Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Opus 4.5 is not the normal AI agent experience that I have had thus far

burkeholland.github.io

Post reply on HN
Print view View original

111–120 of 1001 posts

Re: Opus 4.5 is not the normal AI agent experience that I have had thus far

#111

Post by enum »

vercaemert wrote: post #106
enum wrote: post #103

Earlier quoted context omitted.

I teach at a university, and spend plenty of time programming for research and for fun. Like many others, I spent some time on the holidays trying to push the current generation of Cursor, Claude Code, and Codex as far as I could. (They're all very good.) I had an idea for something that I wanted, and in five scattered hours, I got it good enough to use. I'm thinking about it in a few different ways: 1. I estimate I…

How do you compare Claude Code to Cursor? I'm a Cursor user quietly watching the CC parade with curiosity. Personally, I haven't been able to give up the IDE experience.

I don't think I can scientifically compare the agents. As it is, you can use Opus / Codex in Cursor. The speed of Cursor composer-1 is phenomenal -- you can use it interactively for many tasks. There are also tasks that are not easier to describe in English, but you can tab through them.

Re: Opus 4.5 is not the normal AI agent experience that I have had thus far

#112

Post by manmal »

NitpickLawyer wrote: post #102
manmal wrote: post #32

IMO codex produces working code slowly, while Opus produces superficially working code quickly. I like using Opus to drive codex sessions and checking its output. Clawdbot is really good at that but a long running Claude Code session with codex as sub agents should work well also. The above is for vibe coding; for taking the wheel, I can only use Opus because I suck at prompting codex (it needs very specific instruct…

> I like using Opus to drive codex sessions and checking its output. Why not the other way around? Have the quick brown fox churn out code, and have codex review it, guide changes, and loop? I've actually gone one step further down the delegation. I use opus/gemini3 for plan, review, edit plan for a few steps. Then write it out to .md files. Then have GLM implement it (I got a cheap plan for like 28$ for a year on Ch…

I agree, codex is great at reviewing as well. I think that’s because code is the ideal description of what we want to achieve, and codex is good (only) when it knows what must be achieved, as verbosely as possible.

Currently I don’t let GLM or Opus near my codebases unsupervised because I’m convinced that the better the foundation, the better the end result will be. Is the first draft not pretty crappy with GLM?

Re: Opus 4.5 is not the normal AI agent experience that I have had thus far

#113

Post by cardine »

Papazsazsa wrote: post #12

"Opus 4.5 feels to me like" The article is fine opinion but at what point are we going to either: a) establish benchmarks that make sense and are reliable , or b) stop with the hypecycle stuff?

> make sense and are reliable

If you can figure out how to create benchmarks that make sense, are reliable, correlate strongly to business goals, and don't get immediately saturated or contorted once known, you are well on your way to becoming a billionaire.

Re: Opus 4.5 is not the normal AI agent experience that I have had thus far

#114

Post by oncallthrow »

Yoric wrote: post #88
OldGreenYodaGPT wrote: post #33

Most software engineers are seriously sleeping on how good LLM agents are right now, especially something like Claude Code. Once you’ve got Claude Code set up, you can point it at your codebase, have it learn your conventions, pull in best practices, and refine everything until it’s basically operating like a super-powered teammate. The real unlock is building a solid set of reusable “skills” plus a few agents for th…

You intrigue me. > have it learn your conventions, pull in best practices What do you mean by "have it learn your conventions"? Is there a way to somehow automatically extract your conventions and store it within CLAUDE.md? > For example, we have a custom UI library, and Claude Code has a skill that explains exactly how to use it. Same for how we write Storybooks, how we structure APIs, and basically how we want ever…

When I ask Claude to do something, it independently, without me even asking or instructing it to, searches the codebase to understand what the convention is.

I’ve even found it searching node_modules to find the API of non-public libraries.

Re: Opus 4.5 is not the normal AI agent experience that I have had thus far

#115

Post by manmal »

lawlessone wrote: post #21
delduca wrote: post #13

I agree, it wrote an entire NES emulator for me. https://news.ycombinator.com/item?id=46443767

It cloned one of the many open source ones available is what you mean.

As long as you give it deterministic goals / test criteria (compiles, lints, tests, E2E tests, achieve 100% parity with existing solution etc) it will brute force its way to a solution. Codex will work for hours/days, even weeks sometimes, until it has finished. A person would never work this way, but since this just runs in the background, there’s no issue with this approach except if you need it fast.

Re: Opus 4.5 is not the normal AI agent experience that I have had thus far

#116

Post by layer8 »

cyberpunk wrote: post #37
tannedNerd wrote: post #4

The problem with this is none of this is production quality. You haven’t done edge case testing for user mistakes, a security audit, or even just maintainability. Yes opus 4.5 seems great but most of the time it tries to vastly over complicate a solution. Its answer will be 10x harder to maintain and debug than the simpler solution a human would have created by thinking about the constraints of keeping code working.

You should try it with BEAM languages and the 'let it crash' style of programming. With pattern matching and process isolated per request you basically only need to code the happy path, and if garbage comes in you just let the process crash. Combined with the TDD plugin (bit of a hidden gem), you can absolutely write production level services this way.

Crashing is the good case. What people worry about is tacit data corruption, or other silently incorrect logic, in cases you didn’t explicitly test for.

Re: Opus 4.5 is not the normal AI agent experience that I have had thus far

#117

Post by fragmede »

yardie wrote: post #67

These are very simple utilities. I expect AI to be able to build them easily. Maybe in a few years it will be able to write a complete photo editor or CAD application from first principles.

Then we're really screwed!

Re: Opus 4.5 is not the normal AI agent experience that I have had thus far

#118

Post by blahblaher »

6177c40f wrote: post #60
hoten wrote: post #40

Earlier quoted context omitted.

Mind sharing the bill for all that?

Cheaper than hiring another developer, probably. My experience: for a few dollars I was able to extensively refactor a Python codebase in half a day. This otherwise would have taken multiple days of very tedious work.

And that's what the C-suite wants to know. Prepare yourself to be replaced in the not so distant future. Hope you have a good "nest" to support yourself when you're inevitably fired.

Re: Opus 4.5 is not the normal AI agent experience that I have had thus far

#119

Post by manmal »

waynenilsen wrote: post #9

Once you get your setup bulletproof such that you can have multiple agents running at the same time that can run unit tests and close their own loops things get even faster. However you accomplish that. Not as easy as it sounds mostly (and absurdly) due to port collision. E2E testing with playwright is another leap.

Just let it test in different containers? That’s not the hard part IMO.

Re: Opus 4.5 is not the normal AI agent experience that I have had thus far

#120

Post by manmal »

kachapopopow wrote: post #19

It's also the feeling I have, opus is not a ground-breaking model by any means. However, Opus 4.5 is incredible when you give it everything it needs, a direction, what you have versus what you want and it will make it work, really, it will work. The code might me ugly, undesirable, would only work for that one condition, but with futher prompting you can evolve it and produce something that you can be proud of. Opus…

Off/nearshoring regularly produces worse code. I’ve seen it first hand.
Post reply on HN