Domenic Denicola
@domenic
Working on @googlechrome to make the web better. Mainly as a way to pass the time before the singularity hits.
I've done a full writeup of my experience participating in the @METR_Evals study on AI productivity: domenic.me/metr-ai-produc….
New paper & surprising result. LLMs transmit traits to other models via hidden signals in data. Datasets consisting only of 3-digit numbers can transmit a love for owls, or evil tendencies. 🧵
It’s worth reflecting on just how fast AI progress has been, especially in math. In 2024, AI labs were using grade school math (GSM8K) as an eval in their model releases. Since then, we’ve saturated the (high school) MATH benchmark, then AIME, and now are at IMO gold.
A good post from @ariel_zj with some new thought experiments on personal identity. preservinghope.substack.com/p/new-thought-… (I continue to be a patternist, and thus fine with Star Trek transporters or destructive uploading.)
1/N I’m excited to share that our latest @OpenAI experimental reasoning LLM has achieved a longstanding grand challenge in AI: gold medal-level performance on the world’s most prestigious math competition—the International Math Olympiad (IMO).
A pretty interesting, albeit in-the-weeds, discussion about typographic details when combining CJK languages with non-CJK text: github.com/w3c/csswg-draf… The context is figuring out if browsers should change the default behavior. (The conclusion is they should not.)
The Social Network sequel film opens with Zuck giving a rousing speech to FB employees. Zuck: We need to win the JavaScript framework wars! What is this multidirectional binding garbage. Drop the bidirection - just unidirectional. It’s cleaner. Engineer: But.. but it’s never…
‘THE SOCIAL NETWORK: PART 2’ is in the works. • Aaron Sorkin is set to write and direct • Unknown if Jesse Eisenberg will return as Mark Zuckerberg (Source: Deadline)
I wonder how GitHub ships a regression, for multiple days, that filtering issues by "closed" state does not work anymore. github.com/httpwg/http-ex…
Chrome 138 is here, featuring built-in AI APIs for translation, summarization, and language detection → goo.gle/4ehjP0t Also, check out the new CSS functions and adapt your web layout for foldable devices with the Viewport Segments API.
I don't understand how GitHub is still not showing changes to large files in the review panel. Like, that's really anti-developer? To save a bit of cpu on the backend??