Has anyone explored in a serious way the pedagogical implications of having PhD students use automation in their reasoning work? There is some evidence that use of generative AI degrades critical thinking and other cognitive skills, which is a good enough reason to ask. We could be doing a disservice to students by persuading them to use such tools. Speaking from personal experience, more than once in my life I've written a computer program to solve a puzzle I still had no idea how to solve after writing the computer program and studying its output. Computer mediation does something to the reasoning process, and I'm not so sure we fully understand the implications. People have been teaching math without such tools for thousands of years; why take on an unknown possible risk like this?

More broadly, I cannot help wondering: what's driving the urgency that demands automation in math of all things? Automating the provision of life-saving devices and medicines to people who need them has a justifiable urgency. But math? This is not to say people should tediously calculate every little thing when there are helpful tools available. It's simply to ask: is there a better rationale than variations of "avoid tedium", "avoid mistakes", "go faster" or "everyone else is doing it", which seem superficially appealing but in the end are incoherent justifications (in my view)?

I think the most important question is: why does anyone comply with what one incurious, self-centered, and self-important man, Linus Torvalds, says?

Why is that most important? Because Linux does not belong to Linus Torvalds. It belongs to the world. Torvalds only gets to dictate if the people who are best-positioned to take that power from him fail to do so. From where I sit (as an outsider), he should have been removed from the project, or at least have his dictatorial power removed, long ago, and the primary problem is that there is no structure or will to actually do that.

Under these conditions Linux is a hierarchical, and therefore right-wing, project. The project allows a single man to dictate the future evolution of a global good. Let's drop the illusion that it is "free" or "open" till that changes.

#dev #tech #FOSS #FLOSS #LOSS #linux #kernel #AI #LLM

The USA is engaged in a war of choice to defeat itself. It certainly is a special moment.

#USPol #USWar #TrumpWar #US #Iran #geopolitics

I've taken part in numerous computer science research projects that involved IRB review. I believe the system can work, even for cases like the above post. But Google and Stanford are bad actors that are taken at their word, and no system, IRB based or otherwise, is going to function properly if it cannot appropriately metabolize this reality.

I was once in a meeting where a Google X person gushed about a project that used smartphone sensor data to detect earthquakes. The calculus being set up here is that detecting earthquakes is a social good that outweighs any particular individual's right to privacy and even outweighs their right to know that their smartphone is being used in this way, let alone consent to it in any meaningful way. (*) Let me be clear and say I reject this framing, but it does appeal to a lot of people, even well-meaning ones. A mess.

#Google #IRB #ethics #EthicsReview #ComputerScience #research #ResearchEthics #tech #dev

This one falls flat on its huggingface right out of the gate.

The Alibi of Capital
How We Broke the Earth to Steal the Future on the Promise of a Better Tomorrow

Description
Today, extraordinary wealth seems to arrive from nowhere. The trick of conjuring this unearned wealth is, in fact, the key to understanding capitalism’s origins and a clue to why the catastrophe of climate collapse is upon us: value is created by consuming the future.

The Alibi of Capital explains how this came about through the imperial expansion of the West, encumbering today’s generations with repayments on earlier extractions. Timothy Mitchell identifies the forms of capitalisation, credit, and coercion that turn prospective assets into present income. Rejecting the common idea that claims on the future create only financial or fictitious capital, he traces the terraforming projects – the destruction of rivers, the colonising of territory, the expansion of infrastructure, and the burning of carbon – through which the future has been squandered. Terms such as finance, technology, the economy, and growth function as alibis that conceal this devastating form of extraction.

#capitalism #extractivism #colonization #tech #technology #ClimateChange #ClimateCrisis #EnvironmentalDestruction #EcologicalDestruction #MortgagingTheFuture

• The future doesn’t get predicted; it gets produced by prediction.
• Prophecies are, more often than not, power plays in disguise.
• There is a very close connection between surveillance, prediction, and authoritarianism.
• Surveillance sheds a light on everything but itself.
• When you make a decision on the basis of a prediction, you cannot challenge that decision, and that goes against the fundamental principle of due process.

Chinese AI model takes US tech industry by surprise with abilities rivaling Claude and ChatGPT
https://apnews.com/article/kimi-k3-china-ai-0d8a5e268deb11a673f4d444fc597cc5

See the attached chart for trend. It plots weekly token usage by US firms.

#AI #GenAI #GenerativeAI #LLM #Claude #ChatGPT #DeepSeek #China #US

Alt...

An area chart shared by Global Market Investor based on data from Open Router comparing weekly token usage *by US firms* on Chinese models versus American models. There is a mark where DeepSeek was released in early January, after which China's share steadily climbs to 50% in late March/early April 2026, followed by a bit of decline and now at or above 50%.

• Replikka: https://theluddite.org/post/replika.html
• RealDoll: https://neuromatch.social/users/jonny/statuses/116939614764592845 (and the links)

#maine #wildfires #QuebecWildfires

We have perfectly functional logical reasoning engines that can accurately self-report, and have had these for decades by this point. One way to understand some of these AI systems is as free riders: they use some of the language of proper reasoning engines to give the impression their mess of neural network Play-Doh is the real logical deal when it is not. A proper hybrid might be, as Gary Marcus has argued for instance, but that would run afoul of the "always use neural networks" dictate.

#oil #OilCrisis #energy #Iran #USPol #USWar #IranWar #StraitOfHormuz #ClimateChange #war #COVID

I believe at this stage CloudFlare provides security in a way analogous to how a protection racket provides security.

#CloudFlare #security #cybersecurity #infosec #web

If we take seriously that code is to some extent "perishable", meaning it can bitrot, drift out of date with current conditions so badly that it becomes effectively non-functional, etc., then perhaps an analogy with farming isn't unreasonable. Proposals like "just fork the project" would have you give up a robust vegetable farm that reliably produces every season in exchange for a single vegetable harvest. In principle, if you're clever and well-resourced you could seed the vegetables and buy your own land and equipment and start your own farm. Regardless, that possibility existing does not equate the robust farm with the single harvest and it's disingenuous to suggest otherwise. This is another take on the Omelas analogy but one a little closer to my heart.

I think the question "who are the farmers in open source?" is very interesting. I don't think the answer is the obvious one.

"this is not the place to debate the ethics of AI"

So, the subtext of a statement like this is they are asserting power. They proclaim that your opinion has no place in the decisionmaking process about AI. They proclaim this because they believe they have the power to exclude you, and are attempting to exercise that power.

That's one way I read such statements, personally.

@jens@social.finkhaeuser.de

#maine #dystopia #USPol #MEPol #ClimateChange #wildfires