开启时会模糊预览图，关闭后正常显示
Important post beyond benchmarking: also explains why agentic model commoditization has not happened yet, despite new competitive releases in the open. Serving models you haven’t trained internally is just not trivial.
Artisanal baker of reasoning models @pleiasfr
hmm. since i’m on the look for symbolic logic frameworks, starting to wonder if julia is not undertapped.
Artisanal baker of reasoning models @pleiasfr
Maybe it’s me but being a mess of different historical periods feels pretty on brand for a Homeric adaptation.
Artisanal baker of reasoning models @pleiasfr
Was like the best time ever to do symbolic ai lock-in since the early 2010s but still lost to slopfest.
Artisanal baker of reasoning models @pleiasfr
Meanwhile @_vatsadev is still tackling the actual challenge (now with native int4 training) so wouldn’t exclude a viable 1.44Mb at some point.
Artisanal baker of reasoning models @pleiasfr
We’re cheating a bit so it will be Zip disks with Baguettotron (one possible cover).
Meanwhile @_vatsadev is still tackling the actual challenge (now with native int4 training) so wouldn’t exclude a viable 1.44Mb at some point.