That "growing too large to be useful" was exactly when they were sold to Lenovo.

https://i.k8r.eu/1Pc-Tw.png

And why is that so hard for you to believe?

I'm especially surprised by this fixation on the prompts. You can't have it both, either AI is great because you can just ask for something, then the exact prompts shouldn't matter, or the precise wording matters, then I'd just want a regular, deterministic query language.

Or a search for precise image downscaling in krita (on 2026-08-03) had the following Google AI result:

> To scale down an image in Krita by an exact integer factor with precise pixel sampling, use Scale Image to New Size or the Transform Tool, set your filter to Box, and input exact integer-divided values.

With the source https://forum.kde.org/viewtopic.php%3Ff=274&t=135270.html

Except that that forum thread has the following posts:

> just use the "box" filter when scaling

Followed by

> There is no box filter.

Which is correct, as Krita does not have a box filter for scaling images.

It took two further rounds of back and forth for Google's AI (where it suggested nearest neighbor) to finish with

> You are entirely correct: Krita does not feature a true Box filter, and using Nearest Neighbor to downscale a complex image introduces heavy aliasing, skipped rows, and jagged artifacts.

This could've been avoided if the AI actually checked against the source code and documentation for krita, and checked the ingredients of the cleaning products in the other case together with basic principles of chemistry.

There's this saying:

> If one person says it's raining and another says it's sunny, a journalist is supposed to go out and look at the weather.

I don't need an automated way to summarize random half-truths from forum posts online. If AI is supposed to bring any benefit, it should be finding out the actual truth about the question being asked, not just report whatever opinions are being voiced online.

I love how people have said this for over a year now, and no matter how often I try it, it's still just as broken.

Try it with any task that isn't in the training data. Say, a custom protocol where you need to cross-reference multiple implementations and the specs to really get it, and with no answers on stackoverflow or medium.

At first it's hilarious, but after a while it just gets exhausting. For all these real-world tasks you need to put so much effort in that it's much easier to just write the code yourself, even with the latest (current Gemini) and greatest (Opus 4.6/4.7) models.

If the parts for your project don't already exist, AI can't help you either. And if they do exist, why spend money on AI if the GitHub search is free and you can just fork and modify what you need?

Let's look at

> So let me be plain about it: Claude was not here. No LLM wrote this—not a sentence of it, nor was it outlined, drafted, reviewed, checked, etc. by one, and there is no prompt behind it either

AI would've chosen rule of three (instead of 4 + etc), wouldn't have used the either at the end (doesn't fit into the rhythm), wouldn't have used "let me be plain" (more likely to use "let me be clear" or "let me be frank"), and wouldn't have chosen "Claude was not here".

If you've actually read a lot of essays and AI generated text (and sadly, I have), it's very obvious that this was neither generated nor rephrased by AI.

This is very much like the recent "kurzgesagt got banned by YouTube's AI detector" story, where the humans whose style was copied by AI now get accused of being AI

But this follows human structure and rhythm through and through, whether on the level of sentences, paragraphs, sections, or the whole text.

I'd say it's handwritten.

In Walter Murch's "In the Blink of an Eye" he describes how, when cutting a movie, the reason cuts work at all is because human experience has cuts: we blink.

More specifically, we blink when a thought is complete, to finish one thought and begin another.

When a movie is well-edited, the audience will adapt to the rhythm of the movie, and you know you've placed a cut at the right moment when the audience will blink at the exact moment of the cut.

The same "rhythm of thought" applies to written essays as well, and once you get a feeling for it, you'll notice how AI generated text follows no such rhythm, no such structure.

AI generated text has the rhythm of TV salespeople combined with tiktok brainrot, overwhelming you while at the same time saying nothing at all.

It's so bad, and so everywhere, that I'm genuinely considering switching to a pre-AI stable distro and waiting 10 years for the AI hyped to be over before updating any piece of software ever again.

So it certainly seems like at least some models are being censored.

If the user chooses to free certain apps from restrictions, that should be their choice.

But if a piece of software overrides the users' choice, that certainly qualifies as a CVE.

This is an important distinction! The user must always have final say.

Whether an app breaks out of the sandbox and steals my vacation photos, or the OS sets new restrictions I can't remove, both are wrong.

Any piece of software has to act in service of the user, and ONLY the user. It must not do things or set restrictions without the users consent. No means no.

Which hasn't been an issue since Chip & PIN became required, 22 years ago (at least over here).

Comments URL: https://news.ycombinator.com/item?id=48895280

Points: 5

# Comments: 0

First there is Android's attestation framework. That does actual hardware attestation, as used by GrapheneOS, and supported by literally no app whatsoever.

Then there is SafetyNet, now Play Integrity. Depending on what level of integrity checking is being done, this will do a combination of cursory surface-level software checks, delegation to the aforementioned hardware attestation framework, and several other checks.

Importantly, SafetyNet/Play Integrity rejects some devices that pass hardware attestation (e.g., Graphene OS), and accepts some devices that fail hardware attestation (fairphone, many cheaper devices with broken ROMs, etc).

e.g., fairphone leaked the private key for their attestation, but many of their devices still pass SafetyNet, while some other devices that pass attestation but have known bootloader flaws are blocked by SafetyNet.

Because this isn't strict cryptographic verification, but a mess of heuristics and guesswork, it's a constant cat and mouse game.

What Google really achieved here is to make it expensive enough that no casual user can bypass it to e.g. cheat in Pokemon Go, but only a determined attacker has a chance.

And with "determined attacker" I'm not just talking about states, but even e.g. movie pirates breaking DRM to rip Netflix movies.

Of course, even full cryptographic attestation isn't perfect, and can be bypassed with enough effort. As shown by the famous iPhone hardware jailbreak, where you drill into the SoC and solder directly to the CPU's internal wiring.

That then allows you to do secure NFC credit card payments even on a rooted phone with custom ROM.

The proper solution is one we had with desktop computing for decades. If you keep the key material on your eID or bank card, you don't need a locked down operating system. Which then allows devices to live for much longer.

We're slowly losing the war on General Purpose Computing.

https://media.ccc.de/v/28c3-4848-en-the_coming_war_on_genera...

You do realize that a full kernel vulnerability like this allows you to feed falsified information to SafetyNet? Just like DRM, it gives the developer the illusion of control, but doesn't do anything to actually improve "safety" or "integrity".

It's silly that whenever I see a vulnerability like this, all I can think about is "finally, a way to get control over my own devices back". Once again, Stallman was right.

https://www.gnu.org/philosophy/right-to-read.en.html

Personally, I'll use this to root my Android TV and Chromecast devices and remove the shitty ads in the launcher (which Google added after I bought the devices!).

That's a good point, and personally one of the reasons I disagree with the ethnic definition of "nation".

The other reason are of course the frisian, danish, sorbian, etc and many other similar minorities that have historically lived in the region of modern Germany.

I think the french definition (which defines the nation almost entirely around the language, not ethnicity or origin) is a much more interesting and useful one. Language determines who you can talk to, and what media you can read or watch.