that help us achieve consensus. There is so much that isn’t “automate away humans.”

submitted 15 days ago • 5 comments

WHOA... Look at the replies to this skeet. I'm honestly not sure I've EVER seen this level of vitriol. There's apparently a huge camp of Bsky users who attack any use of machine learning to power the platform. From what I can see, this seems like a fairly benign, open-source use case(?) #MLsky

submitted 16 days ago • 3 comments

We're excited to announce the first Evolving Self-organisation workshop at GECCO 2025! Submission deadline: March 26, 2025 More information: evolving-self-organisation-workshop.github.io

submitted 16 days ago • 4 comments

Awesome survey from Bertoldi lab. Fun fact: you can get metamaterials to compute as well, thus conflating "thought" and "action". Take that Descartes! dl.acm.org/doi/pdf/10.1...

submitted 23 days ago • 0 comments

Regarding the NIH indirect cost cut which was just announced, the motivation for the idea:

submitted 19 days ago • 0 comments

Here's the script I wrote for calling my representative about the latest horrors. Would probably benefit from a better breakdown of how indirects work, but I didn't want it to get too long. Now I just need to decide whether to call now or wait for Monday when I can talk to an actual person.

submitted 19 days ago • 6 comments