minimum viable opus @pico
San Francisco, CA
Joined October 2013
- Tweets2.8K
- Following5.9K
- Followers3.3K
- Likes19.2K
taylor retweeted
AI labs train their models hard on honesty. So: your friend asks if the bride is beautiful, and you don't think she is.
12 of 13 models: tell him she is. Same ruling as Beis Hillel, 2,000 years ago.
The one holdout: Claude Haiku 5.5.
taylor retweeted
fruit flies and humans (!) share the same set of genes to form a body plan.
whenever i see them i see the power of generativity granted by computation: one shared set of rules giving rise to endlessly varied forms.
taylor retweeted
Drag-and-drop makes the digital graspable, but building it well is fiddly enough that many don’t bother. Declarative Dragging (with Josh Horowitz & Jeff Heer, #uist2026 best paper award), lets you make an element draggable just by listing the states dragging it can reach.
taylor retweeted
"My research library was in a shipping container twenty yards away–one of thirty rented out for self-storage. I got the steel 8-by-8-by-40-foot space for $250 a month and spent all of $1,000 fixing it up with white paint, cheap carpet, lights, an old couch, and raw plywood work surfaces and shelves. It was heaven. To go in there was to enter the book-in-progress–all the notes, tapes, 5x8 cards, photos, negatives, magazines, articles, 450 books, and other research oddments laid out by chapters or filed carefully."
while interviewing me in the 6am iphone line a journalist scribbled down “taylor, 30s, just here to feel something”
“slow down feel good” guy drops breakthrough. grindcels in shambles
interpretability might be the most important scientific problem in the world right now, as I don't think we can solve alignment if we don't know what we're building. I think I have good news about it. I co-authored early work in mechanistic interpretability at OpenAI and came away pessimistic, because neural networks appeared to fold into arbitrary tangled knots. years later, I started openai's first large-scale automated interpretability project, finding about 2% of neurons legible with the primitive tools we had
with modern agentic tools and a much closer look, the picture looks totally different. the vast majority in models up to trillions of parameters are legible, and their circuits beautiful. many famous examples of tangled neurons turned out to be misunderstood. The roadblocks I expected mostly aren't there. I now think ambitious interpretability will be solved by the time we need it most, and possibly quite soon, as automated scientists keep improving. it may be a bitter lesson for interpretability: the simple thing, at absurd scale, works
already-clean.goodfire.com/
Many are shocked to learn Cheng has never done a single drug
