I mean to say, that if I have graph of llm nodes and I add another llm node, its not just multiplied benefits in generation because of the new logic, it's also multiplied headache of inexplicable responses. One small hallucination ruins the entire chain. And now we have more nodes so more chaos.

Resend

Nothing too sophisticated. I built local coding agents to see if I can automate some of my repeititive tasks, commit messages, and jsdoc commenter. With langgraph and claude. Occassionally random unsuable outputs. Adding the n+1 th LLM node to an existing system isn't an incremental effort IMO.

That is a good point. ultimately we are dealing with agents which are probabilistic and non deterministic and with more such agents in the pipeline, the error rate is hugely compounded. We still need a lot of supervision.

😂

Shared! 🔥

Added!

Amazing to see indie game industry apps. 🔥

Please add me!

I am currently doing client work while building my products and I am getting to the point where I want to be working on products full time.

Perfect. It’s perhaps a great time for you to prioritise both cash flow as well as freedom to build what you want.

Very nice!

Sheeeeesh.

Wow! That is crazy offer.

That is very interesting. Are you intentionally in a fractional CTO job so that you have more time to build something ?