Jan Dutkiewicz

this is literally what the article says, yes.

Put differently, if these assistants are aggregators of sources and they cannot, as software, accurately represent their sources 100% of the time, whatever else they do with that information, that's a massive cause for concern. Our standard cannot be "this LLM is getting better at *not lying*".

I've seen this & am not convinced. Consider basic outputs like accuracy of quotes, where you have a 20% error rate. It doesn't matter if users ask specifically for a given new source or "news" in general, that the machine can't accurately reproduce a quote 100% of the time is already an indictment.

Largest study of its kind shows AI assistants misrepresent news content 45% of the time – regardless of language or territory An intensive international study was coordinated by the European Broadcasting Union (EBU) and led by the BBC

If any other technology had this sort of error rate we would scrap it.

www.bbc.co.uk/mediacentre/...

Herbal supplements were supposed to make them healthier. Instead, they got sick Common compounds like turmeric and ashwagandha can overwhelm the liver and lead to health complications

People want hacks for health, but it really comes down to this: "proper nutrition is best achieved through diet." Rather than worrying about superfoods, supplements, and other nonsense, food politics should be about making nutritious food widely accessible.
www.theguardian.com/wellness/202...

YouTube video by MOLESTAtv Molesta - Martwie Sie ft. Ero, Pablopavo

I always forget this certified Warszawa banger has animal rights bars. I was looking for an epigraph for my book.

"W kolejce w mięsnym stoimy spokojni
Wiesz jak traktują zwierzęta w ubojni
My obojętni wobec tych zbrodni
Jesteśmy wszyscy jeszcze bardziej głodni"

www.youtube.com/watch?v=GhR8...

you understand safety factor, yeah?

The former are their own reference standards (10X safety factor); the latter is an aspirational target. You're quoting from a fact-checked article. Not sure what you think the gotcha here is, chief.

It may seem that in a fraught political moment, adjudicating the safety of protein shakes seems frivolous; I think the opposite. A lot of people get ideas about politics ("the food system is broken and is trying to poison us") from sources like this one. We need clarity on the science and policy.

The problem of course is that scaremongering travels far and fast; fact-checked correctives take time and move slowly.

Everything you actually need to understand about the Consumer Reports protein powder report (threading a gift link!) -- plant-based protein powders are fine, and do not, ffs, allow them to push you to get whey protein

thanks for reading.

No, your protein powder isn’t poisoning you New testing finds two-thirds of popular protein powders exceed lead limits — especially plant-based brands. What you need to know before your next scoop — and why the reality is different.

You may have read that your protein supplements are giving you lead poisoning. That's not the case. If you want to have protein shakes, that's fine. But whether you need to and whether they're safely regulated is a different story. My latest for @vox.com.

www.vox.com/future-perfe...

From Industry to E.P.A.: Lobbyist Now Oversees Pesticide Rules

The former top soybean lobbyist who fought to keep the tree-wrecking pesticide dicamba on the market now runs EPA pesticide policy

and the agency just proposed re-approving dicamba with even fewer restrictions than before.

Gotta have dicamba to grow the soybeans we can’t sell.

Gift Article:

writing about this now. piece dropping soon.

i don't want to diminish the risk of lead exposure, but Consumer Reports has found the ultimate scaremongering clickbait grift: use impossible lead safety standards and apply them to any food that is dried, powdered, or concentrated to convince the public their perfectly safe food is poisoning them.

lol.

yeah obviously.

this doesn't cite the claim, but i'm assuming he gets the idea from the hayek paper i linked.

that's the paper that i thought of. if you're thinking of another one link it.

More irksome use of stats. CR claims that average Americans are exposed to "up to 5.3 micrograms of lead each day through their diet." They don't link the source, but it's clearly Gavalek et al. (2019), which says mean exposure is 1.7 - 5.3. So that's a big range to not mention explicitly.

matt hayek: iopscience.iop.org/article/10.1...

but that applies only to cattle, and comes with the proviso that matching current consumption levels is impossible given land constrains, never mind cost constraints. in short: current levels of meat could never be achieved without industrial ag.

The Case for a Global Ban on Industrial Animal Agriculture by 2050

The "keep it in the ground" of the food system. Hard to argue against this. Equally hard to imagine it happening.

verfassungsblog.de/global-ban-o...

oh you're one of these.

read the article, which, incidentally, i didn't write.

oh yeah i'd never heard of that place until last summer - so good.

Biggest personal vegan tragedy since Foodswings closed in Brooklyn.

L'Artisane Bakery & Bistro closes shop in the Gables Unfortunate news coming from downtown Coral Gables where L’Artisane Bakery & Bistro has closed its doors for good. The shop – which billed itself as America’s first vegan French bakery – shut down thi...

The place was, in my opinion, the best vegan restaurant in the country.

There goes the only good reason to visit Miami.

theburnmiami.com/2025/09/lart...

what does the phyto mean?

uh, it means bad, joe.

whoa. bad estrogen. that's crazy.

This piece is a good companion to mine from The New Republic last week. The problem isn't the soy; it's that we feed it to factory-farmed animals and burn it for biofuels. Soy itself is a protein- and nutrient-rich food we should eat more of.
bsky.app/profile/jand...