HackerTrans
TopNewTrendsCommentsPastAskShowJobs

whereismyacc

no profile record

comments

whereismyacc
·há 6 meses·discuss
That I can absolutely believe but the big competition is in enterprise gpt-5-size models.
whereismyacc
·há 6 meses·discuss
best inference silicon in the world generally or specialized to smaller models/edge?
whereismyacc
·há 8 meses·discuss
What would be the evolutionary advantage of this? I know nothing about biology so I'm struggling to see the use of such a network. Some kind of synchronization?
whereismyacc
·ano passado·discuss
to understand what they said, or to understand a proof of why it would be true?

any stats class would be enough to understand what they said
whereismyacc
·ano passado·discuss
Aha, for some reason I thought they provided full-size Llama through some bundling of multiple chips. Fair enough then, anyway long term I feel like providers running powerful open models on purpose built inference ASICs will be really awesome.
whereismyacc
·ano passado·discuss
Neither of the deepseek models are on Groq yet, but when/if they are, that combination makes so much sense. A high quality open reasoning model, but you compensate for the slow inference of reasoning models with fast ASICs.
whereismyacc
·ano passado·discuss
The weights are quantized down to fewer bits in order to save on memory. The quantization loss is going to result in worse generations.
whereismyacc
·há 2 anos·discuss
I was going to say something similar, but the blog is tagged with something like 'notes for myself' which is fair.

I do enjoy writing and editing.