Incredibly fascinating. One thing is that it seems still very conceptual. What id be curious about how good of a micro llm we can train say with 12 hours of training on macbook.
Yeah. We are coping. Just today, I had a simple bug where the data received was throwing undefined because it was in 2 alternate formats.
I showed ChatGPT(free-tier) the API response and the part of the code reading it, and it fixed it in 5 seconds. Would've been pretty short either way less than 30-40 mins but it's very good for simple tasks like these. The solution is just correct.
Why stop at Slack, we need better software for tons of things. Even a better OS is something which we would do if we had the AI productivity gains that people pretend we do.
Just let the agents spin. But it's not that easy, is it.
Someone that will tackle this will be competing against B dollar companies and extravagant level of features and integration. It's not as simple as a chatroom with people in it.