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

Scaling long-running autonomous coding

cursor.com

Post reply on HN
Print view View original

101–110 of 203 posts

Re: Scaling long-running autonomous coding

#101

Post by viraptor »

hahahahhaah wrote: post #90
simonw wrote: post #3

"To test this system, we pointed it at an ambitious goal: building a web browser from scratch." I shared my LLM predictions last week, and one of them was that by 2029 "Someone will build a new browser using mainly AI-assisted coding and it won’t even be a surprise" https://simonwillison.net/2026/Jan/8/llm-predictions-for-202... and https://www.youtube.com/watch?v=lVDhQMiAbR8&t=3913s This project from Cursor is the s…

Web browser should be easy as source exists. Fix all SVG bugs in my browser tho...

There are 3.5 serious open codebases of web browsers currently. Only two are full featured. It's not nothing, but it's very far from "source exists so it's easy to copy what they do".

Re: Scaling long-running autonomous coding

#102

Post by torginus »

Personally what I don't like about this now that I think about it, is that they didn't scale up gradually, let's say there there's a ladder of complexity in software, starting at a simple React CRUD app, going on to something more complex, such as a Paint clone, to something even more complex, like a file manager etc, ending up at one of the most complex pieces of software ever made, a web browser.

I'd want to see some system, that 100%s the first task, saturation, does a great job on the next, then does a valiant effort on the third, then finally makes something promising but as yet unusable on the last.

This way we could see that scaling up difficulty results in a gradual decline in quality, and could have a decent measurement of where we are at and where we are going.

Re: Scaling long-running autonomous coding

#103

Post by viraptor »

trjordan wrote: post #8

This is going to sound sarcastic, but I mean this fully: why haven't they merged that PR. The implied future here is _unreal cool_. Swarms of coding agents that can build anything, with little oversight. Long-running projects that converge on high-quality, complex projects. But the examples feel thin. Web browsers, Excel, and Windows 7 exist, and they specifically exist in the LLM's training sets. The closest to real…

> Web browsers, Excel, and Windows 7 exist, and they specifically exist in the LLM's training sets.

There's just a bit over 3 browsers, 1 serious excel-like and small part of windows user side. That's really not enough for training for replicating those specific tasks.

Re: Scaling long-running autonomous coding

#104

Post by NitpickLawyer »

kungfuscious wrote: post #95
bonesss wrote: post #88

Earlier quoted context omitted.

So the chain of events here is: copy existing tutorials and public/available code, train the model to spit it out-ish when asked, a mature-ish specification is used, and now they jitter and jumble towards a facsimile of a junior copy paste outsourcing nightmare they can’t maintain (creating exciting liabilities for all parties involved). I can’t shake the feeling that simply being a shameless about copy-paste (ie cop…

You raise a good point, which is that autonomous coding needs to be benchmarked on designs/challenges where the exact thing being built isn't part of the model's training set.

swe-REbench does this. They gather real issues from github repos on a ~monthly basis, and test the models. On their leaderboard you can use a slider to select issues created after a model was released, and see the stats. It works for open models, a bit uncertain on closed models. Not perfect, but best we have for this idea.

Re: Scaling long-running autonomous coding

#105

Post by throwaway63467 »

I‘m running opus 4.5 which is arguably their best model and while it’s really good for a lot of work it always introduces subtle errors or inconsistencies when left unsupervised as prompts are never good enough to remove all ambiguity for complex asks, so I can’t imagine what it will do to a code base when left alone with it for days or weeks.

Re: Scaling long-running autonomous coding

#108

Post by azan_ »

observationist wrote: post #6
jphoward wrote: post #2

The browser it built, obviously the context window of the entire project is huge. They mention loads of parallel agents in the blog post, so I guess each agent is given a module to work on, and some tests? And then a 'manager' agent plugs this in without reading the code? Otherwise I can't see how, even with ChatGPT 5.2/Gemini 3, you could do this otherwise? In retrospect it seems an obvious approach and akin to how…

Get a good "project manager" agents.md and it changes the whole approach of vibe coding. For a professional environment, with each person given a little domain, arranged in the usual hierarchy of your coding team, truly amazing things can get done. Presumably the security and validation of code still needs work, I haven't read anything that indicates those are solved yet, so people still need to read and understand t…

Could you perhaps share such agents.md? Sounds interesting

Re: Scaling long-running autonomous coding

#109

Post by embedding-shape »

snek_case wrote: post #68
embedding-shape wrote: post #17

Did anyone manage to run the tests from the repository itself? The code seems filled with errors and warnings, as far as I can tell none of them because of the platform I'm on (Linux). I went and looked at the Action workflow history for some pages, and seems CI been failing for a while, PRs also all been failing CI but merged. How exactly was this verified to be something to be used as an successful example, or am I…

I found the codebase very hard to navigate. Hundreds (over a thousand?) tiny files with less than 200 lines of code, in deeply nested subdirectories. I wanted to find where the JavaScript engine was, and where the DOM implementation was located, and I couldn't easily find it, even using the GitHub search feature. I'm not exactly sure what this browser implements and how. Even their README is kind of crappy. Ideally y…

> It's impressive that they got such a big project to be built by agents and to compile

But that's the thing, it doesn't compile, has a ton of errors, CI seems broken since long... What exactly is supposed to impressive here, that it managed to generate a bunch of code that doesn't even compile?

What in the holy hackers is this even about? Am I missing something obvious here? How is this news?

Re: Scaling long-running autonomous coding

#110

Post by embedding-shape »

mrefish wrote: post #19
simonw wrote: post #3

"To test this system, we pointed it at an ambitious goal: building a web browser from scratch." I shared my LLM predictions last week, and one of them was that by 2029 "Someone will build a new browser using mainly AI-assisted coding and it won’t even be a surprise" https://simonwillison.net/2026/Jan/8/llm-predictions-for-202... and https://www.youtube.com/watch?v=lVDhQMiAbR8&t=3913s This project from Cursor is the s…

Time to raise the bar. By 2029 someone will build a new browser using mainly AI-assisted coding and the surprise is that it was designed to be used by pelicans.

> Time to raise the bar

Lets make someone pass the one we have, this experiment didn't seem to yield a functioning browser, why would we raise the bar?

Post reply on HN