Hacker News: fluidcruft

New comment by fluidcruft in "LLMs corrupt your documents when you delegate"

fluidcruft — Sun, 10 May 2026 12:42:24 +0000

On the one hand there's simonw's concept of a "frequent LLM user" and then there's the actual vast majority of people using ChatGPT web app or one of the various Office CoPilots.

New comment by fluidcruft in "A recent experience with ChatGPT 5.5 Pro"

fluidcruft — Sun, 10 May 2026 12:13:21 +0000

I think it's because people walk every model up to its limits and become very aware of a task they can't make work. They do a lot of work simplifying and understanding limitations at that boundary. Then an improved model comes out and they immediately toe that barrier and make swift progress. They will also notice that the new model is natively doing tricks they had done manually.

The reality is likely that everyone is hitting similar barriers and the solutions are somewhat generalizable and get added to training new models.

Eventually people will reach the new limits and the cycle repeats.

New comment by fluidcruft in "Google broke reCAPTCHA for de-googled Android users"

fluidcruft — Sun, 10 May 2026 00:02:57 +0000

What they mean is those are the sites that will require attestation. It's pretty quaint to think that people who don't like bots would rather play wackamole with bots when they can just flip a switch and they're gone.

New comment by fluidcruft in "Google broke reCAPTCHA for de-googled Android users"

fluidcruft — Fri, 08 May 2026 23:30:54 +0000

Possibly... but the extension of this to Android and Apple is going to be the entire internet shuts you out. And everything else will be a giant Dead Internet crawling with bots.

New comment by fluidcruft in "Google broke reCAPTCHA for de-googled Android users"

fluidcruft — Fri, 08 May 2026 22:43:11 +0000

There really isn't much of an option. Apple's just as bad if not worse.

New comment by fluidcruft in "Embedded Rust or C firmware? Lessons from an industrial microcontroller use case"

fluidcruft — Sun, 03 May 2026 14:58:48 +0000

I think people don't like the JavaScript treadmill. People want to think about using tools and getting proficient with them rather than relearning tools. I'm not saying rust is like that, but I do feel that way about python and JavaScript. Those are dynamic languages but it is what all this editions stuff evokes. It's an if it were stable, it wouldn't be changing sort of thing.

New comment by fluidcruft in "I cancelled Claude: Token issues, declining quality, and poor support"

fluidcruft — Fri, 24 Apr 2026 21:36:35 +0000

It was probably the bug about cache getting purged after 5min rather than 1hour. You can review things pretty well within an hour. 5min is a real crunch. 5min doesn't mix with multitasking or getting interrupted.

New comment by fluidcruft in "I cancelled Claude: Token issues, declining quality, and poor support"

fluidcruft — Fri, 24 Apr 2026 17:10:34 +0000

I feel like they are making it harder to use it this way. Encouraging autonomous is one thing, but it really feels more like they are handicapping engaged use. I suspect it reflects their own development practices and needs.

New comment by fluidcruft in "I cancelled Claude: Token issues, declining quality, and poor support"

fluidcruft — Fri, 24 Apr 2026 17:09:01 +0000

My mental model for LLM is I don't expect them to chew gum and walk at the same time. Cleaning code up is a different task from building new functionality.

GLM always feels like it's doing things smarter, until you actually review the code. So you still need the build/prune cycle. That's my experience anyway.

New comment by fluidcruft in "I Cancelled Claude: Token Issues, Declining Quality, and Poor Support"

fluidcruft — Fri, 24 Apr 2026 16:48:12 +0000

It sounds like we have very similar usage/projects. codex had been essentially uncapped (via combination of different x-factors between Plus and Pro and promotions) until very recently when they copied Anthropic's notes.

Plus is still very usable for me though. I have not tried Claude Pro in quite a while and if people are complaining about usage limits I know it's going to be a bad time for me. I had to move up from Claude Pro when the weekly limits were introduced because it was too annoying to schedule my life around 5hr windows.

I started using codex around December when I started to worry I was becoming too dependent on Claude and need to encourage competition. codex wasn't particularly competitive with Claude until 5.4 but has grown on me.

The only thing I really care about is that whatever I'm using "just works" and doesn't hurt limits and Claude code has been flaky as all hell on multiple fronts ever since everyone discovered it during the Pentagon flap. So I tend to reach for ChatGPT and codex at the moment because it will "just work" and there's a good chance Claude will not.

New comment by fluidcruft in "GPT-5.5"

fluidcruft — Thu, 23 Apr 2026 21:00:15 +0000

It would be funny but not very flywheel so the one that gets there is more likely to get a gunner.

New comment by fluidcruft in "Bitwarden CLI Compromised in Ongoing Checkmarx Supply Chain Campaign"

fluidcruft — Thu, 23 Apr 2026 17:18:16 +0000

I seems like we need better standard libraries, but standard libraries turn into tarpits. I sort of like the way python's stdlib works.

New comment by fluidcruft in "Google Cloud customer wakes up to $18,000 bill despite $7 budget"

fluidcruft — Thu, 23 Apr 2026 05:26:35 +0000

It may have changed but there are (or used to be) gym memberships that would only let you cancel if you could prove that you moved and your new primary residence was more than one hour from the nearest branch. (I'm pretty sure there was a law passed to end this practice, but it was widespread)

New comment by fluidcruft in "Kernel code removals driven by LLM-created security reports"

fluidcruft — Wed, 22 Apr 2026 13:22:19 +0000

It's an interesting form of tree shaking.

The overlap of bugs being found, nobody caring enough to bother read the reports or fix the code, and nobody caring that the modules are pushed out of main seems good.

New comment by fluidcruft in "Claude Code to be removed from Anthropic's Pro plan?"

fluidcruft — Tue, 21 Apr 2026 23:35:46 +0000

I think the test is that new sign-ups won't have it and will the loss of five new Pro subscriptions be offset by more than one new Max subscribers.

Plenty of Pro subscribers never touch claude-code.

New comment by fluidcruft in "Claude Code to be removed from Anthropic's Pro plan?"

fluidcruft — Tue, 21 Apr 2026 23:32:57 +0000

How can you run the A/B test with mismatched documentation?

New comment by fluidcruft in "The creative software industry has declared war on Adobe"

fluidcruft — Sun, 19 Apr 2026 15:17:53 +0000

Yeah, theverge is subscription now.

New comment by fluidcruft in "Claude Opus 4.7"

fluidcruft — Thu, 16 Apr 2026 15:29:29 +0000

I generally think codex is doing well until I come in with my Opus sweep to clean it up. Claude just codes closer to the way my brain works. codex is great at finding numerical stability issues though and increasingly I like that it waits for an explicit push to start working. But talking to Claude Code the way I learned to talk to codex seems to work also so I think a lot of it is just learning curve (for me).

New comment by fluidcruft in "Ancient DNA reveals pervasive directional selection across West Eurasia [pdf]"

fluidcruft — Thu, 16 Apr 2026 14:41:38 +0000

The flip side is everything is being degraded by random mutation.

It's like holding a large ball in place on a hill that sees frequent tremors. If the ball is still halfway up the hill it's being held in place, if it's being held in place it's still halfway up the hill. It might be considered a tautology if you're only working with symbols and ignore all the mechanistics.

New comment by fluidcruft in "Google broke its promise to me – now ICE has my data"

fluidcruft — Wed, 15 Apr 2026 21:35:41 +0000

All communities have rules of behavior.