开启时会模糊预览图，关闭后正常显示
RT @heyanuja: navier stokes describes how fluids move the millennium prize asks, can a fluid's energy concentrate into a single point in fi…
Root node of the web of threads: https://t.co/ifH80GcLpo
RT @heyanuja: I made a Goodreads for academic papers! (..and blog posts, substacks, lesswrong, etc) Paper Trails [https://t.co/dPzhlBgo8E]…
Root node of the web of threads: https://t.co/ifH80GcLpo
My AI prediction for 2026 is that the amount of alpha in ignoring 90% of all AI chatter will only increase as the benchmark numbers continue to get further untethered from the hard reality of incorporating LLMs into real features for real customers who are doing real work.
Author. Coder. CTO. θηριομάχης. Building: https://t.co/otXT4Wy6WR. Writing: https://t.co/dBPBtyCIHw.
As a builder in AI, there has just been so much alpha in ignoring AI hype & vagueposting & manic FOMO & threadbois, and treating LLMs as one tool in the toolbox & as yet another REST API that I can call & get a useful response from.
My AI prediction for 2026 is that the amount of alpha in ignoring 90% of all AI chatter will only increase as the benchmark numbers continue to get further untethered from the hard reality of incorporating LLMs into real features for real customers who are doing real work.
In the new year, Symbolic will Big News -- watch this space. At that time, I will take a little mini victory lap, & I'll discuss some of the approach that let us achieve what we'll announce. But to scoop myself, my own secret edge as a CTO is: I just don't believe the AI hype 🤷♂️
I don't believe the hype about "AI" repos with a bazillion github stars. I clone into them & realize there isn't much real work there, & anything that is interesting in there I can just lift & put into my own app w/out adding a dependency.
RT @heyanuja: the local enstrophy function is defined as 0 with a comment "placeholder value; properties axiomatized" then the next lines…
Root node of the web of threads: https://t.co/ifH80GcLpo