@JustMicrock I don't watch TikTok myself but I'm quite fond of the people that try to trick doom scrollers into learning some new trivia; I think it's much better than total brainrot, so I don't know what they're complaining about
I think Claude has done a pretty good job
Last month, we demonstrated high-rate alchemy: turning mercury into real, stable gold, beating CERN’s record by a factor of 10.
Last week, we enriched isotopes in the world’s most powerful plasma centrifuge, beating the Soviet record that stood for 40 years.
Now, we’re building a fusion reactor in San Francisco to achieve record power density.
Another example of things that I believe are only possible in C++. This is a hand-written inner matmul loop.
It compiles 4x (fp16, fp32, for matrices with size matching SIMD width and not). It targets WASM SIMD, Neon, SSE2/AVX/AVX2/AVX512.
There's no branching in the inner loop (e.g. to check for tail). UchenLoad* is template magic that does load for fp32 and load/promote for fp16.
@chiefofautism It runs perfectly on this hardware, it is really smart, and it can do tool calls so you can use it as an agent
I run 4bit quant at max context and if you have a 12GB GPU you can bump that up to 5bit and add multimodal projections as well