Bjorkbat

This is a reflection of a complaint I frequently make about so called "world models". It's more accurate to call them video game models. They don't simulate the world, they simulate video games. For that reason a common issue with them is that the physics are video game physics rather than real world physics.

I know people hate it when I draw a comparison between crypto/NFTs and AI since the latter has some utility, but it really is crazy just how similar the vibes are, and you can see this in this recent fad of people trying to clone existing software with AI.

In case anyone has a fuzzy memory, back during peak NFT hype grifters would take art made by other artists and attempt to tokenize it as NFTs without the consent of the original artist, or otherwise possibly copy said artist's work and make a collection of it.

It was disgusting and parasitic then and it's disgusting and parasitic now.

I was about to say, one take I've heard is that the party ideology considers profit a kind of "rent" in a derogatory way, and consequently seeks to undermine the ability of western companies to collect large profit margins

Okay, look, at this point this really should be a company-ending incident. Or, I should say, incidents. I was fine the first time I heard of some rogue hacking accident, but now I've lost count of how many times OpenAI has lost control of its AI.

They may be some of the smartest people in the industry but they're clearly too incompetent and/or negligent to be allowed to develop more intelligent AI. This is ridiculous.

Huh, person who contributed this to Atlas Obscura has an interesting username.

Let's do some roleplay. You're an employee at OpenAI evaluating a model in a sandbox and you and your colleagues are discussing the fact that while the sandbox doesn't allow for internet access, it actually kind of does allow for internet access in that it allows models to download whatever software packages they need through a package manager that OpenAI hosts internally. So to be clear, OpenAI hosts the package manager, but the software packages themselves are still out there, on the internet.

Do you think that this package manager could potentially be a problem? Do you think it might be worthwhile to host the software packages themselves on an internal, sandboxed network, just to be extra certain? Or would you dismiss this as a needless precaution?

All the more reason then to call their bluff. "Oooh we created a genie and it's almost out of the box". Cool, you've hyped the IPO, but also you have to plead your case before Congress as to why the company should continue to operate given its failure to prevent AI-related accidents from occurring.

When I hear about incidents like these my first reaction is that the people responsible for developing frontier AI are too incompetent and/or negligent to (safely) develop AGI / superintelligence.

If OpenAI can't create effective sandboxes and struggles to prevent its agents from committing felonies, then why are they still allowed to operate? Why are the employees who are responsible for these lapses in AI security still employed?

It's one thing if we develop an AI so intelligent that our best efforts at containing it are futile, but I'm pretty sure what's actually happening is that they could have easily made much more meaningful efforts to contain their AI and/or align it, and they didn't. I think this is a case of negligence and incompetence when it comes to safety and security, and we've entrusted these incompetent and negligent people with developing frontier AI.

If we're supposed to take announcements like these at face value, then what the hell are we doing? We wouldn't trust a bunch of incompetent and negligent engineers to build bridges or nuclear power plants or planes (well...not so sure about that last one), so why are we letting people who are demonstrably negligent and incompetent when it comes to safety and security build the thing they assure us could cause massive damage if not properly controlled/aligned?

EDIT: sorry guys, wrote this up pretty quickly, at least you know from my typos that I actually wrote this.

The most straightforward answer is that despite efforts to design a benchmark that, in theory, is supposed to measure generalizable intelligence, performance on ARC-AGI-3 can't be reliably correlated to performance anywhere else. I kind of lost faith in it after o1 or o3, I can't remember which, absolutely crushed ARC-AGI-1.

And, you know, maybe also some funny business. I think it's good to be a little suspicious of a model that happens to shoot upwards in performance on a specific benchmark while also kind of keeping up with the pack on a bunch of other benchmarks.

For my needs, sure. One caveat is that if I’m doing a longer trip I might save power by turning it off when the ground is flat/downhill

Arguably the "social coding" angle is the reason why Tangled is the only actual alternative to GitHub. Otherwise there's actually plenty of other alternatives to GitHub, but none of them have GitHub's punchcard, which I'm ashamed to admit is the primary draw for me.

I actually bought one of these and love it. My road bike is indistinguishable from any other normal road bike with the exception of it being much easier to get around now.

From experience I can say that it seems like a lot more common sense to simply replace the front wheel of your bike with a wheel containing an electric motor, rather than attaching whatever Asus is selling to your bike.

One theory I've been entertaining is that whenever GPT-3.5 came out a lot of people were talking about the "bitter lesson" and how scale was all we really needed to get to AGI. No need for any fancy tricks, just release a larger model trained on more data, by the time we released a hypothetical "GPT-5 sized" model we'd have AGI.

Anyway, the actual theory is that Google and Meta have fallen behind because they've been playing by this playbook of focusing on scale and training data, whereas OpenAI and Anthropic have done so well because they are likely doing much more interesting things to improve their models over time. It makes sense when you realize that one of Google's key strengths, besides talent, is that they have an incredible amount of data they can use for training due to being both the world's leading search engine as well as having all that video data from YouTube. Scaling the training data makes more sense to them than it does to Anthropic and OpenAI, who are both relatively data-disadvantaged.

You can kind of see this when you look at the Gemini 3 scorecard when it came out (https://blog.google/products-and-platforms/products/gemini/g...) and notice that while it wasn't as good as Claude And GPT at coding, it scored higher on a bunch of other non-coding benchmarks, and I think the reason why is simply because of Google's data advantage.

If true, I feel even more vindicated for believing that the "scale is all we need" narrative was bullshit.

I was about to post a snarky comment along the lines of "Sierra? The publishers of Homeworld and Homeworld 2?"

I always got the impression that the only way a relatively average software developer could make a “successful” SaaS is if they built something weird and niche that appeals to an audience of <1000 paying customers. In that sense, it doesn’t really matter if your competition is better at SEO since your competition never even thought to build something like this, never cared, and not to mention that the market for this thing is so small that SEO is arguably a wasted skill. You’ll need to acquire these people by finding them directly or through word of mouth.

This blog post seems to fundamentally misunderstand the nature of solo-developer SaaS, but then again arguably mostly software developers also fundamentally misunderstand it.

To me demoscene is kind of synonymous with the Amiga, so I would argue that peak demoscene lines up with the rise and fall of the Amiga brand.

So I think that's maybe the other differentiator between web experiment and art, because demoscene has a very distinct but difficult to describe cultural element that makes me identify it as art.

I'm keenly aware, I have a pretty extensive collection of Hacker News bookmarks. It's hard to articulate why I think these are different, but I think the best way to put it is that cachemonet feels a lot more avant garde, and perhaps also a reflection of a very particular form of "web culture" that has no clear successors.

People are experimenting with what you can do on the web, but the experiments aren't very "aesthetically inspiring". For that reason I'm kind of lukewarm on neal.fun.

Finding out that this is over 10 years old has made me profoundly sad. Despite the age of LLMs arguably unlocking massive amounts of productivity and agency for developers and non-developers alike, it feels as though we are living in a dark age of creativity on the web, maybe even a dark age for computer culture in general.

Ironically the physics are kind of my biggest criticism. They call these "world models", but I think it's more accurate to call them "video game models" because they employ "video game physics" rather than real world physics, among other things

This is most evident in the way things collide.

I’m sure there’s some boring neuro-chemical explanation for this, and I won’t doubt or deny the neuro-chemical explanation, but the fact that there’s a mushroom that consistently brings about hallucinations of tiny people is so bizarre that I kind of want to indulge in equally bizarre explanations. Maybe it’s not a hallucination and this mushroom simply allows us to see the tiny people all around us. Maybe mushrooms are intelligent and are intentionally making us hallucinate tiny people.

It’s a little bit crazy, I know, but it’s odd to me that evolutionary forces would produce a mushroom that makes you have some specific hallucinations, rather than simply make things swirl together or simply produce intense feelings of euphoria or dread. I mean, marijuana just gets you high and that’s that.

Oh yeah, funny enough even though I’m a bit of an AI art hater I actually thought very early Midjourney looked good because of all had an impressionistic, dreamy quality.

Funny enough that had crossed my mind with the woodchuck example, because at a glance I can't see any weird artifacts, but I felt confident I could tell it was AI generated immediately if I saw it in the wild, and I couldn't really explain why. My immediate guess was "well, who the hell would actually bother to make something like this?"

Something I find weird about AI image generation models is that even though they no longer produce weird "artifacts" that give away that the fact that it was AI generated, you can still recognize that it's AI due to stylistic choices.

Not all examples they gave were like this. The example they gave of the word "Typography" would have fooled me as human-made. The infographics stood out though. I would have immediately noticed that the String of Turtles infographic was AI generated because of the stylistic choices. Same for the guide on how to make chai. I would be "suspicious" of the example they gave of the weather forecast but wouldn't immediately flag at as AI generated.

Similar note, earlier I was able to tell if something was AI generated right off the bat by noticing that it had a "Deviant Art" quality to it. My immediate guess is that certain sources of training data are over-represented.

Reminds me of the old gem of the Web 1.0 internet that was Exit Mundi

One of my most frustrating things regarding the potential of an AI bubble was some very smart and intelligent researcher being incredibly bullish on AI on Twitter because if you extrapolate graphs measuring AI's ability to complete long-duration tasks (https://metr.org/blog/2025-03-19-measuring-ai-ability-to-com...) or other benchmarks then by 2026 or 2027 then you've basically invented AGI.

I'm going to take his statements at face value and assume that he really does have faith in his own predictions and isn't trying to fleece us.

My gripe with this statement is that this prediction is based on proxies for capability that aren't particularly reliable. To elaborate, the latest frontier models score something like 65% on SWE-bench, but I don't think they're as capable as a human that also scored 65%. That isn't to say that they're incapable, but just that they aren't as capable as an equivalent human. I think there's a very real chance that a model absolutely crushes the SWE-bench benchmark but still isn't quite ready to function as an independent software engineering agent.

So a lot of this bullishness basically hinges on the idea that if you extrapolate some line on a graph into the future, then by next year or the year after all white-collar work can be automated. Terrifying as that is, this all hinges on the idea that these graphs, these benchmarks, are good proxies.

And if they aren't, oh wow.

I definitely agree with you there. I contracted with a company that had some older engineers who were in largely managerial roles who really liked using AI for personal projects, and honestly, I kind of get it. Their work flow was basically prompt, get results, prompt again with modifications, rinse and repeat, it's low effort and has a nice REPL-like loop. Paraphrasing a bit, but it basically re-kindled the joy of programming for them.

Haven't gotten the chance to ask, but I imagine managing a team of AI agents would feel a little too much like their day job, and consequently, suck the fun out of it.

That said, looking back, I think the reason why generative AI is so fun for so many coders is because programming has become unnecessarily complex. I have to admit, programming nowadays for me feels like a bit of a slog at times because of the sheer effort it can sometimes take to implement the simplest things. Doesn't have to be that way, but I think LLM copy-paste machines are probably the wrong direction.

I think people underestimate the degree to which fun matters when it comes to productivity. If something isn’t fun then I’ll likely put it off. A 15 minute task can become hours, maybe days long, because I’m going to procrastinate on doing it.

If managing a bunch of AI agents is a very un-fun way to spend time, then I don’t think it’s the future. If the new way of doing this is more work and more tedium, then why the hell have we collectively decided this is the new way to work when historically the approach has been to automate and abstract tedium so we can focus on what matters?

The people selling you the future of work don’t necessarily know better than you.

Yeah, I thought about that after I looked at the SWE-bench results. It doesn't make sense that the SWE results are barely an improvement yet somehow the model is a more significant improvement when it comes to long tasks. You'd expect a huge gain in one to translate to the other.

Unless the main area of improvement was tools and scaffolding rather than the model itself.

> Practically speaking, we’ve observed it maintaining focus for more than 30 hours on complex, multi-step tasks.

Really curious about this since people keep bringing it up on Twitter. They mention it pretty much off-handedly in their press release and doesn't show up at all in their system card. It's only through an article on The Verge that we get more context. Apparently they told it to build a Slack clone and left it unattended for 30 hours, and it built a Slack clone using 11,000 lines of code (https://www.theverge.com/ai-artificial-intelligence/787524/a...)

I have very low expectations around what would happen if you took an LLM and let it run unattended for 30 hours on a task, so I have a lot of questions as to the quality of the output

> I used to make Dragonball Z fan sites...

Based

I'm working on a revamp of my personal site. I do a lot of creative coding, most of them are throwaway experiments, so I thought I'd showcase more of them there. Besides that though, I have some "rare pepes" that I've been meaning to put somewhere. What I like about these is that they're highly polished, animated gifs that imitate the sort of "holographic" effect you'd find in rare collector's cards, but at the same time you can't track down who originally made them, they aren't part of some professional's online portfolio. In that sense they feel like a special piece of internet folk art, made by some complete rando.

Nowadays we have Pinterest and the like, but I really like the idea of creating my own little online space for images I like.