the official explanation given for why the rayban branding was dropped is nonsense (trying to say it's about luxury pricing when the cheapest of these is still more than a pair of normal raybans) so i'm going with the more likely explanation of "they finally noticed the pr disaster in the distance"

AI is forcing a wedge between those who see any given craft as a means to an end vs. those who have learned to enjoy and value the skillful execution of the craft itself

I really think that unless something changes very soon active--effective--direct action that will stop data center build outs and functionality is inevitable.

you can already see this with some kids who grew up on sites with (intentionally) terrible search functions like instagram, and to call them deeply incurious would be putting it politely if a guide or literal person does not tell them what to seek out, they are not physically capable of doing so

will admit it's really funny they're doing the tech/gaming company strat of providing a service that should be free, "initially for free" so they can charge an offensive amount after enough people are locked in

i'm very fascinated by more recent developments in these malware attempts there's such a huge boom in late zoomers who didn't get a real laptop/pc until very recently, and i've noticed a *lot* of tricks, this included, designed to prey on them not knowing the usual quirks of a desktop experience

i am dead serious when i say this: put fluorescent, reflective orange on these and i would say this is the exact way i want all smart glasses to be regulated to look like

also, firefox updates routinely un-flip the kill switch, without your consent, so that mozilla can try pushing a.i. garbage on you again, something people regularly complain about this makes any numbers mozilla provides inherently untrustworthy

this is an excellent comparison, in that cars are really terrible at their purported main purpose compared to other alternatives, scales poorly to mass-use, and would only last if modern infrastructure effectively forces you to use one against your will

in a timeline where the government wasn't just doing this to arbitrarily punish a company that did something they didn't like, trying to make the legal argument that *all* large language models are like this, and *none of them* have stable guardrails, would've had some very funny consequences

i've gone to multiple workshops and presentations in recent months where it is *really* obvious when a.i. was used, and you could visibly see the moment each individual person in the crowd realized it was, one-by-one, until the entire group is mentally checked out

a lot of the current moment can simply be described by everything in the economy having turned into a sort of bitcoin

This is the actual public sentiment around AI. Every corporate survey I’ve encountered is careful to ask “how we can help you use AI” instead of “how do you honestly feel about AI” for precisely this reason.

yeah it's weird that we've already had a above-13 age gate on most of the internet, and now that someone sharpie'd over the 13 with a 16 we have to pretend this necessitates a new surveillance state but unfortunately we need a few more months of social media discourse to draw the line on this one

as a general rule, whenever this idea is proposed, you should assume - the one proposing it knows it won't accomplish anything healthy - they are very well aware that doing so provably enables harassment on others - they want it because *they want to do the harassing*

every once in a while someone remembers that large language models can't differentiate data and instructions, and then everyone is shocked, and then 3-6 months later someone else remembers that large language models can't differentiate data and instructions, and then everyone is shocked, and then---

everyone who lives here will agree to this, and we'll get a test project reaffirming that it's a good idea, and then 10 of the least pleasant people you'll ever meet will get it scrapped because they are not capable of imagining a world where they have to leave their f150 to experience joy

Your periodic reminder that people are *more* abusive online when they're using their "real names" than when they're using a persistent pseudonym and this effect has been replicated in literally every study that has ever been done

one thing i think bluesky users (being old enough to have deeply tech-incompetent parents) struggle to understand is how phones/computers, in 2026, are mostly a tool of surveillance to let parents spy on their kids all day the era of the internet being a place for independence died a while ago

remember when facebook artificially inflated the viewcount of its videos and everyone obviously knew this was the case (you'd have videos with >1m views but very few comments, indicating a "view" didn't involve much interacting) but every media company decided to lose billions on humoring it anyways

microsoft has realized that people who play games hate a.i. and everything it stands for, and they're already stripping the copper wire off of all the other actually-successful things so they can fund more a.i. that doesn't make money, so this is just a two birds one stone situation

this is up there with the fact that dishwashers are better for the environment than handwashing, in terms of being very obviously true if you think about it for more than 5 seconds, but for some reason every single person you'll ever meet will be very hostile to the possibility

i know of someone who didn't have a single means of accessing an online space, during an era in which everyone had social media, until they were 16 and perhaps not so coincidentally they got so one-shot by instagram reels that they basically needed their partner to ban them from the app forever

bluesky has a very natural protection against teens being on here, in that they'll have to exist on the discovery timeline and be harangued by 45 year olds until they leave

New: Researchers have quantified how easy AI search is to manipulate. Just 13 words buried in a random Reddit comment can poison AI search results. They suggest this is not easy to stop: "The way you can attack these systems is so much dumber than you think it is" www.404media.co/it-is-trivia...

this is your routine reminder that at most, 10% of gimp's lack of popularity is due to anything in the application itself, and the other 90% is because nobody is using a program named that

One of the main (only) reasons I've advocated for tech ppl to engage with the technology is that while some continue to marvel stupidly at what it does, if you critically engage with it for any amount of time it's really deeply disillusioning

also note, the moment they went from marketing chatbots as quirky little personal assistants to something they needed to sell as drop-in replacements for real workers, companies were suddenly very keen to strip all female aesthetics out, sometimes replacing them with more "useful"-feeling male ones

There's a common reactionary anti AI claim that the shift to agentic AI is because it increases token use and it's dead wrong. It's much dumber than that. If you ask an LLM what time it is, it can't tell you. If you give the LLM access to the time command that you could have run yourself, it can.

people with actual positions of authority believe their marketing for barely even a few hours and, mysteriously, all the self-consciousness of the computer disappears and it goes back to just being some model funny how that works out

will vouch for this; i knew i had cognitive issues triggered by being upright, but doctors wouldn't believe me because my "resting heart rate" was >110bpm, so i was "naturally unhealthy" fitness tracker data revealed my resting to actually be *70*. i had p.o.t.s. so bad sitting upright triggered it

something that never crops up in these conversations is just how much teachers abuse a.i. as their "cheating" students are the number of times i'll see people complain about their students supposedly abusing it, *as they're talking about feeding essays into one*, is genuinely embarrassing

really is beautiful how even the slightest whiff of possibility that you could lay off a few workers --- not even evidence, the *possibility*, an i.o.u. of usefulness --- has led every single industry to reveal they're willing to no longer have standards in the name of doing so

you can tell programming is a real trade because it's even having a moment where the bosses and their dumbest workers don't seem to understand you're not meant to be speeding up the slow parts you're *supposed to be* taking it slow with things like new features. this is what happens when you aren't

"We're shipping so much more code with AI!" "Good, stable code?" "Oh heavens no. No no no. not at all. We couldn't possibly"

entering new levels of "what will a guy do to avoid learning anything about css" previously thought unforeseeable

burning twelve dollars of shale gas (probably more given how heavily subsidised these things are right now) in a residential memphis neighbourhood for two lines of CSS is not something i would personally describe as “fascinating”

love how federal government websites are now malware

this has been a known thing for large language models for a while, how even with all the data in the world, they all eventually devolve into the same patterned blobs as another example: image generators all converging on the same 12 images, irrespective of what you first put into them

"We already live in a world where you walk outside and there are billboards and signs that are obvious AI slop. That’s become part of our visual reality. To me, generative AI feels less like innovation than a symptom of a broader cultural and economic rot." I like this kid.

"stop using markdown", they said. "it's inefficient. nobody wants to open your .md file just use a better word processor. why are you writing in a text editor anyways" well for one it doesn't have a "feature" that lets all my peers play the role of a boss that thinks me to be insufficiently human

a big reason people want you to think large language models are human-like is so they can project a abdication of legal responsibility onto it, a way to make it so when technology screws up, it's "someone else's" fault every legal reminder that this obviously isn't true is a good thing

it's very funny that - there *is* a real thing with models being able to fuzzy search for bugs, and this *has* far-reaching security consequences - nobody at anthropic can actually capitalize on that because none of them are used to doing anything other than pretending the computer is people

this is also why they struggle to understand most people don't really buy into the a.i. stuff they just assume everyone is lying as often as they are, that everyone knows things as little as they do. they cannot comprehend that some jobs require guys who objectively know things to be true or untrue

there really is a case for punishing social media companies who are known for psychologically ruining kids' lives except because doing that might require governing, every government has simply decided to reward them by giving them all the perfect excuse to lead a massive surveillance state

In Mississippi, lawyers on both sides of a court case were caught using AI, essentially two LLMs arguing against each other. The judge dismissed all lawyers from the case and canceled the trial: www.404media.co/judge-learns...

people kept trying this defense with rsync, a tool that others only noticed being vibe coded precisely *because* it suddenly started breaking en masse, for no reason, after years of stability but to be pro-a.i. requires some degree of delusion on what good software even is, so it's not shocking

“if you don’t like gen ai code then i have bad news about all the software you like” no you fool, that is not news. we have all noticed that the software sucks now. and we don’t like it