Mastodon Feed: Posts

Mastodon Feed

pzmyers@freethought.online ("pzmyers 🕷") wrote:

The "anti-war president" now says the only thing the federal government should do is war. Definitely not child care. Or medicaid/medicare. Or science.

https://freethoughtblogs.com/pharyngula/2026/04/04/the-man-is-insane/

https://www.youtube.com/watch?v=%5FGpxR8D3wK0

Mastodon Feed

dysfun@treehouse.systems ("gaytabase") wrote:

as they say,, a good landing is one you can walk away from, a great landing is one where they can reuse the plane afterwards.

Mastodon Feed

dysfun@treehouse.systems ("gaytabase") wrote:

(yes of course the plane was a write-off)

Mastodon Feed

dysfun@treehouse.systems ("gaytabase") wrote:

nobody died though. nobody even suffered any injuries somehow. incredible

Mastodon Feed

dysfun@treehouse.systems ("gaytabase") wrote:

LOL, most boeing story ever. they positioned some circuit breakers where they could easily get knocked, saw it happened and offered a solution only as an optional extra. this later caused a crash landing.

Mastodon Feed

pzmyers@freethought.online ("pzmyers 🕷") wrote:

Minnesota…"April". It does not mean what you think it means.

https://freethoughtblogs.com/pharyngula/2026/04/04/this-was-the-morning-i-was-supposed-to-sleep-in-then-grade-papers/

snowy yard

Mastodon Feed

EmilyEnough@hachyderm.io ("Emily 🏳️‍🌈🏳️‍⚧️") wrote:

I have never heard of Ethel Cain before her tdov instagram post, but my takeaway is this:

Damn I wish I could sound like that when I sing. I was singing bass in the choir at 14 though, so I don’t think there’s any amount of voice training in the world that could get me into her range.

Mastodon Feed

dysfun@treehouse.systems ("gaytabase") wrote:

state nomad

Mastodon Feed

Boosted by soatok@furry.engineer ("Soatok Dreamseeker"):
yuril@furry.engineer ("挧日 | Yuril") wrote:

RE: https://mastodon.sakura-star.net/@KitsuneofInari/116336733010511672

"No programmer really fully understands the code they are working with" is the most truest part of this

Mastodon Feed

Boosted by soatok@furry.engineer ("Soatok Dreamseeker"):
halcy@icosahedron.website ("halcy​ @ :revision_ol:") wrote:

on this laptop, I have a choice of many WSL distros, such as Round Ubuntu and also Square Ubuntu

two WSL distros in Windows Terminal, one uses a round logo and one a square one, both Ubuntu (they are in fact actually the same)

Mastodon Feed

dysfun@treehouse.systems ("gaytabase") wrote:

it's been a week since i wrote code and i'm getting a hankering to do something.

tempted to get back to the HOT, even though c is kinda impossible for data structures.

Mastodon Feed

dysfun@treehouse.systems ("gaytabase") wrote:

inadvertently lewd graph, youtube

diagram of performance vs arousal. it's a semicircle through under arousal, optimum arousal and over arousal.

Mastodon Feed

Boosted by slightlyoff@toot.cafe ("Alex Russell"):
maxleibman@beige.party ("Max Leibman") wrote:

Between the conservatives’ desire to avoid giving anything to the undeserving poor and the liberals’ desire to avoid giving anything to the undeserving rich, there is no social good we can’t come together as a nation to stop.

Mastodon Feed

Boosted by slightlyoff@toot.cafe ("Alex Russell"):
SeanCasten ("Sean Casten") wrote:

I will stop quoting Ozymandias in relation to the Trump administration when they stop running their administration as a direct rip off of Shelley's pithier more thought-provoking original work.

Mastodon Feed

Boosted by slightlyoff@toot.cafe ("Alex Russell"):
filippo@abyssdomain.expert ("Filippo Valsorda") wrote:

There was no good way to see what CT logs are actually used by CAs, so I made a dashboard of Censys data on exe.dev.

There are some interesting patterns, but the main one is that Let's Encrypt is the only CA that evenly spreads load. Other CAs are mostly using older logs, or their own logs and Google's.

(Of course, LE is 50% of issuance, and GTS is 25%, so the rest don't matter much.)

https://groups.google.com/a/chromium.org/d/msgid/ct-policy/718571cb-a841-4102-bcfa-3fe3feab63ae%40app.fastmail.com

Mastodon Feed

Gargron ("Eugen Rochko") wrote:

RE: https://mastodon.social/@pojntfx/116345677794218793

This is not acceptable, plain and simple.

Mastodon Feed

aredridel@kolektiva.social ("Mx. Aria Stewart") wrote:

You can also tell that the text will be bland and inoffensive and say nothing of interest beyond those facts.

Mastodon Feed

aredridel@kolektiva.social ("Mx. Aria Stewart") wrote:

I love how “high quality, topical AI generated image" means factually accurate and completely oblivious to social context in an article. You can see someone's epistemic position so nakedly now.

Mastodon Feed

dysfun@treehouse.systems ("gaytabase") wrote:

built by clowns

you will never guess which aeroplane manufacturer this is about

Mastodon Feed

dysfun@treehouse.systems ("gaytabase") wrote:

mount everest is way more dangerous than you'd think

no i'm pretty sure i filed it under "ludicrously dangerous" already.

Mastodon Feed

baldur@toot.cafe ("Baldur Bjarnason") wrote:

Cat photos for Easter #caturday

Grása, a very furry grey and white cat, contemplates the snow that's covering my neighbour's driveway
A colour photo of a black and white cat sitting in contemplation about why our spring is covered in snow.

Mastodon Feed

dysfun@treehouse.systems ("gaytabase") wrote:

petition to rebrand LLMs as 'mechanical bastards'

Mastodon Feed

baldur@toot.cafe ("Baldur Bjarnason") wrote:

More about the preview

https://www.baldurbjarnason.com/2026/easter-sale-and-the-toy-factory/#also-included-a-preview-of-my-upcoming-book

Mastodon Feed

baldur@toot.cafe ("Baldur Bjarnason") wrote:

The preview to The Magic Toy Factory is the first of my ebooks where I use Typst (via pandoc) to typeset the PDF version and it’s honestly been excellent.

https://typst.app/

Mastodon Feed

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
osnews@mstdn.social ("OSNews") wrote:

How Microsoft vaporized a trillion dollars

This is the first of a series of articles in which you will learn about what may be one of the silliest, most preventable, and most costly mishaps of the 21st century, where Microsoft all but lost OpenAI, its largest customer, and the trust of the US government.
↫ Axel Rietschin

It won't take long into this series of articles before you start won

https://www.osnews.com/story/144729/how-microsoft-vaporized-a-trillion-dollars/

#Microsoft

Mastodon Feed

baldur@toot.cafe ("Baldur Bjarnason") wrote:

“Miseducative Experiences”

https://2ndbreakfast.audreywatters.com/miseducative-experiences/

> But to frame any opposition to technology as a "moral panic" is a rhetorical sleight of hand in which one side gets to invoke "science" and "research" while dismissing the other as mere "hysteria."

Mastodon Feed

adele@social.pollux.casa ("Adële 🐁") wrote:

Web eco-design

Best practice guidelines

Our digital uses already represent 40% of our annual sustainable budget, both in terms of greenhouse gas (GHG) emissions and the depletion of the last reserves of non-renewable resources (abiotic resources). That's 10 times too much!

Eco-designing digital services allows us to better control and limit their environmental and health impacts...

https://rweb.greenit.fr/en

#smolweb #ecodesign #smallweb

Mastodon Feed

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
KitsuneofInari@sakura-star.net ("宮城巴惠") wrote:

Krita’s Maintainer is awesome!

Krita Maintainer Halla posting that reads: “Speaking as the Krita maintainer, I don’t want us to accept any LLM-generated code into Krita. Some may say, as long as the contributor understands the code it’s fine, but that’s a fallacy. No programmer really fully understands the code they are working with (otherwise bugzilla would be empty); but with LLM-generated code they haven’t even put any thought into it. That code will be unmaintainable. And that’s even completely apart from the fact that using LLM’s to generate code lowers your cognitive capacity, burns up the planet, destroys water sources all over the world and takes away from people what fits them best: being creative together with other humans. The right thing to do could be to, after you’ve experimented in this environmentally-disastrous way, to describe the feature you reached and wanted and make a wish item in bugzilla, or discuss it here, and then let people who can code implement it in a responsible way.”

Mastodon Feed

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
jonny@neuromatch.social ("jonny (good kind)") wrote:

It was only a matter of time before some AI-addled guy saw me quoting fucked up parts of the Claude code source, thought it was my code, and say it was just a skill issue and I just need to {prompting superstition} but it finally happened.

Mastodon Feed

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
jonny@neuromatch.social ("jonny (good kind)") wrote:

man what is going on.

[conversation thread between tuban__muzuru (t for short) and me] t: Well, what can you expect from a code vending machine anyway. I've been writing with Claude for a year. See, n00bs ask for code. Pros ask for spec. I've been coding for 40 years now and Claude writes great Rust and C t: Show me the spec you gave Claude to write this menu. me: Bro before you try going down the skill issue route, fucking anthropic wrote this menu and this is the Claude code source. t: After all that work defining EFFORT_LEVELS as the single source of truth, this function just... writes them out again. The includeMax branching is the tell — whoever wrote this didn't think to reach for EFFORT_LEVELS.slice(0, -1) vs EFFORT_LEVELS. const levels = EFFORT_LEVELS.slice(0, includeMax ? undefined : -1);
me: Yeah! It sucks really bad! Anthropic must really suck at using Claude code to write Claude code and they should have just written more markdown files that tell it to write better code! t: Say there, little buddy - you're sposta write the spec. But I use Claude to hammer our my spec before I do things. You should, too. t: Here's why you need the training, not the LLM: this is no different than riding a big horse. Data first, talk to the LLM about your data and your users and their needs and what the work you're doing is trying to solve. Then you start writing spec. The LLM will restart each time, like fucking Groundhog Day. You will have to lead this new instance to your decisions/ and tell it to read everything. That is the ONLY WAY TO DO THIS. me: Man what part of THIS IS THE CLAUDE CODE SOURCE CODE AND IT WAS WRITTEN BY ANTHROPIC is not getting through here
t: Stop yelling. You fucked up and left some dead code and had the temerity to paste it in here me: Man. This is not my code. This is anthropic's code. This is the source code of Claude code. How many different ways can I say this. t: You asked for the code. In two passes. It also explains why the crack is so localized. If the values were scattered everywhere you'd be looking at the "fractally bad" scenario. Instead it's one function that predates the abstraction. That's a totally normal and forgivable state for a codebase to be in — it just needs a one-line cleanup and ideally a lint rule or PR comment culture of "if you touch EffortLevel, check for strays." me: I did not ask for anything. this is the source code for Claude code. this code was written by anthropic. i am reading their code. I did not write this code. Anthropic wrote this code. This is the source code for the Claude code tool. This is the Claude code source code. This is not my code. I did not write this code.