Continuing to fiddle. I've discovered that data visualization is extremely susceptible to "procrastiwork". I should just stop and release my notebook.

We still don't really know what drove Tyler Robinson, or if he even holds a coherent worldview. The tone certainly shifted when he did not match the hoped-for demographic profile. The civil war was cancelled, at least for now, to be replaced by Young Male Mental Health Appreciation Week.

did a little media criticism www.theverge.com/politics/777...

More of this kind of reporting, please!

Glenn Beck’s media outlet The Blaze just gave an interview last month to the person who killed the politicians in Minnesota

As for the claim that "half of AI scientists think there is 10% chance or higher that we'll be wiped out by AI": See my account of what really happened in this survey at aiguide.substack.com/p/do-half-of...

The Nazi early prohibition on Jews from owning guns has been an NRA talking point for decades.

Not to mention the enormous amounts of ML paper slop on arxiv, stuff that is full of stuff that is just completely wrong, but has enough stuff that resembles math or experiments in it. My point is, this seems like an arbitrary line to draw about what is good enough for arxiv

When you think about it, we are basically just computers - *person who works with computers for a living*

Kinda wow: the mystery model "summit" (rumored to be OpenAI) with the prompt "create something I can paste into p5js that will startle me with its cleverness in creating something that invokes the control panel of a starship in the distant future" & "make it better" 2,351 lines of code. First time

East Asia friends! I will be in Seoul next week, then Tokyo ~2 weeks, then Taipei for a weekish! Let me know if you're around and want to grab a lunch, coffee, or beer!

this is one of the worst long-term trends for US political stability forming in media res

🧪 BREAKING (good news): Senate subcommittee says NO! to Trump's proposed slashes to NASA & NSF funding. Today, the subcommittee said to keep NASA + NSF funding at $33.9 billion, the same as in FY24. See 7:15 below. Full Senate appropriations committee meets tomorrow about it. 🧵 1/3

I'm not going to repost any of the insane antisemitic conspiracy bullshit that grok is spewing today, but it highlights how absolutely essentially is that we not let LLMs become a form of epistemic grounding for our society.

For the longest time I've been using Google Translate as a gateway to explain machine translation concepts to people as it's a tool that everyone knows. Now I get to contribute over the summer. 🌞 If you're near Mountain View, let's talk evaluation. 📏

Can coding agents autonomously implement AI research extensions? We introduce RExBench, a benchmark that tests if a coding agent can implement a novel experiment based on existing research and code. Finding: Most agents we tested had a low success rate, but there is promise!

Crazy peer reviewing proposal: 1. Decouple longform reviews from acceptance decisions 2. 10 or 20 reviewers give a simple upvote/downvote after quick readthrough 3. 1 reviewer per paper writes a longform critique w/o score 4. AC uses (2)s score, (3)s crit Lighter workload + more eyes on each paper

ngl the more coerced and transactional reviewing becomes the less eager I am to review when I didn't submit

Multicultural text-to-image work requires costly, subjective human evaluation. Some of my projects have stalled because no automated, quantified "visual cultural attribution" metric existed. BITS undergrads Siddharth and Arnav Yayavaram, @simi97k.bsky.social, @gneubig.bsky.social, and I made one.1/

in elementary school I happened to be in the same strange school play as future YouTuber Jaiden Animations. It was such a strange play she made a video about it, including a clip of childhood me singing "you are ugly" Thanks to YouTube rewind my Bacon number is 4 (Will Smith, Jeff Goldblum, Bacon)

Really hitting home after hearing gunshots at the SLC march. Apparently the suspected gunman arrested after shooting & critically injuring a man. The idea of violence against objects is a category error.

And here is the remote pre-recording of my talk for the wultiwodal meta-evaluation tutorial at CVPR youtu.be/ymwPz1sioJI

I love this film so much, because when you think about it the government's response to discovering Kaiju are real being a bunch of bureaucrats immediately trying to establish that "dealing with Godzilla does not fall within the purview of my department" is actually the most realistic take yet.

I have at least two funny meme LM evaluation project ideas for any interested collaborators

Breaking News, I am just getting word that they are in fact called "Multi-", not "Wulti-modal models" Thank you @shaily99.bsky.social for informing me

Check out our CVPR tutorial "Evaluating Large Wulti-modal Models: Challenges and Methods" on Wednesday from 1-5pm in room 109! Unfortunately, I won't be there to present but my labmate Xiao will share some of my slides during her section :) lmm-understand.github.io

broooo what if some of those that work forces are the same that burn crosses

This piece rhetorically asks: "Should the climate movement start demanding that everyone stop listening to Spotify? Would that be a good use of our time?" unfortunately I think many would say 'yes'. andymasley.substack.com/p/individual...

"The cost of living has increased, but the cost of owning has increased more" says the rent hike letter from a landlord who had funded 0 repairs since I've lived here, in CA with frozen property tax

A cheeseburger uses a lot more water than a ChatGPT request 🍔 Actual farms, not the data center variety, are sucking up groundwater more quickly than surface water, explains @markgongloff.bsky.social 🎥

Proposed to cut number of people involved in NSF activities by 70%. We are literally on the chopping board. Call your reps.

To be honest, I kinda love grok? (when it isn't being Elonbotomized to be a racism machine) So many rightoid maniacs query it expecting to see their conspiracist beliefs echoed back at them only to repeatedly get gently corrected with factual information lmao

I cannot stop thinking about Andor. Masterpiece, must watch for pretty much everyone imo

Sent my thesis in to my committee this week, will defend June 2 at 1pm PT! If you're interested in catching it on zoom, here's a calendar link! calendar.google.com/calendar/u/0...

Despite clickbaity title this is a great level-headed piece from a real scientist who tried working in AI for science. The key point that AI is a tool not an all encompassing revolution is common sense but the details are interesting and illuminating open.substack.com/pub/understa...

my junior coauthor who realized too late they didn't provide the license for every individual GLUE task: *sobbing* are you sure it's ok?? me: please come down off the ledge

If we just add a few more annoying tasks for authors and a few more for reviewers we can fix peer review in AI!

According to a 2021 report, the University of California system: • generated $82B in economic activity in California • supported 529K jobs in the state • generated $21 in economic output for every $1 received Public divestment from higher ed makes no sense, even in the narrowest economic terms.

Michael News! I will be joining the Tech Policy Lab at the University of Washington @ischool.uw.edu and UW NLP working with @aylincaliskan.bsky.social as a postdoc in the fall, to work on situated evaluation, multimodal/lingual/cultural genAI, and new directions in safety, fairness, and alignment!

"Women are PIs on 58% of the canceled grants, although they are PIs on only 34% of all active NSF grants. Similarly, Blacks are PIs on 17% of the terminated grants, although they make only 4% of the total pool. Hispanic PIs and those with disabilities were twice as likely to lose a grant."

There's no escape! Even in my sister's bar admission ceremony the bar president starts talking about AI 🤣

We were interviewed for IEEE spectrum about reasoning models! spectrum.ieee.org/chain-of-tho...

What is it about the City of Berkeley and Country of England that makes interest in AI safety and weirder fringe stuff like AI consciousness so prevalent? Like why are these topics so big there and not in like Seattle or Pittsburgh??

Finally a study mix I can get behind www.youtube.com/watch?v=0tR5...

"LLM on way to replace doctors" gets published in Nature. meanwhile "LLM judgement not as good as human MDs" gets a spot in "Physical Therapy and Rehabilitation Journal".

Mic drop from @tallinzen.bsky.social: "It’s sometimes confusing when we pretend that there’s a scientific conversation happening, but some of the people in the conversation have a stake in a company that’s potentially worth $50 billion."

Very interesting oral history -- interviews with some top NLP folks on the effects of GenAI on their field: www.quantamagazine.org/when-chatgpt...

We won an outstanding paper award!! 2025.naacl.org/blog/best-pa...

PSA for NAACL peeps from a southwest boi (sadly I won't be there): be sure to find a place to eat New Mexico style stacked enchiladas. You can get it "Christmas style" where its served with both red and green hatch chile. The hatch chile is integral, do not skip. Not photogenic, but very delicious

I wondered if it could really be all that bad from the beginning, after all users are signing up to publicly interact with each other on a forum but woof, I don't think I would have signed off on this broad of a "the LM is allowed to impersonate this" policy