Codeberg's philosophy sounds really good, I build open source scientific research projects, would you advise me using this platform instead of Github for making my code public?
If I understand correctly the conversation, ChatGPT had to compute for quite some time, meaning a large amount of computations. What kind of resources would be we needed to achieve the same results with a local LLM? Is it even feasible with current open models?
For instance, would it be affordable for a research lab to not rely on OpenAI?
Thanks for the references, I'll have a look. I can tell you what I don't want: books that throw theorems one after the other without any context, like Baby Rudin [1] for instance.
I tried once to read Terence Tao's Analysis I [2], it's really good but my main problem was that I wasn't able to know if my proofs are correct when I do an exercise. So maybe the solution is to get a teacher.
I'm looking for mathematics books that take the time to explain with words and sentences what is actually going on when they introduce a new theorem, something that focuses on meaning.
Honestly I understand OP, I also run GrapheneOS but most of my social life happens on whatsapp and I know it's stupid but I can't change dozens of people's choice to install Signal, even though most of them know the issues with Meta.
My hope is that one day or another whatsapp will enshittify so bad that people will be more prone to move to Signal in the same way that recently Windows got so bad that many users moved to Linux.