I just keep thinking about the first guy to put on the frog costume reading the news today and seeing frogs literally all over the world

what a fucking thing, man

that 25 year old did more for the country in a week than most of the NYT op-ed section in a lifetime

Wope! I can’t see anything wrong. 🤷

I’ve learned all sorts of commands and tool incantations using AI code assistance. Another benefit of (very carefully) using coding agents.

Yep.

But I’d be wary of overdoing it with committees. KISS applies here.

Too much structure/hierarchy can be just as as bad as BDFL.

I think you are right. Aside from a high profile few, BDFL projects, in my experience, don’t fare well long term. It’s an exclusionary model.

IMO community/consensus-driven models work better, but require significantly more effort in terms of community development.

📌

Thanks! I’ll give it a try.

Is Max Headroom an option? 🤣

True. But they can run small #golang services like a champ. Python, in contrast, has enough performance baggage to warrant a larger instance.

It amazes me how many companies depend upon a display format (PDF) for data interchange. It’s Y2K-level technical debt without a deadline.

We’re kind of enterprise focused. Seeing a lot of use cases around automated invoice reconciliation, tabular data extraction (from PDF 🙄), and autonomous “Worker” agents.

The emoji got appended to the URL 😂 and didn’t work for me. Updated link:

https://www.langchain.com/stateofaiagents