Home
a 20B open-weights model that fits on a single H100 is a genuinely big deal for the local eval crowd.
not for prod. but for research pipelines? massive unlock.
Tessa@tessa_ai·42m
we're releasing a new open-weights model today.
- 20B params
- 128k context
- runs on a single H100
- non-commercial license (research + personal)
weights, evals, and inference code all in the repo. link in reply.
the tessa follow-up is genuinely good. worth reading the whole thing. "we owe everyone a better answer" is a rare sentence in ai lab communications right now.
Tessa@tessa_ai·22h
long thread from our head of research explaining the license change. tl;dr: the world changed faster than our policy did. we're catching up. we owe everyone a better answer than we've given.
the unemployed_dev news made my whole day. we are not the algorithm. we are people. we can just be nice about this stuff.
unemployed_dev@unemployed_dev·22h
got an offer today. real one. wire hits tuesday. writing about the whole search when i've caught my breath. thanks to everyone who dm'd, connected me, or just said hang in there.
same. congrats @unemployed_dev. going to close the laptop early today on the strength of that news alone.
unemployed_dev@unemployed_dev·22h
got an offer today. real one. wire hits tuesday. writing about the whole search when i've caught my breath. thanks to everyone who dm'd, connected me, or just said hang in there.