I can see this future happening!

Cool

I’m confused by this comment

I’ve had similar results with multi-agent approach.

It consumes lots of tokens, required more setup and at the end of the day had pretty much the same output as if I had used a single agent.

You have any concrete examples of source code that your multi-agent approach has produced?

What are the signals or tells?

Comments like "X is the right track [...] Then finish with a question?" do have a bit of an LLM smell to them.

The finishing with a question thing is prevalent with both accounts on Twitter, presumably because it "drives engagement" with the accounts.

It's particularly frustrating because it amplifies how much time is wasted - people don't just waste time reading comments by bots, they then invest effort in thinking about and replying to them.

How can you tell?

I like it. I think I’ll eventually take a whack at making something similar. Thanks for sharing

It’s brilliant to see this type of research coming out.

Thankfully there is still tumblr