hashfyre

hashfyre

@hashfyre.dev

Discordant demi-geek. Makes maps. Eats myths and fantasy. My politics is class politics. Burnt-out Leftist Techie who is probably going to quit being one. Opinions are solemnly mine. http://pronoun.is/he +LSN-ASD

649 followers764 following3.5K posts

Pinned

I am really shit to AI users, promoters, re-skeeters, justifiers, and tech-bros. I'll not be nice to you or yours. IDGAF about how you felt about our interaction. See this post, block and never interact with me, if you fall in those categories. I'll be downright cruel to you.

hashfyre reposted

Something benefiting you personally is not necessarily a good thing. Something that inconveniences you personally is not necessarily a bad thing.

hashfyre reposted

hashfyre reposted

"AI agents are agents of the person or organization that deploys them—and should be treated by the law as such. ... To allow businesses to hide behind the excuse of faulty AI ... would introduce disastrous incentives for corporate misbehavior" www.schneier.com/blog/archive...

schneier.comAI and Liability - Schneier on SecurityEarlier this month, a German court ruled that Google is liable for its AI search summaries. Rejecting defenses like “users can check for themselves,” and that they generally know “that information gen...

hashfyre reposted

Replying to @mewefree.bsky.social

This is nuts. He was already placed in a league with 5yo. Saw in an article on huffpost that pro footballer Eddie Johnson has enrolled him in his youth soccer program for the winter because of this incident, so this kid hopefully can move on from these jerks and continue having more fun.

hashfyre reposted

Comrade Parvatibai Bhor: A Fighter's Story October 30, 2026 5.30 - 7.30 pm Press Club Parvatibai Bhor went from a child bride at 9 to a "lal bawta wali", reading Lenin & leading strikes in the 1930s. Her autobiography has just been translated into English by senior journalist Lina Mathias.

hashfyre reposted

restaurateur says you should taste the shitburger before you decide how you feel about it

The A.V. Club@avclub.com

Ava DuVernay's 14th, about the 14th Amendment, uses AI to animate old photos that are in the public domain, and the filmmaker wants people to see the documentary for themselves before deciding how they feel about it. tinyurl.com/c6nj6vf7

hashfyre reposted

If you're an indie Indian music fan with some basic graphic design / copywriting / social media chops who's currently un- or under-employed, hit us up, we have an idea we're working on

hashfyre reposted

The 👏 Point 👏 Of 👏 Education 👏 Is 👏 Not 👏 To 👏 Make 👏 Sure 👏 Enough 👏 Essays 👏 Are 👏 Produced 👏!!

Carl T. Bergstrom@carlbergstrom.com

"AI can draft an essay for you, especially if you give it a high quality prompt." Really? This is what MIT thinks high schoolers should learn? MIT's "Day of AI" dayofai.org provides materials for K-12 teachers and claims to have reached over 5,000,000 students to date. 1/2

hashfyre reposted

So much to say about Brazil, not least of which is the role that Erika Hilton, a Black trans leader and most voted woman politician in Brazil, played in mobilizing folks for Lula. She was behind the critical, concrete, economic promise of shorter work-week. theprint.in/feature/erik...

theprint.inWho is Erika Hilton? Black trans politician was one of Lula’s loudest campaignersErika Hilton’s electoral triumph comes amid a hostile environment. Brazil has the highest number of reported murders of transgender people in the world, with 80 recorded in 2025.

hashfyre reposted

I want all trans people to know that Erika Hilton is a black transwoman and the most-voted congresswoman in the history of Brazil, with 1.5 millions votes. She campaigned for Lula over the last three weeks. WE TRANS PEOPLE CAN EXIST IN POLITICS. It’s not a dream. It’s already a reality in Brazil!

hashfyre reposted

A new peer-reviewed estimate of what data centres will pump out in climate pollution - Between 300 to 1,300 mtco2-e per year, by 2035 (IEA scenarios) Mckinsey scenarios: - 3,000 by 2035 - 30,000 by 2045 arxiv.org/pdf/2608.14421 limited.systems/articles/ai-...

CJ@mateloos.be

“Zelfs de uitstoot volgens de conservatieve IAE-scenario’s met beperkte groei ligt in de orde van grootte van 1 GtCO₂e per jaar en is onaanvaardbaar in het licht van het wereldwijde koolstofbudget”, schrijft [professor Wim Vanderbauwhede] in zijn paper. 48hrs #Climate #AI gift link:

hashfyre reposted

The right doesn’t fight in good faith. They have proven repeatedly that they are fuelled by raw hatred against imaginary enemies made out of very real people. Conceding to them does no good. Ever. No need to be kind to racists, please.

George Monbiot@georgemonbiot.bsky.social

The lessons from Brazil and Serbia? Don't concede one inch of ground to the far right. Don't appease. Don't triangulate. Don't grovel. Don't say they have "legitimate concerns". Mobilise your base, and fight them with all the democratic means at your disposal.

hashfyre reposted

NEW: "Donald Trump was an active participant in Jeffrey Epstein’s sexual abuse of young women during the mid-to-late 1990s," per new House Judiciary Dem report. Report has new allegation that Trump raped a 24-year-old and then promised the victim a visa & NYU admission, demanding she stay quiet.

hashfyre reposted

This is probably the best ad I've seen against using generative AI: Mayer, who was an engineer at Google and a chief executive of Yahoo, has rotted her brain so much she can no longer classify her daughter as "empathetic, creative and an animal lover". The slide of competence is incredible.

David Gura@davidgura.bsky.social

cool, cool

hashfyre reposted

Literally two of my three most pertinent character traits are “ADHD” and “vehement loathing of generative AI and the ghouls pushing it”

This post is unavailable.

hashfyre reposted

Replying to @disabilitystor1.bsky.social

And also, like, I know literal dozens of neurodivergent parents who could answer those questions in five seconds using their brains, which are perfectly functional for things like "knowing and loving their kids".

hashfyre reposted

Replying to @70sbachchan.bsky.social

This really cant be said enough: Bolsonarists are so anti-national interest that it sounds unbelievable for a rightwing party to be so unpatriotic They begged Trump to sanction & tariff their own country & cause economic pain to millions of Brazilians so their guy could win. x.com/rojasrjuand/...

hashfyre reposted

Replying to @calligiovanini.bsky.social

Especially if you are American, voting alone won't save you. You don't have a left wing party. The Democrats are making anti trans ads. Erika Hilton, a black trans woman, is the most voted congresswoman in Brazil and this is a video of Lula kissing her on the cheek.

hashfyre reposted

A wrong lesson to take from the Brazilian elections is that voting can defeat fascism. That's not true at all. Democracy is an incredibly frail system and the US almost took over the country through it. ORGANIZED PEOPLE DEFEATED FASCISM

Calli Giovanini COMMS CLOSED@calligiovanini.bsky.social

It's another day, so let's talk about this. First of all, Lula didn't win the election alone. He won because more than half of the country FREAKED THE FUCK OUT and built a popular movement that organized several actions both irl and online.

Cue the centrist bullshitters. "Profound" says a random bearded man, grapsing at authenticity.

404 Media@404media.co

Despite how "shitty" chatbots can be, we're already seeing AI change the way our world works. This week on the pod, @emanuelmaiberg.bsky.social sits down with longtime tech reporter and author @mikeisaac.bsky.social to talk about the AI bubble. Watch now: www.youtube.com/watch?v=1sNB...

Tech media can't seem to pick a fucking lane, and will keep playing both-sides until the dust settles.

hashfyre reposted

Guide two children through a hand-painted nightmare. - Explore painted woods - Outwit deadly traps - Escape the witch It's called Gretel & Hansel. Would you enter the forest? #indiegames #WhatToPlay #PointAndClick #HandDrawn #FairyTales

hashfyre reposted

Your monthly reminder that if you're using ChatGPT to do your science, you're failing your colleagues, community, and all of us. Here, ChatGPT makes up data and presents it as the real thing — then apologizes when confronted.

hashfyre reposted

If you are a cybersecurity reporter looking for a topic for an article I’d say there are four things I haven’t seen much written about: * Surge in Japanese breaches * South Korea targeted by Qilin (& NK affiliates) * France cybersecurity woes * U.S. healthcare sector hammered