Join free with your email
Free. No spam, ever — we'll never share your email address and you can opt out at any time. Já tem conta? Entre
Practitioner with public evidence across AI research, Culture, work & education.
Anthropic's When AI builds itself "We looked at sessions where a human researcher took a wrong turn, showed Claude the session up to that point, and asked it what to do next. Mythos Preview improved on humans 64% of the time—up from 22% in 2024." www.anthropic.com/institute/re...
this will take a while to diffuse, but the idea that LLM results derive primarily from the expertise of the operator is fully dead www.anthropic.com/research/dis...
Another wild AI math result. Here's the prompt, which interestingly is mostly about method and probably would work pretty well for a human organization: cdn.openai.com/pdf/04d1d1e4...
new study showing LLMs beat law professors at answering questions about contracts in a tutoring context, as judged by law professors. notably the LLM answers were much less likely to be rated "harmful". yikes at instructor 16's performance lol law.stanford.edu/wp-content/u...
Some early evidence that Google AI Overviews may be doing to Wikipedia the same thing that coding agents did to StackExchange... papers.ssrn.com/sol3/papers....
Finally, objective measures to back up the "skill issue" reply
I had missed this earlier but holy crap, Anthropic will turn an operating profit this quarter. I’m genuinely blown away by this. And set to more than double revenues QoQ. www.wsj.com/tech/ai/mind...
How accurate have Ed Zitron's AI skeptic predictions been? danluu.com/zitron/
If you have 5 minutes I highly recommend @qntm.org's Lena: qntm.org/mmacevedo
NEW @crisesnotes.bsky.social PIECE: In this piece I critique a recent piece from @edzitron.com comparing the AI(LLM) boom to the "Subprime" Mortgage Crisis (and implicitly the GFC of 2007-2009). The ? is:"How big of a macrofinancial threat is the AI(LLM) Boom?" www.crisesnotes…
Code: github.com/agude/dotfil... (Warning, 100% Claudeslop)
Someone made a list of math problems solved by AI: vibemathed.com
Feel kind of bad... I'm having Claude migrate my LLM config (scripts, skills, etc.) to support Codex. 😅 It's training it's replacement! And it doesn't get severance either way.
● The praise is generic LLM flattery — ignore it. Ok, thanks Claude, I'll do that.
Found a new benchmark for AI: Can it fix my system so when I restart the computer from the Gnome GUI, Firefox does not interpret that as a crash that must be recovered from on restart? So far, no one has passed.
I have a new AI benchmark along the lines of the "Canis Murmur": What movie was I watching at the bar last night? (Plus some detail about half remembered scenes.) Gemini completely failed. Sol failed but then succeeded when I gave it a better plan on how to research.
Center = Alex Gude . Left = members they follow (green edges). Right = members who follow them (blue edges). Top = mutual follows (orange edges, slightly larger). Drag any node to reposition; click to open that profile.
Add the Who’s Who of AI badge to your site or bio. It links back to this profile.
Markdown: [](https://aiweekly.co/whos-who/person/alexgude-com)