HNHacker News
TopNewBestAskShowJobs

IshKebab

20,326 karma · joined August 23, 2014

submissionscomments
IshKebab··on Why We're Banning AI
Uhm yeah probably. Are you suggesting that LLMs have been stagnant despite investment?

I don't understand your point. They're getting better so quickly it's hard to keep up.

IshKebab··on I Love Software Rendering
Unreal Tournament also had a software renderer IIRC, with an interesting dithering effect.
IshKebab··on New 'super ice' 10 times stronger, bringing it close to the strength of concrete
I hate how researchers always have to pretend there's a practical application for their inventions. This is clearly never going to replace concrete in any situation. But it's still interesting.
IshKebab··on New Withings Users Must Pay $99/Year for Apple Health Syncing
I guess this will lead to more companies relying on strong crypto to maintain control over their devices rather than relying on the amount of work required to reverse engineering them.
IshKebab··on I know why the mathematicians weep
They're generating by prediction. That's how they work - the predict the most plausible next token.
IshKebab··on '500-hour' Order of the Sinking Star is as complex and exhausting as it sounds
I felt like that in the demo. It's waaaay too slow to do anything difficult. Basically everything in the hour or so that I played was trivially easy.

Compared to Baba Is You which gets difficult and interesting after like 10 minutes.

IshKebab··on I know why the mathematicians weep
There are two reasons:

1. LLMs aren't run deterministically - they use a bit of noise to improve the output. So even with the exact same input you're going to get a whole distribution of outputs.

2. Feedback from the environment. LLMs (and humans!) work best when they can test ideas. That's why coding agents work so well. If you ask an agent 1000 times to just think of a compression algorithm, it'll probably spit out very samey answers just like a human would. If you give it a test corpus and let it try out different things it will absolutely be able to come up with novel ideas.

IshKebab··on I know why the mathematicians weep
I think they're referring to the fact that neural networks are more or less black boxes. Their internals are blurry and not explainable in the same way that traditional linguistics tries to explain things - with rules.

Of course, ultimately that's because language itself doesn't follow simple rules like linguists would prefer. They're more like guidelines.

IshKebab··on E Ink Kaleido 4
This is their colour filtered black and white product. Quite muted colours in the past and it only really works well with a front light, which kind of ruins the point IMO.

I'm much more interested in their Spectra and Gallery displays which use color pigments. You can even buy e-ink gallery style displays for a totally unreasonable price!

https://inkposter.com/products/inkposter-tela-28-5?variant=4...

Actually to be fair that is not a totally crazy price.

IshKebab··on Why We're Banning AI
Yeah it's pretty obvious now. But look at the improvements Anthropic made to Opus 5.5. It went from physically painful to read, to kind of annoying but tolerable. I wouldn't bet on it taking a long time to reach high quality prose.

Btw it is kind of weird how it still can't write prose normally but it can write impressively high quality code now. Must be something to do with their training. Or maybe it's easier to quantify good code.

IshKebab··on Muse Spark Open
I tried Muse Spark 1.3 for a while because it's so cheap, but it's definitely in a lower league than Astra, Opus 5.5 and Argon. Still produces slop if you try and vibe code with it, whereas the more capable models write good code.
IshKebab··on Steam Client is now faster than Epic's new vibe coded launcher
They're posting this as some kind of gotcha against AI, as if Valve aren't using AI too.
IshKebab··on David Rudnick on "borderline impossible" process of designing Arsenal typeface
Whatever typeface they're using on this website is kind of awful. It's got totally random weights on different parts of the letters. Look at the bottom of the e for example.
IshKebab··on Maybe a better way to track the release history of a Git repo
[flagged]
IshKebab··on Anthropic conjures threat that doesn't exist yet to stoke fears of AI apocalypse
Uhm yeah I don't think they're saying it's likely now, just that it's very likely in future. The Register is being quite dumb here.

Though I suspect the way we'll get "wild AI" first is through cloud services paid for by cybercrime rather than IoT botnets - they're too underpowered.

IshKebab··on Cyberattack halts construction of tunnel between Denmark and Germany
Even if you do it won't help. You need to be off-grid. The vast majority of solar panels are deliberately designed to shut down when there's a power cut so that they don't feed power back into dead wires (that might be in the middle of being repaired).
IshKebab··on Redesigning source control at Anthropic
[fan out agents]
IshKebab··on Redesigning source control at Anthropic
Well that's actually true though. The very very few investors that are good at investing do invest their own money. The vast majority of investors are no better than index funds.
IshKebab··on Gemini 4 Argon
To be fair they dramatically improved how annoying Claude is with Opus 5.5. It's not perfect but it's not nearly as awful as it was.
IshKebab··on 11 year old unsoundness bugs found in Rust borrow checker
> Most systems these days have over commit

Sure if you ignore the most popular desktop OS in the world.

Also overcommit (at least how Linux does it) is dumb. It leads to broken OOM behaviour (killing a totally random process, or more often locking up the system and hard rebooting). And some apps do need to handle OOM properly, e.g. formal verification systems which can use enormous amounts of memory. Jasper has to resort to monitoring page fault stats to know when it is using too much memory.

IshKebab··on Firefox Enterprise
Maybe they're really non-technical users who figured things out in 2005 and haven't learnt a single thing since then.
IshKebab··on AI coding agents generate more code, but not more software
Yeah this is nonsense. Maybe true a few months ago but it's going to look hilariously short sighted in 6 months.
IshKebab··on AI is a tool and using it well takes skill
Yeah I don't think so. It takes some knowledge today if you want to have control over architectural decisions. But honestly even if you don't the latest models still make fairly reasonable decisions.

I am pretty sure my wife could vibe code website or whatever at this point, or if not she will be able to in 6 months.

IshKebab··on Fingerprints are the new cookies
https://fingerprint.com/blog/disabling-javascript-wont-stop-...
IshKebab··on Fingerprints are the new cookies
> I still don't understand why the highest entropy APIs aren't gated behind browser consent.

Last I checked it's because it was generally considered that preventing fingerprinting is practically impossible anyway. It's not like you need a lot of entropy for fingerprinting, and the web API surface is enormous. To actually stop fingerprinting you'd need to make your web browser more or less useless.

IshKebab··on Show HN: PS2 games statically recompiled to WebAssembly
Not explicitly but it's clearly all done using AI. Read the prose.
IshKebab··on Total Annihilation, rebuilt for modern computers
Yeah I'd be surprised. Law has a certain amount of common sense and if there's a new technique that just lets you easily clone any software they aren't going to say "aha good loophole you win".

Plus I think clean room process needs two separate teams and this is just one guy.

I did find this post from actual lawyers about it, but basically it's a grey area that hasn't been decided yet.

https://www.beankinney.com/news-and-insights/blogs/512000-li...

IshKebab··on Total Annihilation, rebuilt for modern computers
Interesting. Looks like he has made a "clean room" clone of TA by pointing AI at it, getting it to reverse engineer it and write some docs, and then have another AI write an implementation based on those docs.

I would be surprised if that stood up in court but I doubt anyone cares for TA at this point anyway.

IshKebab··on 3rd Yandex Cloud Data Center Was Hit
Some do, but many don't. They aren't immune to propaganda.
IshKebab··on I'm sorry, but you still have to think
I dunno. You still have to test, but in my experience if I noticed unreliability during high load I could just tell Claude "I noticed unreliability during high load" and it would easily have set up a high load benchmark, discovered this broadcast channel issue and fixed it.

It definitely improves the results if you do think but I don't think "you still have to think" is a safe space that's going to save us from unemployment.

← PreviousPage 7 of 34Next →