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

91–100 of 203 posts

Re: Scaling long-running autonomous coding

#91

Post by pleurotus »

qingcharles wrote: post #76
gordonhart wrote: post #24

Earlier quoted context omitted.

Web browsers are insanely hard to get right, that’s why there are only ~3 decent implementations out there currently.

The one nice thing about web browsers is that they have a reasonably formalized specification set and a huge array of tests that can be used. So this makes them a fairly unique proposition ideally suited to AI construction.

As far as I read on Ladybird's blog updates, the issue is less the formalised specs, and more that other browsers break the specs, so websites adjust, so you need to take the non-compliance to specs into account with your design

Re: Scaling long-running autonomous coding

#94

Post by orlp »

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…

> Long-running projects that converge on high-quality, complex projects

In my experience agents don't converge on anything. They diverge into low-quality monstrosities which at some point become entirely unusable.

Re: Scaling long-running autonomous coding

#95

Post by kungfuscious »

bonesss wrote: post #88
snek_case wrote: post #68

Earlier quoted context omitted.

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…

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.

Re: Scaling long-running autonomous coding

#96

Post by gaigalas »

There's a clear conflict between SKILLS, tools and multi-tasking.

I think "intra-context" tooling is already dead. It's too narrow.

It's all "extra-context" now: how one instruments for multiple agents, at multiple times, handling things.

Personally, I think the best tool in this realm will come from open source, and be agnostic (many agents from many places interacting), in order to leverage differences between subtle provider qualities (speed, price and so on).

Building a browser is an interesting and expensive experiment. How much did it cost?

Re: Scaling long-running autonomous coding

#97

Post by rlt »

xmprt wrote: post #12
cheevly wrote: post #9

Earlier quoted context omitted.

2029? I have no idea why you would think this is so far off. More like Q2 2026.

You're either overestimating the capabilities of current AI models or underestimating the complexity of building a web browser. There are tons of tiny edge cases and standards to comply with where implementing one standard will break 3 others if not done carefully. AI can't do that right now.

Not only edge cases and standards, but also tons of performance optimizations.

Re: Scaling long-running autonomous coding

#99

Post by torginus »

I'm kinda surprised how negative and skeptical anyone is here.

It kinda blows my mind that this is possible, to build a browser engine that approximates a somewhat working website renderer.

Even if we take the most pessimistic interpretation of events ( heavy human steering, relies on existing libraries, sloppy code quality at places, not all versions compile etc)

Post reply on HN