I assume it's deliberate - you're not supposed to know what it's doing. It's a black box that either completes the task or spins forever trying.
replyI agree. I was completely sold on Claude models for a year. 4.6 vs OpenAI codex in same period? It was night and day. Opus I could talk to about api design, tradeoffs, etc. codex was mechanical, used “load bearing” constantly, and unsettling brief.
replyNow it’s flipped. Sol emits thoughts as it works, which help as I’m scrolling through and see it’s made a bad assumption. It can be directed but still push back. Opus? It’s seems to inherit the unsettling silence of Fable and waits till the end to give you its authoritative “here’s how it is. I even end up having 4.6 “translate” what it says back to English. I hate having to instruct an llm to “talk to me”.
Yes there’s ways of getting it to talk more plainly, “don’t overwhelm me”, not be as nit picky and anxious “we are bold and fearless”. But didn’t have to do that before.
Think of these status updates as progress spinners.
replyIt wasn't a status update, I asked it to explain something to me.
replyWhat was your question?
reply"Please look at this repo and give me a high-level explanation of how the app works", more or less.
replyAnd all it replied with was what you wrote ("The loop Write. A file, applied. Properties go under data.properties, never on data:")?
replyOr was there more to the response?
LOL, is it trying to speak in Haikus?
replyPain. Sufferance. Inevitable is, the Yodaization of LLM output.
replySpike. Applied it has been.
replyThat's definitely what it feels like.
replyA few generations from now, everyone will talk like a beat poet. Jazz speak.
reply