Frey told Ars he was shocked to see a platform as big as Discord relying on such poor support infrastructure.

It’s exactly what we can expect from tech companies these days.

When it does a different crazy thing every time and you have no idea why, it means you’re a genius and have created life.

If a physics simulation doesn’t agree exactly with experimental data, it is often difficult to figure out why and tweak the model until they agree. With AI, incorporating a few experimental examples into the training process is a lot more straightforward, and it’s not necessary to understand where exactly the model went wrong.

That’s not too bad if it’s only ever used as a rough guide in the early stages of design, with proper testing done later. But do we trust corporations not to get lazy and pressure their engineers to skip the accurate tests altogether, especially when they can then brag to their investors that AI is replacing expensive engineer time? What would Boeing’s management want to do with this tech?

There is a port to C++ underway, which still seems to be active:

https://github.com/attermann/microReticulum

I wonder what made them imagine an LLM could answer questions that involve reasoning about firewall configurations. It’s just not the kind of thing LLMs are good at, since they can’t parse something specialist like that or create a model then simulate applying the rules one by one to evaluate the end state. An LLM will improvise a plausible-sounding answer based on similar-looking questions and answers in its training data, and that’s not trustworthy at all.

The whole world can see this and form a view of Donald Trump and the USA. He’s really a textbook case of an American having no clue how they look to everyone else. This isn’t the look of a strong leader who’s in control, and taunting your enemy by making fun of their religion will not win you any sympathy or any advantage. Trump has never looked so desperate.

Just 5% said they regret their vote and would vote differently if they could.

The media are forever looking for evidence MAGA is breaking up, but this isn’t it.

Unfortunately it’s also critical for MRIs.

No, thanks for the tip. I’ll try that.

Edit: It worked! Disabling exploit protections got my bank’s app working. Thanks again.

The “advanced flow” with a one-day wait is just Google realizing they need to boil this frog a little more slowly to prevent a backlash. They still want a fully boiled frog in the end.

Mine refuse to run in GrapheneOS. Everything runs except my banks’ apps and eBay. It’s a slight inconvenience but the tradeoff has been worth it.

Edit: Thanks to [email protected] I got my bank’s app working under GrapheneOS.

I’ve been using NewPipe without problems. Is it supposed to have gone away?

We need development of Linux OSs for phones to ramp up. And we need Linux distributors and backers to fight back against “age verification” laws that are actually ID verification laws. There’s a global attack underway on multiple fronts against free software, private computing, and user ownership and control of devices.

In the meantime, for a stopgap, there’s GrapheneOS, but that doesn’t fix the problem of developers having to choose between Google’s way or unpaid obscurity.

Taxing the billionaires could pay for some of those wars you love to do.

I’d rather weed out the assholes.

From the article:

Google says it’s removing XSLT to address security vulnerabilities. The underlying library that processes XSLT in Chrome (libxslt) is an aging C/C++ codebase with known memory safety issues. Chrome’s team argues that because only about 0.02% of page loads use XSLT, it’s not worth the maintenance burden.

It’s debatable whether Google, with all its resources, really needs to do this, especially given that 0.02% of all page loads is still quite a lot. But there are certainly times when it’s better to just delete seldom-used old code from your project to lower the maintenance burden and reduce the surface area for attacks.

Previously they would have had to encounter a person who wanted to manipulate them. Now there’s a widely marketed technology that will reliably chew these vulnerable people up.

Worked fine for me, but I block ads and trackers on my home network so that probably helped.

The link sent Rademacher to a page on WebinarTV.us which featured a full recording of the Zoom recording, an AI-generated video summary of the meeting, “chapters” that sent the viewers to different parts of the meeting, and an AI-generated episode of the “Phil & Amy Show,” in which two AI-generated personalities discuss the content of the call, including quips and rapport between Phil and Amy.

So their business model is to steal other people’s meetings and add an overlay of shit? I hope it fails miserably for them.

It’s not easy, particularly if you developed it and have spent months immersed in all the detail. To emerge from that and imagine coming to it as a new user is pretty hard. I don’t have much to add but I like your advice. I need to rewrite the docs for one of my projects and I’ll be bearing your points in mind.

Maybe one other point I’d add is: have a clear idea of who you’re writing for, and have different levels and styles of documentation for different types of users. Don’t try to satisfy everyone in the same document. Divide the documentation up by intended readership.