Tootfinder

Bra artikel kring vad ultraprocessad brukar betyda och problemet med någon form av regler mot det utan att man ändra den definitionen.
Men ändrar man definitionen så gäller inte längre forskningen som finns direkt på det man reglerar mot.
https://www.politifact.com/ar…

What is ‘ultraprocessed’ food’? Brazil has an answer
Health Secretary Robert F. Kennedy Jr. promised the U.S. would have a definition of “ultra-processed foods” by April. Experts have long disagreed about the term, but Brazil might have an answer.

Donald Trump is teaching the world a lesson
- but not the one he thinks.
The attack on Iran was meant to be a dazzling display of military supremacy.
It has instead illuminated chinks in the US’s armour.
The US president’s formidable arsenal cannot summon up an insurrection from Iran’s tyrannised and leaderless opposition.
It cannot force merchant ships to run a gauntlet of missile and drone attacks in the strait of Hormuz.
The government in Tehran and th…

Source: the Pentagon is discussing plans to set up secure environments for AI companies to train military-specific versions of their models on classified data (James O'Donnell/MIT Technology Review)
https://www.technologyr…

The Pentagon is planning for AI companies to train on classified data, defense official says
The generative AI models used in classified environments can answer questions but don't currently learn from the data they see. That could soon change.

I woke up angry today.
ICE spokesperson Lindsay Williams is a liar, who is bad at lying.
“Asked if the ICE officers chased Lopez, Williams said: ‘Chased? I wouldn’t say that. They followed him until he crashed.’”
Wouldn’t say that? How about: ICE was committing crime (their standard procedure), someone died, they’re the killers. Your people killed a school teacher, Williams.
Have the day you deserve, you apologist scum.

Driver fleeing ICE traffic stop crashes, kills teacher near Georgia school
Authorities in Georgia say a Guatemalan driver fleeing a Georgia traffic stop by federal immigration officers crashed into another vehicle, killing a teacher who was headed to work. Oscar Vasquez Lopez, the 38-year-old driver accused of causing the Monday crash just outside Savannah, remained jailed Tuesday on vehicular homicide and other charges. Court records don't list an attorney for him. An Immigration and Customs Enforcement spokesperson says officers were looking to enforce a 2024 deport…

We just found out our AI has been making up analytics data for 3 months and I’m gonna throw up.
Support
So we’ve been using an AI agent since November to answer leadership questions about metrics. It seemed amazing at first fast answers, detailed explanations, everyone loved it.
I just found out it’s been hallucinating numbers this entire time.
Our VP of sales made territory decisions based on data that didn’t exist. Our CFO showed the board …

"Across a variety of tasks, including mathematical reasoning and reading comprehension, we find that although AI assistance improves performance in the short-term, people perform significantly worse without AI and are more likely to give up."
https://arxiv.org/abs/2604.04721

AI Assistance Reduces Persistence and Hurts Independent Performance
People often optimize for long-term goals in collaboration: A mentor or companion doesn't just answer questions, but also scaffolds learning, tracks progress, and prioritizes the other person's growth over immediate results. In contrast, current AI systems are fundamentally short-sighted collaborators - optimized for providing instant and complete responses, without ever saying no (unless for safety reasons). What are the consequences of this dynamic? Here, through a series of randomized contro…

So to follow up on this, I've caught it in action. Models, when quantized a bit, just do a bit more poorly with short contexts. Even going from f32 (as trained) to bf16 (as usually run) to q8 tends to do okay for "normal" context windows. And q4 you start feeling like "this model is a little stupid and gets stuck sometimes” (it is! It's just that it's still mostly careening about in the space of "plausible" most of the time. Not good guesswork, but still in the zone). With long contexts, the probability of parameters collapsing to zero are higher, so the more context the more likelihood you are to see brokenness.
And then at Q2 (2 bits per parameter) or Q1, the model falls apart completely. Parameters collapse to zero easily. You start seeing "all work and no play makes jack a dull boy” sorts of behavior, with intense and unscrutinized repetition, followed by a hard stop when it just stops working.
And quantization is a parameter that a model vendor can turn relatively easily. (they have to regenerate the model from the base with more quantization, but it's a data transformation on the order of running a terabyte through a straightforward and fast process, not like training).
If you have 1000 customers and enough equipment to handle the requests of 700, going from bf16 to q8 is a no-brainer. Suddenly you can handle the load and have a little spare capacity. They get worse results, probably pay the same per token (or they're on a subscription that hides the cost anyway so you are even freer to make trade-offs. There's a reason that subscription products are kinda poorly described.)
It's also possible for them to vary this across a day: use models during quieter periods? Maybe you get an instance running a bf16 quantization. If you use it during a high use period? You get a Q4 model.
Or intelligent routing is possible. No idea if anyone is doing this, but if they monitor what you send a bit, and you generally shoot for an expensive model for simple requests? They could totally substitute a highly quantized version of the model to answer the question.
There are •so many tricks• that can be pulled here. Some of them very reasonable to make, some of them treading into outright misleading or fraudulent, and it's weirdly hard to draw the line between them.

Via: L'avventura: Pet Detective @amcnal.bsky.social
“over on x the everything app some Chicago high school kid named Danny Spud tried to disrupt his school's anti-ICE protest by yelling about how much he loves ICE and then filmed himself getting walloped by a fellow student lmao
the quote tweet making fun of him has 3x the likes”
I’ll post the video next.
i’m not totally 💯 it’s not staged, but either way, remeber;
violence isn’t THE answer
But it su…