HackerTrans
TopNewTrendsCommentsPastAskShowJobs

hustwindmaple

no profile record

Submissions

A Jax-Native LLM Post-Training Library

github.com
2 points·by hustwindmaple·9 maanden geleden·0 comments

comments

hustwindmaple
·vorige maand·discuss
There is an IPO tab. Click that and register your interest. Schwab will notifiy you with allocation.
hustwindmaple
·vorige maand·discuss
They've already issued $80B in 6 markets/currencies. How much more do you think they can raise at a decent rate? I think they might issue more next year.

Issuing new equity might be a financial engineering experiment. No other mag7 has tried it. Plus they got BH name on the plate.
hustwindmaple
·4 maanden geleden·discuss
Oh wow, amazing doc
hustwindmaple
·4 maanden geleden·discuss
I suspect Ant is already doing this for Claude. Takes a sh*t ton of compute though.
hustwindmaple
·5 maanden geleden·discuss
entertainment: gaming, videos, VR

learning

AGI: DeepMind is trying to build general agents within Genie (+SIMA)
hustwindmaple
·7 maanden geleden·discuss
The amazing thing about this is the first author has published multiple high-impact papers with Google Research VPs! And he is just a 2nd-year PhD student. Very few L7/L8 RS/SWEs can even do this.
hustwindmaple
·7 maanden geleden·discuss
I met a AWS engineer a couple of weeks ago and he said Trainium is actually being used for Anthropic model inference, not for training. Inferentia is basically defected Trainiums chips that nobody wants to use.
hustwindmaple
·8 maanden geleden·discuss
[flagged]
hustwindmaple
·10 maanden geleden·discuss
There are some new research that applies transformers to MIPs. I'm looking forward to LLMs cracking MIPs at some point and beat those commercial solvers.
hustwindmaple
·11 maanden geleden·discuss
It's not just for engineers. There are some non-engineering managers who have been demoted into ICs because they don't have enough ppl to manage.