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

Scaling long-running autonomous coding

simonwillison.net

Post reply on HN
Print view View original

21–30 of 112 posts

Re: Scaling long-running autonomous coding

#21

Post by mrbungie »

ramraj07 wrote: post #16
halfcat wrote: post #10

So AI makes it cheaper to remix anything already-seen, or anything with a stable pattern, if you’re willing to throw enough resources at it. AI makes it cheap (eventually almost free) to traverse the already-discovered and reach the edge of uncharted territory. If we think of a sphere, where we start at the center, and the surface is the edge of uncharted territory, then AI lets you move instantly to the surface. If…

The fundamental idea that modern LLMs can only ever remix, even if its technically true (doubt), in my opinion only says to me that all knowledge is only ever a remix, perhaps even mathematically so. Anyone who still keeps implying these are statistical parrots or whatever is just going to regret these decisions in the future.

> Anyone who still keeps implying these are statistical parrots or whatever is just going to regret these decisions in the future.

You know this is a false dichotomy right? You can treat and consider LLMs statistical parrots and at the same time take advantage of them.

Re: Scaling long-running autonomous coding

#22

Post by pseudosavant »

ramraj07 wrote: post #16
halfcat wrote: post #10

So AI makes it cheaper to remix anything already-seen, or anything with a stable pattern, if you’re willing to throw enough resources at it. AI makes it cheap (eventually almost free) to traverse the already-discovered and reach the edge of uncharted territory. If we think of a sphere, where we start at the center, and the surface is the edge of uncharted territory, then AI lets you move instantly to the surface. If…

The fundamental idea that modern LLMs can only ever remix, even if its technically true (doubt), in my opinion only says to me that all knowledge is only ever a remix, perhaps even mathematically so. Anyone who still keeps implying these are statistical parrots or whatever is just going to regret these decisions in the future.

But all of my great ideas are purely from my own original inspiration, and not learning or pattern matching. Nothing derivative or remixed. /sarcasm

Re: Scaling long-running autonomous coding

#23

Post by ukuina »

halfcat wrote: post #10

So AI makes it cheaper to remix anything already-seen, or anything with a stable pattern, if you’re willing to throw enough resources at it. AI makes it cheap (eventually almost free) to traverse the already-discovered and reach the edge of uncharted territory. If we think of a sphere, where we start at the center, and the surface is the edge of uncharted territory, then AI lets you move instantly to the surface. If…

Single-idea implementations ("one-trick ponies") will die off, and composites that are harder to disassemble will be worth more.

Re: Scaling long-running autonomous coding

#24

Post by NitpickLawyer »

heavyset_go wrote: post #20
ramraj07 wrote: post #16

Earlier quoted context omitted.

The fundamental idea that modern LLMs can only ever remix, even if its technically true (doubt), in my opinion only says to me that all knowledge is only ever a remix, perhaps even mathematically so. Anyone who still keeps implying these are statistical parrots or whatever is just going to regret these decisions in the future.

Yeah, Yann LeCun is just some luddite lol

I don't think he's a luddite at all. He's brilliant in what he does, but he can also be wrong in his predictions (as are all humans from time to time). He did have 3 main predictions in ~23-24 that turned out to be wrong in hindsight. Debatable why they were wrong, but yeah.

In a stage interview (a bit after the "sparks of agi in gpt4" paper came out) he made 3 statemets:

a) llms can't do math. They can trick us with poems and subjective prose, but at objective math they fail.

b) they can't plan

c) by the nature of their autoregressive architecture, errors compound. so a wrong token will make their output irreversibly wrong, and spiral out of control.

I think we can safely say that all of these turned out to be wrong. It's very possible that he meant something more abstract, and technical at its core, but in the real life all of these things were overcome. So, not a luddite, but also not a seer.

Re: Scaling long-running autonomous coding

#25

Post by gjadi »

NitpickLawyer wrote: post #24
heavyset_go wrote: post #20

Earlier quoted context omitted.

Yeah, Yann LeCun is just some luddite lol

I don't think he's a luddite at all. He's brilliant in what he does, but he can also be wrong in his predictions (as are all humans from time to time). He did have 3 main predictions in ~23-24 that turned out to be wrong in hindsight. Debatable why they were wrong, but yeah. In a stage interview (a bit after the "sparks of agi in gpt4" paper came out) he made 3 statemets: a) llms can't do math. They can trick us with…

Have this shortcomings of llms been addressed by better models or by better integration with other tools? Like, are they better at coding because the models are truly better or because the agentic loops are better designed?

Re: Scaling long-running autonomous coding

#28

Post by NitpickLawyer »

gjadi wrote: post #25
NitpickLawyer wrote: post #24

Earlier quoted context omitted.

I don't think he's a luddite at all. He's brilliant in what he does, but he can also be wrong in his predictions (as are all humans from time to time). He did have 3 main predictions in ~23-24 that turned out to be wrong in hindsight. Debatable why they were wrong, but yeah. In a stage interview (a bit after the "sparks of agi in gpt4" paper came out) he made 3 statemets: a) llms can't do math. They can trick us with…

Have this shortcomings of llms been addressed by better models or by better integration with other tools? Like, are they better at coding because the models are truly better or because the agentic loops are better designed?

100% by better models. Since his talk models have gained more context windows (up to usable 1M), and RL (reinforcement learning) has been amazing at both picking out good traces, and taught the LLMs how to backtrack and overcome earlier wrong tokens. On top of that, RLAIF (RL with AI feedback) made earlier models better and RLVR (RL with verifiable rewards) has made them very good at both math and coding.

The harnesses have helped in training the models themselves (i.e. every good trace was "baked in" the model) and have improved in enabling test time compute. But at the end of the day this is all put back into the models, and they become better.

The simplest proof of this is on benchmarks like terminalbench and swe-bench with simple agents. The current top models are much better than their previous versions, when put in a loop with just a "bash tool". There's a ~100LoC harness called mini-swe-agent [1] that does just that.

So current models + minimal loop >> previous gen models with human written harnesses + lots of glue.

> Gemini 3 Pro reaches 74% on SWE-bench verified with mini-swe-agent!

[1] - https://github.com/SWE-agent/mini-swe-agent

Re: Scaling long-running autonomous coding

#29

Post by cheevly »

shubhamjain wrote: post #15
simonw wrote: post #3

One of the big open questions for me right now concerns how library dependencies are used. Most of the big ones are things like skia, harfbuzz, wgpu - all totally reasonable IMO. The two that stand out for me as more notable are html5ever for parsing HTML and taffy for handling CSS grids and flexbox - that's vendored with an explanation of some minor changes here: https://github.com/wilsonzlin/fastrender/blob/19bf103…

Why attempt something that has abundant number of libraries to pick and choose? To me, however impressive it is, 'browser build from scratch' simply overstates it. Why not attempt something like a 3D game where it's hard to find open source code to use?

Assets are very hard to produce and largely unsolved by AI at the moment.

Re: Scaling long-running autonomous coding

#30

Post by retinaros »

Agentic coding is a card castle built on another card castle (test time compute) built on another card castle (token prediction) the mere fact that using lot of iterations and compute works maybe tells us that nothing is really elegant about the things we craft.
Post reply on HN