made you look
If one day clang of gcc ships with a local LLM to lint code for security problems I would probably run it occasionally even on my hobby code despite being an AI hater.
Compilers already have deterministic linters built in, devs often turn them off anyway.
Quite a lot of the bugs being found in existing codebases show up as soon as you re-enable them.
And it’s also not like the hardware is improving much either, we’ve effectively hit the wall performance wise with processors and RAM.
That’s literally what all these datacenters are for, we can’t scale the individual computer performance up, so instead we build more and more of them.
The LLMs are wicked fast at reviewing code
That’s because it’s not actually reviewing the code, rather just producing text that statistically looks like review comments.
So it does look vaguely useful, e.g. if code has null checks, review comments usually won’t ask for them to be added, but it’s never actually reviewing, or understanding, or reasoning, or anything else people claim.
It’s like an actor playing a doctor in a TV show, they’re not actually diagnosing patients, they’re just following a script that looks like it.
“Prompting” isn’t a skill, that’s just asking for what you want. Same way as asking a barista for a drink doesn’t require the same level of skill as making it.
But in general yes that’s how arguments and decision-making works, for any law. These things aren’t mathematical proofs interpreted by computers, they’re people you convince of your point.
Ehh, it seems pretty targeted to me, it’s defined to only apply to automatic apps or services, e.g. it specifically exempts manually created images.
Bans X from allowing grok to make naked pictures of celebrities but doesn’t apply to people using photoshop to do the same.
I agree with you, but the cash example is a bad one because there is a push to move entirely to electronic payments.
There’s less microplastics now than in the 70s?
Some newer radiation hardened stuff is 10x larger than that, older gear even more so. But that just reduces the risk, not sure it’s possible to negate it entirely.
An easier way is to just include more CPUs as part of the system, run them in lockstep, then compare the results by majority rule. If 2/3 CPUs say one answer and the third says something else, you discard the result of the third and go with the other CPUs.
The ladybird devs are currently in the process of switching language again from Swift to Rust, using LLMs.
This isn’t sending your packets anywhere but their closest datacenter, not sure I’d trust MS (Or rather, Cloudflare) with your porn rather than your ISP who you’re actually paying.
The original use case for this stuff was unencrypted HTTP with a public WiFi connection, in which case your ISP is the owners of whatever shop you’re in and yeah they could see everything.
If you’re at home or whatever it offers effectively no benefits, doesn’t “block trackers” or whatever nonsense like Nord claims, but I don’t think Microsoft ever claimed that it did.
IPFS has gateways though, so you can link to the latest version of a page which can be updated by the owner, or alternatively link to a specific revision of the page that is immutable and can’t be forged.
Seems like we need to switch to URLs that contain the SHA256 of the page they’re linking to, so we can tell if anything has changed since the link was created.
IPFS says hi
It’s the same person running all of them, so yeah it is.
No, Nokia do own a bunch of patents on it, I’m pretty sure they also created (and have patents on) the HEIF format used in HEIC/AVIF as well.
Edit: search results were failing me, but here’s a couple.
https://blogs.windows.com/devices/2013/03/18/h-265hevc-high-quality-video-at-half-the-bandwidth/ https://mspoweruser.com/nokia-details-its-contribution-to-h-265hevc-hints-at-integration-in-devices/
Much in the same way that laws don’t prevent crime, a project banning AI contributions doesn’t stop people from trying to sneak in LLM slop, it instead lets the project ban them without argument.
Compared to e.g. pushing a button in VS code and having your browser pop up with a pre-filled in github PR page? It’s clunky, but that doesn’t mean it’s not useful.
For starters it’s entirely decentralised, a single email address is all you need to commit to anything, regardless of where and how it’s hosted. There was actually an article on lobsters recently that I thought was quite neat, how the combination of a patch-based workflow and email allows for entirely offline development, something that’s simply not possible with things like github or codeberg.
https://ploum.net/2026-01-31-offline-git-send-email.html
The fact that you can “send” an email without actually sending it means you can queue the patch submissions up offline and then send them whenever you’re ready, along with downloading the replies.
Sourcehut uses it, it’s actually the only way to interact with repos hosted on it.
It definitely feels outdated, yet it’s also how git is designed to work well with. Like git makes it really easy to re-write commit history, while also warning you not to force push re-written history to a public repo (Like e.g. a PR), that’s because none of that is an issue with the email workflow, where each email is always an entirely isolated new commit.
SpaceX wrote in its July permit application — under the header Specific Testing Requirements — Table 2 for Outfall: 001 — that its mercury concentration at one outfall location was 113 micrograms per liter. Water quality criteria in the state calls for levels no higher than 2.1 micrograms per liter for acute aquatic toxicity and much lower levels for human health
Cool, you can drink the mercury water, but I’ll pass thanks.
The big one for me was rsync, maintainer started vibecoding it and the first release with those changes had a huge amount of regressions.
Yet it keeps getting held up as an example of “using AI tools properly”.