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.
I grew up in Southern California - so the concept of continuously free flowing fresh water all year was a thing of hard-to-believe imagination.
All these decades later, now that I live in the mountains around the Monterey Bay, I look with daily amazement at at the ever-running stream that runs past my garden.
I kinda wonder whether climate change will cause my stream to dry up or become larger.
I'm glad my house is roughly 40 to 60 feet above (and well back from) the cr…
Heute 9 Uhr startet das Heart Times Bike Race in die zweite Ausgabe.
#DotWatching : https://www.followmychallenge.com/live/heart-times-2026/
In her introduction to an event on climate communication, initiator Mare de Wit quoted a man in the movie La Haine.
While falling from a tall building, he kept saying "Jusqu'ici, tout va bien."
So far, so good. A metaphor for climate change, for those of us who weren't hit hard by it yet.
Filing: Anthropic hired Ballard Partners, a lobbying firm with strong ties to Trump administration, days after DOD designated the company a supply chain risk (Bloomberg)
https://www.bloomberg.com/news/articles/2026-04-13/anthropi…
Wish I could share the full photo with you but I don’t share identifiable faces from protests. When I asked if I could take a photo, she held up her sign with a determined, slightly sad smile. That feeling could sum up the event. Joyful, yes. But also a recognition that there is hard work left to do. #nokings
Sometimes I worry that having an autoimmune disease will eventually cause me to ignore something serious, because there's ALWAYS something feeling off or hurting. Currently wondering if I am just getting a feedback loop of nasty symptoms that makes it hard to get rid of them, or if there's a more serious root cause. But stress from everything could definitely also be the reason. Sigh. #AutoimmuneDiseases
I finally got to layer 3 on Casualties: Unknown.
The random spikes falling from the ceiling were spooky. I also fell past a turret thing and it gave me a damn heart attack, but it didn't hurt me.
Then I saw a land mine. I stepped on it by accident, then ran as far as I could, and... kaboom. A few dislocated bones, got debris stuck EVERYWHERE, and quite a bunch of bleeding. Surprisingly, I could heal everything (which took a while), except for their poor left eye... but it is…