Mastodon Feed: Posts

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

I find this almost exactly as offensive as AI-generated stuff being shoved into everything.

Boosted by cstanhope@social.coop ("The Luddites were right"):
GossiTheDog@cyberplace.social ("Kevin Beaumont") wrote:

The day after Stephen Colbert's show was cancelled, he went on community access television and did an episode with Jeff Daniels, Jack White, Eminem and others.

https://www.youtube.com/watch?v=jJTXB5uT%5FC4

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
keithwilson ("Dr Keith Wilson 💭") wrote:

“In 2024, data centres accounted for more than a fifth (22% according CSO figures) of all metered electricity in the country, which was proportionally much higher than elsewhere.” #Ireland #BigTech https://www.rte.ie/news/ireland/2026/0523/1574815-ireland-electricity-prices/

Boosted by kornel ("Kornel"):
urlyman ("Jonathan Schofield") wrote:

In Geek mythology, Gemini is associated with the myth of the twins, Faster and Bollocks

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
jmeowmeow@hachyderm.io ("Jeff Miller (orange hatband)") wrote:

@aredridel And many highly paid, highly motivated financial analysts looking for angles of attack, to be packaged by well-practiced advocates.

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

Personalization systems are actually the AI technology I have the most beef with: I don't think we've fully reckoned with the algorithmic manipulation of information they do. Suddenly the info sphere is not an object you can interrogate, but you become the object, and the infosphere around us is made into a mirror.

Weirdly, LLMs used well are actually _better_ about this because they're somewhat more able to be interrogated (though naive questioning is not probing so much why as one would hope). But the breadth of the corpus and the freedom we have to query within it are currently new capabilities. Not to say that some companies aren't well on their way to researching how to personalize that into a reflection again. Google must be stopped, and I trust OpenAI just as little. People's chats still end up adding quite a reflective lens on things, but the raw access is still there if we want to make use of it.

Boosted by brib@bribstodon.xyz ("brib :neofox_floof:​ :Nonbinary:"):
geekysteven@beige.party wrote:

brib@bribstodon.xyz ("brib :neofox_floof:​ :Nonbinary:") wrote:

I generally like #Vivaldi a lot as a browser, but... not a huge fan of the colour contrasts (or lack of them) in the 8.0 update. Sometimes the tab/URL bar text doesn't contrast well to the background (e.g. light text on light orange background), and I can't see the boundaries between tabs.

How do I get the old colour scheme back? Any theme suggestions that fix this?

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

Job loss? That one really seems weird. The numbers are not adding up to the stories we hear yet. But the numbers are not dire yet. Of course, our actual tracking of those numbers is increasingly suspect because of the US fascist project.

The big tech companies are absolutely pulling shenanigans, and using "AI" as an excuse for layoffs in a great number of cases, and in many cases, openly admitting they don't have any good ideas, so they'll lay people off instead of allocate them better. We are absolutely seeing the shifts in what are viable careers in tech though. And that was underway before AI: the UX bootcamps, for example, destabilized those job function, devaluing the workers, and the resulting shifts absolutely wrecked those job titles and the pay for them, leaving a bunch of people in the lurch.

But behind that too was our broken education system in the US: it was a shortcut around the massive debt that going to college can produce for people. But it didn't last, because it was vocational training being used purposefully as a wedge to change the labor market. Similar, sometimes intentional effects have happened in other aspects.

A huge amount of the "AI" problem is actually labor problems coming home to roost. We can blame Reagan for a lot of this.

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

Energy impacts? We must build more solar. We _are_ building more solar, and the US not being in the middle of a fascist takeover would sure help. But the problem is not really "AI", so much as that is an inciting incident, but some of the real problem is the US political structure being deeply intertwined with oil.

Looking globally, the conversion to renewable energy is well underway. Later than ideal for sure, but, we're doing it.

20% of UK power was wind last year!

California just installed _19 nuclear reactor-equivalents_ of battery capacity. That enables solar for base load generation in a whole new way.

it's not enough but the process is underway, and the economic model of it is actually really optimistic. Solar is _cheap_. And scalable such that we could just make every parking lot a solar farm. We need to be asking why not at every planning meeting. We need to be starting companies to just go do this.

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

Infosphere pollution? That was already well underway with the incentives to produce slop articles — remember the quaint critique of everything turning into listicles? It's the same effect, just with vastly increased supply now. That's a material difference, but the incentives were already there.

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

Datacenters? They're exploiting the absolutely frayed fabric of local planning process and democratic decisions there, and the ossified rules from higher levels of government that prevent localities from actually dealing with it. Plus they're preying on the ravaged small town economies that farm consolidation and manufacturing offshoring has created. Not every data center is bad, when they're integrated well in a community and not stressing its systems, and not too large and running their own generators, they're extremely boring buildings. Heat is a problem many places, and electricity's fungibility makes non-local impacts there, so we need to deal with that.

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

I really do think the "AI crisis" is actually mostly a cover for the democracy polycrisis underneath. (And, I actually think it is a useful foil for people wanting to distract from that.)

brib@bribstodon.xyz ("brib :neofox_floof:​ :Nonbinary:") wrote:

I should really catch up on Jessie Gender's videos.

So many videos I only watched halfway through (largely because they're 2+ hours long)

brib@bribstodon.xyz ("brib :neofox_floof:​ :Nonbinary:") wrote:

I keep on wringing my hands over life direction but I think I have something vaguely resembling an answer.

There's two things I could do that would probably do a decent job of scratching the "interesting meaningful work" part of my life: write countless essays about how tech companies ruin everything (basically, a more refined and reasoned version of what I keep doing on Mastodon), or actually learn how to make proper independent video games (I used to do it a lot in the past, dropped it largely for self-sabotage reasons, have been verrryyyyy slowly picking it up again). I don't think I could juggle both at once so I sorta need to pick. At the moment my writing is stronger than my gamedevving, but my friends say I actually look happy when developing/showing off games.

Neither of these options have (from the perspective of an onlooker) particularly wonderful money-earning prospects, especially not if I'm breaking into the career, so.... I also need a way to pay at least some of the bills, ideally which gives me the time and energy to pursue the writing or gamedev thing. I don't really know how to do this part, even before AI doing commercial software dev was sufficiently soul-sucking to consume my energy

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
testobsessed@ruby.social ("Elisabeth Hendrickson") wrote:

AI is shifting the economics of software development so much. 2 examples:

1. I stopped teaching ATDD because product people generally preferred to react to what they didn't want than declare what they did and the connection between human language & code was too fragile.

2. I encountered Clean Room in the 90s, but dismissed it as impractical for consumer packaged software.

Both are having a moment now because declaring intentions is now the bottleneck. FASCINATING.

Boosted by brib@bribstodon.xyz ("brib :neofox_floof:​ :Nonbinary:"):
algernon@come-from.mad-scientist.club ("Vengeful Mouse") wrote:

The new Luddite movement

...is probably not going to be featured on the Financial Times.

#algernonReviewsHackerNews

Boosted by brib@bribstodon.xyz ("brib :neofox_floof:​ :Nonbinary:"):
jerry@infosec.exchange ("Jerry 🦙💝🦙") wrote:

There's an RCE vulnerability in nginx, so go patch. There's also another RCE in nginx that hasn't been patched, so commence hand wringing and keep an eye out for the new new patch when it is released.

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

God it's weird having little beef with AI-as-technology and beef with nearly every consumer-facing instantiation of it.

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

RE: https://mastodon.social/@koen%5Fhufkens/116622529210179804

And it's not really AI that's done it, this is the long tail of ALL of the search engine's practices. It's just real notable when it drops to near zero.

Boosted by aredridel@kolektiva.social ("Mx. Aria Stewart"):
koen_hufkens ("Koen Hufkens, PhD") wrote:

RE: https://social.vivaldi.net/@everton137/116613317111584835

You work for free, we make the money. That's how AI works.

cwebber@social.coop ("Christine Lemmer-Webber") wrote:

We don't know what the courts will rule, what laws will pass. But we know what the incentives of Disney, the RIAA, have on the copyrighted *inputs* side.

And what about those running the models? What about them?

They might have a harder case, people bring up the monkey with a camera case, but this is a risk also. OpenAI's CFO explicitly said in an interview that it would like to pull a Unity, and suddenly upend the situation customers have come to expect and get a licensing cut of every piece of profit assisted by them. https://www.businessinsider.com/openai-cfo-sarah-friar-future-revenue-sources-2026-1

The problem with all this is that the legal area is *uncertain*. And it's even more uncertain because it will be ages before there is *international consensus* on how to handle these things.

cwebber@social.coop ("Christine Lemmer-Webber") wrote:

Spotify and UMG strike a deal: users can pay extra to make their own "AI remixes" of licensed artists https://www.billboard.com/pro/spotify-and-umg-strike-licensing-deal-for-ai-covers-remixes/

And so. What does this mean for FOSS?

What a question! What do I mean even?

Well what does it mean for projects which have accepted AIgen patches? UMG and spotify are going to want to make sure that users *have to* pay a premium for such a service. And so what do you think they're going to push for, copyright-wise?

That's right, if they can make a legal case that the slurry of inputs *can* result in copyrighted output, well...

What does that mean for AIgen patches applied to FOSS projects? What legal terms, from their slurry of inputs, carry forward? What about attribution, a requirement in even the most permissive of most FOSS licenses? To say nothing of copyleft, and also code released without a license, etc.

FOSS projects welcoming AIgen patches may be sitting on a ticking time bomb.

Boosted by cwebber@social.coop ("Christine Lemmer-Webber"):
travis_nice@bne.social ("Travis Nice") wrote:

@cwebber Delivered within CURLE_OPERATION_TIMEDOUT or it’s free!

Boosted by brib@bribstodon.xyz ("brib :neofox_floof:​ :Nonbinary:"):
emily_s@mastodon.me.uk ("Emily_S") wrote:

Some serious work to do on the text filtering and formatting when reading wikipedia articles but look! I searched for something and got a useful first result!

The dumbest and simplest algorithm in the history of search engines. Literally sum up the appearance counts for all the terms you typed in and return the pages sorted so the highest count comes first.

Boosted by baldur@toot.cafe ("Baldur Bjarnason"):
iris_meredith ("Iris") wrote:

My second article in three days! This bounces off @ali's recent article to discuss where exactly in the tech space is ground for LLMs to grow. My conclusion is in the title:

https://deadsimpletech.com/blog/docs-shit-bad-at-teaching

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

Note: glibc provides no wrapper for pidfd_open(), necessitating the use of syscall(2).

well that's encouraging...

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

That hideous swollen lump is my knee. Trust me, it's getting better fast!

https://freethoughtblogs.com/pharyngula/2026/05/23/recovery-update/

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

just having a totally normal summer covering a window in tinfoil to survive the afternoon :blobfoxlaughsweat:

jsonstein@masto.deoan.org ("Jeff Sonstein") wrote:

@hrheingold

I thought you might find some of this piece to sound familiar ;^}

https://terrygodier.com/the-boring-internet