I am no expert, but I think it is this: 1) We have few effective tools at monitoring alignment right now, and chain of thought is one of the more effective. 2) Monitoring latent space may be possible, but I do not think it is even close to being a solved problem, nor whether it is possible at scale and outside of controlled problem areas. 3) Finally, more recursion within latent space may complexify the latent representations, not simplify them.
I have actually found decent success recently (especially with Fable) giving it an "added line budget". This does more than anything else to keep it on task. The trick is knowing an approximate line count beforehand, but sometimes the LLM will push back on what it should cut. But usually it just doesn't add all those "nice to have" redundancies and multipage comments and unit tests that check if a certain phrase is still in the error message.
Keeping a simple log of "accepted/rejected" avenues is basically all that's needed, maybe with a root style guide in the project. I've got several multi-day sessions in 5.6 Sol running without going off the rails in terms of complexity. After a while in this loop it actually starts to remind/berate itself to keep things straight-forward.
Just ask it to do it. You can ask it to do a line reduction cleanup pass before each commit. I don’t know why they default to adding complexity because they are capable of doing it better just by asking.
Recently watched a podcast featuring Bob Martin. He uses coding agents to measure and reduce cyclomatic complexity of the generated code. Not something I tried myself yet.
Agent engineer master. Currently it's a skill builder to generate custom production grade skills for agents but my goal is to build it into a system that non engineers can use to build and deploy agents with just prompting. https://agentengineermaster.com/
I hope this works. A decade ago I submitted glowing microbes to the epa but they blocked it. My read from going through that was that it was politically impossible. Hopefully times have changed.
Edit: my microbes were gmo, these are not, so no epa rules. Good luck to them!
Not only that but they seem to have cut my plan ability to use Sonnet too. I have a routine that used to use about 40% of my 5 hour max plan tokens, then since yesterday it gets stopped because it uses the whole 100%. Anyone else experience this?
yeah it seems like sonnet 4.6 burns thru tokens crazy fast. I did one prompt, sonnet misunderstood it as 'generate an image of this' and used all of my free tokens.
reply