Matthew Di Ferrante
@mattdf.bsky.social
software engineer, vuln finder // into machine learning, formal methods, cryptography, pure maths, finance, game theory // links/contact: https://mattdf.com // twitter: @matthewdif
Reposted by Matthew Di Ferrante
Took longer than anticipated, but good riddance :)
May 3, 2023 at 10:01 AM
Took longer than anticipated, but good riddance :)
hope that within another 3-6 months we get an open source LLM that is actually on par with GPT-3.5, given the recent pace of progress on finetuned LLaMa models that seems realistic.
April 30, 2023 at 11:06 PM
hope that within another 3-6 months we get an open source LLM that is actually on par with GPT-3.5, given the recent pace of progress on finetuned LLaMa models that seems realistic.
glad i held out and never bothered to join mastoshit because this feels like it will finally make that terrible protocol irrelevant
April 13, 2023 at 6:02 AM
glad i held out and never bothered to join mastoshit because this feels like it will finally make that terrible protocol irrelevant