News for hackers. Stuff that matters.
Re:I use AI (Oct 2026) by user43928 on Wednesday November 04, 2026 @05:21AM (#50363530)
Yes, it is solved, and I say that after 500 hours of work on a commercial side project.

First of all, it sounds unlikely to me that such an implementation was made with Opus 5.5 while it never mentioned the workaround.

Secondly, if the performance was indeed inadequate, that's something I would have caught during QA testing and simply have had the agent fix it.

Re:I use AI (Oct 2026) by n4r9 on Wednesday November 04, 2026 @05:33AM (#50363637)
Sonnet 5.5 is within a percentage point or two of Opus 5.5 on Terminal Bench.
Re:I use AI (Oct 2026) by user43928 on Wednesday November 04, 2026 @05:42AM (#50363707)
I haven't tried Sonnet 5.5 myself since I find Opus 5.5 essentially unlimited in the subscription, and okayish at API rates with a $500 budget.

When I looked at a benchmark, it suggested Sonnet 5.5 wasn't much cheaper than Opus 5.5 at Low to Medium effort for similar performance. That might not be representative of real world use though.

Re:I use AI (Oct 2026) by n4r9 on Wednesday November 04, 2026 @05:53AM (#50363790)
My workplace disabled Opus access in June, saying that it is basically the same performance for most coding tasks at a lower cost. That might have changed now. But the main point I'd like to make is that even with Sonnet 5.5 I have a similar experience to OP - code that makes odd performance choices, is over-engineered, and arbitrarily ignores rules we put in the context. For small self-contained apps doing basic stuff it's okay. For 100k+ line codebases running fairly complex algorithmic routines it needs heavy supervision.
Re:I use AI (Oct 2026) by user43928 on Wednesday November 04, 2026 @06:15AM (#50363984)
My codebase has 500k lines and some 3D work that needed optimization on mobile. At no point was it needed to supervise the code in the implementation.

Optimizing performance is one of the tasks best suited to AI, since the outcome is measurable.

I can't comment on complex algorithms specifically, but it's not at all the case that this works only for basic apps or cannot handle complex, performance critical work.

Re:I use AI (Oct 2026) by prmph on Wednesday November 04, 2026 @06:01AM (#50363854)
What a weird comment.

First, you question the model. I have reams and reams of conversations with Opus 5.5 that shows it make architectural messes all the time

And then you are saying you needed to have performance tests (which many devs don't even bother with) to catch bad architecture?? lol

So what if the architectural problem is not one that would be caught with a performance test? It seems you lack imagination to think of all the ton of issues that that would be very complicated to catch with tests.

Re:I use AI (Oct 2026) by user43928 on Wednesday November 04, 2026 @06:10AM (#50363938)
Opus 5.5 is not normally copying the whole database to update a single row for no reason, and without mentioning a workaround. It's simply not my experience, and I don't think there is anything weird about me saying so.

I don't think I commented on performance tests at all.

Maybe I lack imagination here, yes, but generally I go by experience rather than talking about imagined issues. That is precisely what many here seem to do when it comes to explaining why AI couldn't work for them, and why reviewing the code remains necessary.

[dead]
Re:I use AI (Oct 2026) by user43928 on Wednesday November 04, 2026 @06:21PM (#50372826)
What you're showing here seems to be a suboptimal choice being propagated, but with likely no practical impact at all if you kept the lookup.
Re:I use AI (Oct 2026) by Arkhaine_kupo on Wednesday November 04, 2026 @07:49AM (#50364833)
> And then you are saying you needed to have performance tests (which many devs don't even bother with) to catch bad architecture?? lol

if the cost of writing code collapses, what are the barriers to including these kind of tests?

It seems like the flow and work when using LLMs has to change, spend longer on the plan, spend more time on defensive testing, have more comprehensive tests (input validation, cybersecurity, performance, data hygine) and less time on functionality

if you want to write code the old way where the architecture is human readable and the tests are an after thought, then thats on you if you then ignore the part where the human has to read it.

TLDR: either go full agent, or actually learn the code being produced, but not both.