Post #3895445

2026-07-17 05:19 UTC

So, I've used Claude at work now for like 3 weeks? I use it a few times a day but some days not at all. General conclusion: it's still meh. It's still nowhere near as good as people keep saying it is. Helpful? Occasionally. Reliable? For easy stuff, it's alright. Otherwise, flip a coin. I truly do not understand how people can use it all day, every day. At some point, you're going to be doing an absolute shit job if you do that.

Replies (1)

@pointlessone@status.pointless.one 2026-07-17 21:14

@wwahammy@social.treehouse.systems I agree that it's mostly meh but also the output greatly depends on the model and effort. Mythos on max trends towards OK-ish more often than not. And it’s definitely noticeable compared to any Sonnet on mid. Though, it's extremely token hungry and max effort take forever to get through. A few times I got bored to the point that I just aborted it and went and written the thing myself.

Open ##3895444