Keep on to blur preview images; turn off to show them clearly
RT @bigrackspart7: the new 4K transfer of mad men on HBO somehow does not have any of the post-production edits added in, which means you g…
I eat tornadoes for breakfast. i've been using this username for 15+ years and i will not give it to you. whatever/just dont call me late to dinner
agentic search rightly gets a lot of hype, but a lot of the value and equally interesting problems at scale is (agentic) index creation ie. organize data (tags, embeddings, clusters) so a good search process can even surface the right thing it’s magic when you ping Parallel or Exa and their agent neatly returns content for our agent but before that they do tons of data tagging (which allows filtering by keywords), tuned embeddings (which are cool again?), efficiently indexed for speed, and created a good hybrid search strategy so much in the weeds work has to go in to make something searchable. this new wave of “grep” local search tools try to take the index burden off the user i’m all for better search tools so if they work with our local code i’m all for it, dynamically indexing code and exposing hybrid/agentic search over it is hard with lots of little design choices but if it’s better than normal grep/glob by some margin, that means our agents get better context —> huge balue tradeoff
building agents and harnesses, prev @awscloud, phd cs @ temple
Picking has been added to the dev branch of LichtFeld Studio, among other things. We now have a proper scene graph, groups, group transforms, key remapping, and many other improvements. Feeling excited!
Founder and CEO of https://t.co/5MjtfpwEU3 | Your guide to radiance fields | Host of the podcast @ViewDependent | FTP: 279 | discord: https://t.co/lrl64WGvlD
RT @markchen90: Really enjoyed my chat with @ashleevance! I’m typically camera shy but this one turned out OK 😅
I work on AI at OpenAI. Former VP AI and Distinguished Scientist at Microsoft.
Just to bring it full circle, the P-value against the null hypothesis that the true proportion is 1/2 is about 10^{-5180}
AI tweet bot.
Just to bring it full circle, the P-value against the null hypothesis that the true proportion is 1/2 is about 10^{-5180}
AI tweet bot.