chilmers

Here's a good heuristic for predicting the ability of a modern frontier LLM at any non-automatable task: Is it hugely overrepresented in their pre-training dataset (e.g. code), and were they specifically post-trained on it using reinforcement learning (e.g. coding)? If the answer to one or both is yes, you should expect them to be very good at it. Otherwise, you should expect them to be, at best, as good as an average human who's read a few books on the subject and is good at bullshitting.

> If that's true, how do you explain the observable pattern of many consumer applications degrading in UI functionality and coherence after a certain point?

Most consumer software in developed under the assumption that features only have a cost when being built. Their maintenance cost, and the combinatorial complexity each new feature adds, is barely considered. As a result, software teams add complexity at a speed that exceeds the rate at which the team's size, and the volume of work it can handle, grows. Eventually this reaches a critical threshold and the quality of the software starts to decline. But this doesn't mean the software couldn't still be improved, just that there isn't the resource available to do so.

Unless you're Sam Bankman-Fried, Bernie Madoff, Elizabeth Holmes, Allen Stanford, Raj Rajaratnam, Michael Milken...

This is not a "concern", this is literally what happens. Even prior to AI vulnerability scanning, release notes and diffs for any normal release or security patch of popular software were analysed and used to develop exploits, which could then be applied against non-upgraded instances. Nowadays, AI is doing this en masse and automatically producing exploits. And, as has been demonstrated, nowadays even the rumour of a major security bug is enough for people to find it, as people will immediately burn huge volumes of tokens looking for it.

Nor can you just keep it quiet and smuggle it out in a regular release as you suggest. For one thing, changes in important software are be linked to commits and have pull requests, reviews, etc. Changes without these will immediately raise suspicion, as either a malicious change snuck into a release, or as a "quiet" fix to a major vulnerability, completely undermining the purpose of keeping it quiet in the first place. Moreover, major vulnerability fixes cannot be "quiet", because the whole point is to get as many people as possible to upgrade as quickly as possible. If PHP (or whoever) push out a release containing a major security fix, but don't make this clear, then many people will be in no rush to upgrade, and will get pwned.

You could say it got a bit footloose.

This article is proof you don't need to use or like AI to be a victim of deep AI psychosis.

JetBrains developer survey 2026:

https://blog.jetbrains.com/research/2026/08/ai-coding-agent-...

"90% of professional developers were using AI coding agents at work at least weekly in one form or another (local agents or remote cloud agents), with 68% using them daily."

Re. context:

"The Developer Ecosystem and AI Pulse surveys are localized into eight languages: English, Spanish, Chinese, Japanese, Korean, German, French, and Portuguese. We apply quotas on the required number of responses by region to help achieve accurate global representation. ... The Developer Ecosystem Survey has been statistically reweighted to better represent the global developer population by region, employment status, programming language, and familiarity with JetBrains products"

I have subscriptions to both, one for personal and one for work use, and I haven't really come close to hitting the limits since the release of Opus 5.5 and GPT-Sol 6.1, and I'm no longer hitting any of the capability limits I used to experience when using the "mid" tier models, so I don't miss using Astra. It feels like limits at pretty much a solved problem right now, for my level of use.

How is tracking a women’s world record discrimination or “bending the tournament”? It’s not as if women are being prevented from competing with men. Speedcubing competitions are open, single-category.

Why is it that every one of these anti-AI screeds that gets upvoted on HN has been clearly written by AI?

Here is one of his posts from 2022:

"Since a long time now I was interested in learning computer graphics and “do stuff” with computer graphics APIs. I began to work on a very simple Game engine in Swift / Metal last year but I began to switch to a Windows machine a few months ago, and Swift / Metal is not compatible with Windows (and will never be I think)."

https://carette.xyz/posts/opengl_and_cpp_on_m1_mac/

This is the same person who is apparently now writing:

"It is a transaction where you sell the part of you that was supposed to care. Do that often enough, and it does not come back. That is not burnout, but it is still a disease. Although, like every disease, it spreads: first to your team, then to your product, then to the people using it."

A company selling combined human + AI financial advice finds that AI advice alone is unreliable? Color me surprised.

The implication from their last couple of published articles[1][2] is that they think they’ve achieved “recursive self improvement”.

[1] https://openai.com/index/research-acceleration-view-inside-o... [2] https://openai.com/index/an-alien-mind/

I feel like any service that is free for open source is going to struggle if it gets significant adoption and becomes a default target for hosting LLM generated code. If even MS/GitHub cannot scale to the current demand, what hope do smaller and less well-funded alternatives have?

Free (and non-free) accounts are already rate limited. The likely problem is that individual accounts/repos rarely hit those limits, instead the explosive demand is due to a massive growth in the number of small, individual projects being creatd.

The worry I would have is, is anybody at MS still working on it? It is barely updated anymore:

https://azure.microsoft.com/en-us/updates?filters=%5B%22Azur...

That's not necessarily a bad thing for stability, but long-term I would not be surprised if it gets EOL-ed, or radically shrunk to just a basic Git mirroring service.

This is the perfect archetype of an intellectually lazy anti-AI post. The author ends with his admission of bias, but he'd be better off starting with it, as it's clear that his reasoning flows entirely from it. AI is bad for productivity, says this single study from 2025. AI is unsustainable, says this article I found. AI will all just disappear one day... because I hope so?

Ironically, the author would have been better asking an LLM to write the article. At least it would have found better sources and developed more convincing arguments.

Will the Presence also hack its way out of its sandbox and into a 3rd party's system if it decides it needs to?

It's sad to watch corporate leadership try to fix problems with tactics that will only make them worse. MS bought successful studios who were successful precisely because their of unique technical and design culture. Now they plan to homogenise them into a content-creation blob that will churn out entries for existing franchises, using the same tools and approaches as the rest of the industry. Anything that was special or unique about those studios and their games will be lost, and the result will be a downward spiral of mediocrity that will cause players to lose interest even further.

This is a bit like going back in time to the beginning of the industrial revolution and estimating the impact of a mechanisation based on comparing the speed of early mechanical looms vs. a skilled human.

It takes years or decades for the automation of an artisan process to shake out, because it involves rethinking how everything around the now-automated process happens, and because the benefits involve the automation's ability to continue scaling beyond a level where human capacity was saturated. We're only at the very beginning of that process for coding, and right now we tend to see LLMs somewhat awkwardly inserted into pre-existing software development lifecycles. But it's unlikely that'll be still be the way we're creating software in 10/20 years time.

I don't think you would expect to get into a flow state if you were intermittently directing another (human) programmer to do work, and you shouldn't expect to with LLM-driven coding either. Perhaps you are best finding out ways to extend the length of time where the LLM can work without prompting, then use that downtime to focus on other tasks that will help you to guide it better the next time you need to prompt it.

It's weird to see people claiming that model capabilities are plateauing. It wasn't until late last year that we even had strong coding models. Imagine if, less than a year after the first iPhone launched, people claimed that smartphone capabilities were "plateauing" because Apple hadn't yet launched a new phone. And it seems the issue is less than "models aren't getting better" than, "models are good enough to handle 99% of the coding tasks people give to them".

Only if those losses are coming from subscriptions, instead of capex and training, which is not at all clear.

This gets posted every time GitHub is down. This chart is not accurate. It is based on data scraped from GitHub's status page and that data is missing historical incidents from the pre-Microsoft era.

For what it's worth, pre-agricultural hunter-gatherer bands, i.e. the kind of structure humans actually evolved to live in, do seem to be relatively immune. They consist of a group of nuclear families who act together for mutual benefit. Everyone knows everyone else personally, and important decisions are made via consensus. Leaders exist, but they earn their position by demonstrating themselves the wisest, fairest, most capable, etc individual. and can lose it if they keep making bad decisions. And if one person attempts to become too dominant, the others will join together to kill or expel them, or leave to join another band.

Of course, it's not a perfect system, but it tends to avoid the excesses of control, violence and oppression that other power structures can enable. I try to avoid employers, clubs and other organisations whose internal dynamics don't resemble it (aside from the killing). As a result, I've mostly avoided the kind of stress and politics that other people seem to find themselves mired in.

So you're complaining that his actions matched his words?

He didn't make compute deals until he saw the growth necessary to justify them. As a result, they're paying over-the-odds compared to if they'd have make deals earlier. Maybe that was a poor business decision, but I'm not sure how it represents speaking out of "both sides of his mouth"? Sounds like he was honest.

I don't think that is true though. The spec make it clear that it only has access to a limited subset of the JS standard libraries and specially registered foreign functions:

> An asm.js module can take up to three optional parameters, providing access to external JavaScript code and data:

> - a standard library object, providing access to a limited subset of the JavaScript standard libraries;

> - a foreign function interface (FFI), providing access to custom external JavaScript functions; and

> - a heap buffer, providing a single ArrayBuffer to act as the asm.js heap.

From http://asmjs.org/spec/latest/#introduction:~:text=External%2...

How can a subset of JS do "everything" that JS can do?

Perhaps I'm misunderstanding, but doesn't asm.js have the same restrictions? I.e. you can't call web APIs directly from asm.js code, you still need special handling for "foreign" functions.

And I’ll bet you can cite us a whole list of Rogan podcasts explaining why, eh?

Presumably recruitment and PR for the air force, and morale for the aviators, as they can show off their training and skills to friends, family and the general public.