News for hackers. Stuff that matters.
Re:NYC's "Click to Cancel" rule is now in effect by troupo on Monday October 12, 2026 @10:19AM (#50056838)
lol. AI was never set to subsume products. The marketing around AI pretends it is.
Re:NYC's "Click to Cancel" rule is now in effect by TeMPOraL on Monday October 12, 2026 @10:47AM (#50057205)
Sure it is. That's literally the main use of it anyone has at this point.
Re:NYC's "Click to Cancel" rule is now in effect by troupo on Monday October 12, 2026 @10:56AM (#50057344)
People in tech should learn to talk to real people at least once in their life.

Or to actually try and maintain/change/fix vibe-coded software for longer than a few days.

No, AI isn't subsuming products. Not now, not in the forseeable future.

Re:NYC's "Click to Cancel" rule is now in effect by TeMPOraL on Monday October 12, 2026 @11:49AM (#50058168)
Sure it is. I know of e.g. people doing graphics and copywriting that barely even open Canva / Figma or Word anymore, because they learned how to get 90% of the work done directly inside ChatGPT. And it's not low-quality work, it's just the same or better work, done much more efficiently, and without the need of some previously-used products.

And software developer's perspective on this is exactly how future looks, because for me, Claude Code now manages my personal computer fully - installs new software as needed, and uses it as needed.

Re:NYC's "Click to Cancel" rule is now in effect by troupo on Monday October 12, 2026 @05:30PM (#50063313)
> And it's not low-quality work, it's just the same or better work, done much more efficiently, and without the need of some previously-used products

Oh, I highly doubt this. Since I use ChatGPT for generating images quite often. There's no chance in hell they are doing consistent high-quality work in ChatGPT 90% of the time. Or your (and their) definition of high quality work is far from a common definition.

This is especially true in the past few months when OpenAI removed any way to directly prompt their image generator. So ChatGPT is often stuck in the same generator seed, the generator never gets the full prompt, regenerates images (inconsistently), messes up details, ignores input (because the model never generates the prompt to the image generating model correctly) etc. etc.

Good enough to generate a restaurant slop-menu or a quick TTRPG image. Consistent high quality? Lol.