Felienne Hermans

something incredible about techies not being able to comprehend a website that isn't a startup

thanks voor de tip Rabobank!

Lula won!!!!!!!!!!!!!

One thing that pisses me off to no end about AI bros in academia is when they cherry pick their favourite problem, where AI may or may not actually work, while proceeding to ignore the global net negative effect of AI usage on the rest of the world. (Full ivory tower, but pretending it’s not.)

"Coding is solved"

but in truth all of us have this property

The tragedy of the boom and bust cycle is that you are not allowed to say "this is a boom and bust cycle" during the boom part, even if it is obvious. Meanwhile the people who were saying "this boom will last forever" will now smoothly pivot to "the bubble has popped, like I always said it would"

The sharpest executives in the industry are noticing this only 2 years after ordinary workers. This is why they get paid the big bucks.

One of the many things I love about El-Sayed is that he's willing to engage with someone hostile if it gets them to admit the premises behind that hostility, to the extent that it becomes impossible to avoid the underlying bigotry. You want to go there, fine, but he's gonna make you own it.

We hadden weer de eer om @felienne.bsky.social te gast te hebben in de podcast. En we kregen even een spiegel voorgehouden. Leerzaam maar ook leuk!

> this is the same man who berated people for suggesting using C++ in the kernel back in 2007, because “a lot of substandard programmers use it, to the point where it's much much easier to generate total and utter crap with it.” More at sfba.social/@drahardja/1...

They should also take up meth.

Ook deze week in de nieuwsbrief...! Jouw kans om het mooie nieuwe boek van Jacquelien Bulterman te winnen. Stuur me jouw grootste faal voor de klas, en win!! (Mag ook als reply hierzo!) www.felienne.nl/2026-41/#de-...

Surely people's response to this will be "we need more powerful AI to prevent hacks like this" rather than "we need fridges to not have wifi" or at least "we need fridges to have a local override button" www.bbc.com/news/article...

Which is why Wikipedia famously uses "I" in all articles

The tone of LLMs only makes sense when you consider that they are intentionally set up to fool you into thinking there is a mind there. (And i guess also when you consider that they are designed to spend as many tokens as possible.)

In fact, seen in this light, it’s clear that the impersonal and non chatty tone of the ship computer is a better design for the task it was set up to do. How insufferable would it be if it was constantly responding to users with “That’s a great question. I can help you with that.”

The ship’s computer in Star Trek is a much better comparison to LLMs than Data from TNG or the Doctor from Voyager. The makers of LLMs desperately want us to think of those characters because that’s more exciting, but their technology is closer to the non-sentient computer that’s just an interface.

I agree and it always makes me think of how people interact with the ship’s computer in Star Trek. It never uses first person pronouns. Instead of “I can’t do that,” it says “unable to comply.” Instead of “I don’t know,” it says simply “unknown.”

So much of the trust people give chatbots comes from that "I"

Lived Experience of Software Engineering. An independent research group studying the human experience of making software. leselab.org via @felienne.bsky.social

If we want to regulate AI, why not start with a simple ban on LLMs using "I"? Easy to build, easy to verify and deflates a hype while still allowing people to use chatbots (so no one can claim we are limiting anything) My latest for @volkskrant.nl in English! www.felienne.nl/2026-10-08/#...

Hier is te weinig verontwaardiging over dus ik gooi hem er nog een keer in. De vuurwerkbranche wordt gecompenseerd (!) met 80 miljoen die voor spooronderhoud was bedoeld en 20 miljoen uit de klimaatpot. WTF. www.nrc.nl/nieuws/2026/...

Dutch editorial cartoonist Martijn van der Wiel finds it barbaric that Defense Secretary Pete Hegseth has ordered the firing-squad execution of convicted Fort Hood shooter Nidal Malik Hasan to be live-streamed for the public to watch.

Ik zou het wat preciezer zeggen: programmeurs maken ze zo dat ze een ik-vorm gebruiken. Ze doen het niet uit zichzelf, dat is nu juist de personoficatie waar we ons tegen moeten verzetten!

Computer Science Academia: There's nothing we can do. Let's accept and do research with LLMs. Mathematics Academia: We did not ask for this work to be done. OpenAI is an evil company. Mathematicians have a particular vision of progress that is informed by history and field-specific considerations

Man who won’t give his employees piss breaks claims many piss breaks on the horizon

Anthropic banning “abusive or cruel behavior” toward Claude in its new usage policies is proof to me that tech companies are going to moderate violence against AI before they ever moderate violence against women and minorities

ah damn, RIP Margaret Hamilton. she took safety-critical software engineering seriously, and got people to the moon and back. news.mit.edu/2026/margare...

the tech industry is about to invent books

NEW: October 17 marks a global event called “Origin,” focused on Earth regeneration. Rather than a movement to directly influence climate policy or polluting companies, Origin promises change through synchronized prayer and body movement. Behind it? A fossil fuel heiress and an anti-vaccine doctor

one of the ways the media shows which side they prefer, is which politicians they quote uncharitably out-of-context and which they clean up for even when what they actually say and do is offensive (the Biden story ran on the front page less than a week before the 24 election)

Gavin Newsom after vetoing another anti-discrimination bill

Hard to think of anything less abstract and more materially consequential than being pregnant