hanska

6 ms·

1.
by hanska 9mo ago
yes if you tell llm to do things it will be too verbose. either explicitly instruct the length ("add 5 lines bulletpoints, tldr format") or just write it yourself.
2.
by hanska 1y ago
koreans knew after all what is up https://en.wikipedia.org/wiki/Fan_death
3.
by hanska 1y ago
it's just considered marketing cost. they might not make $ from Bob, but they increase their client base by 1 any non-referral users, SweepIQ still keeps their amazon profit (and uses that to pay for the gpu's) + in future they mi
4.
by hanska 1y ago
The Groq interview was good too. Seems that the thought process is that companies like Groq/Cerebras can run the inference, and companies like Nvidia can keep/focus on their highly lucrative pretraining business. https:/&#x2
5.
by hanska 1y ago
i've tried it, it works, it's nice :) got a mini mmo running with several people at once running around
6.
by hanska 2y ago
not anymore! with this tool, we can get rid of all of that fluff and go straight back to pure content only.
7.
by hanska 2y ago
"The only reason I can think of to buy a higher expense fund when a much lower expense variant exists is in the name of tax loss harvesting" perhaps it's because they receive kickbacks from that specific fund.
8.
by hanska 3y ago
not common knowledge I think. all usd swift wires get routed through 2 of the major banks - think it was JP Morgan Chase & Wells Fargo. this allows them to check all of these transfers for aml/cft and other sanctions blacklists. he