Old Fucking Punk @lwriemen@librem.one

You are telling people that it's okay to type confidential information into Microsoft Copilot.

That. Is. Not. True.

Microsoft admits that they *will* violate local laws in order to share confidential information hosted on non-US servers with the US government.

There is no version of Microsoft Copilot that isn't subject to American courts.

That is the most important thing to know about LLMs!

This article starts with a story from someone who tried Claude Code and found it amazing, but then switches to the same person a few months later seeing what a disaster it’s been. This quote is key:

The problem is that code produced by an AI agent looks reasonable, but can contain ‘hard-to-spot bugs’ that end up causing major problems

LLMs, by their nature, generate statistically plausible output. That is often a set that overlaps with correct output. But the things that are not correct look exactly the same as the ones they are.

Learning to review code is hard. You look for the bugs that you expect to be possible implementing it by using your theory of mind for the person writing the code and the kinds of things that they might overlook (not necessarily a specific person, but the kinds of things people miss) and also common bug classes.

And the big help is that the person writing the code is not thinking adversarially. They are not trying to sneak bugs in. Normally. Unless they’re a supply-chain attacker, and we’ve a depressing amount of evidence that code review doesn’t catch supply-chain attacks.

An LLM is not trying to do anything. It has no intent. But it is a machine that is trained on code that made it past code review. The kind of bugs that it will generate are ones that look like code that appeared in production. This is exactly what an attacker would do: try to write code that looks correct but is subtly wrong.

I’m only being slightly flippant when I say LLMs are a mechanism for brining supply chain attacks in house.

https://xcancel.com/aakashgupta/status/2087036537930424351?s=46&t=

Maybe the Pentagon should start a GoFundMe to get more missiles. https://patreon.com/BrianMcFadden

As I wrote on Destructured, AI companies' current investment in data center and generator construction is the single best metric for assessing AI's future energy demand: https://futurism.com/artificial-intelligence/amazon-data-center-largest-pollution-source-texas-usa Anyone who suggests that the energy consumption of AI is overstated is ignoring exhibit number one.

«they simply replace skill barriers with dependence on proprietary services, while users of these services often trade full community participation and all its benefits for interactions mediated by LLMs trained on community-built open source projects. Our communities are already noticing a shift away from collaborative stewardship and toward transactional interactions with projects»

If FLOSS keeps accepting this state, it's digging its own grave.

https://greatscottgadgets.com/2026/06-17-llms-generate-code-communities-build-projects/

“it’s just a tool and the future and you’re an SJW for not wanting it” torvalds screamed over the sound of rsync, a previously very stable C codebase that lots of infrastructure relies on, melting into vibed slop that can’t reliably do anything anymore

it’s fucking wild to see another foundational project break so hard a bunch of dependencies either dumped it or stayed on an old version and go “oh me too” and implement the exact same “tool” and development processes built around that “tool” and not see the train about to hit you when your releases start to look exactly like that other project’s releases

you know, assuming software engineering wasn’t just a lie we told ourselves:

in software engineering, unintentionally changing release cadence and having a ton more shit than you planned in each release is a sign that something is direly wrong with how your project is being run

and again I’ll point out how fucking quick this shit is happening

torvalds posts a defense of AI slop in the kernel with a right-wing conspiracy theory as the reason

a bunch of other open source projects just go “well the kernel says it’s ok so we’re doing slop now”

barely a couple of weeks later the kernel is changing its release cadence and quality standards entirely and the man in charge is shrugging and literally going “what can you do”

you can cut the crap, asshole. it’s your literal job to cut the crap.

nobody requested this, nobody wants gigantic kernel updates with very low quality code in them. the kernel of all things needs a slower pace and more care, because it’s the fucking kernel.

so who do you imagine is holding torvalds’ feet to the fire on this? maybe it’s the LLM companies pouring fucktons of money into the Linux foundation? nah can’t be. must be all the downstream projects that are going to get fucked over by this, they surely want the worst possible version of Linux

Newsletter: Firms close up shop amid a continued crypto winter, a Trump venture ditches its crypto plans, and the crypto industry–friendly Clarity Act market structure bill fails to reach a vote before August recess — likely dooming it.

https://www.citationneeded.news/issue-108/

#crypto #cryptocurrency #USpolitics #USpol #CitationNeededNewsletter

To teach my students the fallacy of "I have nothing to hide", I simply ask them to tell the whole class about the last time they masturbated.

It is surprisingly efficient, they immediately grasp it. I never had someone argues with me on the subject after that.

ICE plans to spend $20 million tax payer dollars to give gloves that deliver electric shocks to their agents.

They’ve murdered three US citizens.

They’ve killed countless immigrants.

They’ve used excessive force to demean, terrorize, disable & harm.

They’ve traumatized entire cities.

They’ve forced people to hide in their homes for fear of being rounded up and disappeared.

They don’t need gloves that can deliver electric shocks, they need to be abolished and prosecuted.

They need to be stopped.

https://www.huffpost.com/entry/ice-shock-gloves_n_6a7bac5fe4b07b8d62257705

#uspol #ice #abolishice #immigration #fascism

Speaking to people in the 'conventional tech services' industry, I get the impression that *nearly all of them* hate the technologies - mostly from US Big Tech - they feel forced to push. Seems soul destroying. People talking about shifting gov't departments from IBM to Microsoft dependence to save $15million annually. When, in most cases, they could be saving almost all of it by using a local provider running libre software... Seems like an industry swamped with cognitive dissonance.

“We are constantly told that universal programs are unrealistic while enormous subsidies, tax breaks and giveaways to corporations are treated as the natural order of things.”

Francesca Hong talks democratic socialism, worker power and governing with Rep. Shaun Scott.

https://inthesetimes.com/article/francesca-hong-democratic-socialism-wisconsin

Hello Fediverse - I'm looking for libre/FOSS/open source Sysadmim, DevOps, & Dev contract work while our (with @isaac) new project, Magnificent.nz, is ramping up...
Capabilities https://davelane.nz/abilities
CV https://davelane.nz/CV
Repo https://forge.davelane.nz
Refs on request!
Any libre leads, worldwide, gratefully accepted!
#GetFediContracted #GetFediHired

@glyph I'm Danish. In my culture, if someone praises you like an LLM chatbot does, that means they're either 1. lying through their teeth and actually despise you, 2. trying to sell you something you both know you don't want, or 3. a tourist.

Put out of the cartoon business by actual photographs of our decaying mess of a President.

AI promises so much, but one of the things it truly delivers is fraud at scale. Example: https://www.404media.co/company-offering-100-human-written-never-ai-peer-review-is-entirely-ai/?ref=daily-stories-newsletter

I get the strong impression that there's broad disillusionment across the 'conventional' (i.e. proprietary) side of our industry - the feeling among practitioners that they've sold out, and are mercenary and largely uninvested in what they're doing. They're biggest wins appear to be 'making things slightly less bad', e.g. the move from IBM (obscenely expensive) to Microsoft (extortionately overpriced), as opposed to doing 'good'. Here's hoping https://magnificent.nz can help turn that around.