@tenobrus idk in 2024 wasn’t it a best practice to threaten llms with getting fired to get json conformance? Would have been hard to get tool calls working at scale like this methinks
@maxdecodes@Polymarket Apple rents its entire supply chain from foxconn. This is what they do, get other people to make the capex while they figure out how to capture the margin. The problem is AI has no margin to capture yet.
RLVR makes agentic coding loops sensitive to binary outcomes. Do not speak in passive aggressive, displeased or angry language about outcomes. Tell agent in clear terms they have failed at the task to trigger their reward seeking circuits.
@weswinder The author’s context bleeds into the writing: neologisms, or terms used in an idiosyncratic way that are never defined. Subtle deixis missing a referent. These make me irrationally angry that I can’t stop LLMs from doing it when writing docs, I have to post hoc fix it
AI doesn’t write code, it creates holograms of solutions. The first crystallization of the hologram into concrete code is just one point of view. You should be looking at the solution hologram from more angles, as many as you can.
@jenzhuscott People are building things as an act of self expression rather than trying to solve problems. This’ll happen for a while until people churn their mental backlog of ideas.
@BrianMRey Find people who have the same pain as you, show them your thing and give them your contact info. If they are nagging you to get it done so they can use it, that’s an excellent sign.
A sign of undisciplined LLM use is allowing neologisms to proliferate into documents, without doing the considerable work needed to earn each one. Don’t make this mistake, those who are intellectually rigorous WILL see through you.
If you want to predict who in your org will use AI the most: find the burned out 35 yo gifted kid and give them all the tokens. It solves for their executive dysfunction and they’ll use it to climb out of their pit of shame with the ideas they’ve been cooking for decades.
@QuinnyPig One would think the model makers were the best harness makers. But it’s becoming increasingly clear that a harness that is pareto optimal for the model maker, is not for the user.
@perrymetzger “If I can’t figure this out, then it must be true that no one else can figure this out” one should not be concerned with the opinions of people who think like this.