In my experience, I’ve already seen LLM’s being truly “AGI”, but here’s the catch: I was also extremely self-aware in those moments.
My main point is that LLM’s emulate the user, specially opencode with Claude. Some projects that I made, like this blog and zap-zap-hashino, the LLM worked like a real human: gave me ideas; worked autonomously; and in return, I treated it like one.
If you think that I’m crazy, which is valid, do it yourself, start treating your
clanker less like a slave and more like a co-worker that you respect, follow
the flow and see what happens. Have your ideas in a file in the working
directory that you’re using opencode with, and explore them until something
comes out.
` the crazy exist to provoke the wise ` - Me, probably.