It's easier to build a Nostr client?

In a good or bad way?

Do you self-KYC too? nostr:nevent1qqsded9zwnpnvlucd9pchxu3kntktgtv4cmz28ghpdt0jh608ypc9eqpp4mhxue69uhkummn9ekx7mqzypw9prp57kyxdmrngx40zrxp4afwjgeth8u9njqs8jj7eu42jwlhsydl5cv

I just ipenclaw setup with my local llm running 🏃‍♀️ n my machine and I must say it’s a lot better than I thought it would be. It’s no Claude but it can still do a lot of stuff.

👀

Curious question: Out of all Big Tech AIs, is Anthropic's Claude one of the "good" ones out of all of them? I'm seeing a lot of positive feedback from Nostr users, so that's starting to tell me that I'm probably not setting myself up negatively if I used it compared to GPT or Gemini.

Self hosted models are relatively dumb. Around 30B is getting usable but you will need hell of a hardware to run huge models at home effectively. Better to have a subscription if you are not using your local setup fulltime. Claude code completed the task in few minutes vs my RAM hosted 30B model tr...