AI-written page, and it gets worse as it goes on:

> The modes are the load-bearing piece:

lol

Aren’t all the big chrome vulnerabilities type confusion?

They’re ZDR and the OAI ones aren’t

Does anyone know if it is not ZDR on the codex app usage too. I am assuming it isn't

WTHIT?

ZDR = Zero Data Retention — they don't store your inputs/outputs.

More like the business tax

> Astra improved a term in a bound on these gaps that had remained unchanged for more than 80 years. We’re sharing the proofs and abridged chain of thought and verification materials for both results.

Looks like they listened to Terry Tao’s request for CoT in his talk on LLM use in mathematics?

Kind of sad how all these technical blogs just reek of Claude text these days. Hard read when it’s obviously padded by an LLM…

Since it's only discounted on the standard "OpenAI," non-ZDR route (old pricing on Azure), I'm guessing a lot of users won't see this benefit? Since a lot of users enable a global "ZDR-only" toggle on OR

It would seem that getting lots of data is exactly the reason to discount this.

OpenAI says they don't use any API data for training.

(there's probably going to be a reply about 'but how can you trust them'; I'm just stating what they say)

Does OpenRouter say the same?

Yea the "When the fact lives outside the model, a wrong answer has an address" sentence seems aggressively AI written. Saw that and my senses went off.

Senses of what? LOL. The whole Internet is AI generated by now and we all contribute to that on daily basis. get used to it or dull your senses ...

Ah Ah Ah – how we did't like this. Downvoting me will surely help! Can't you see how much slope is already around? Don't we – you and me – contribute to that, especially at work? Isn't "dull your senses" standard answers of most expensive shrinks? So what did you disagree with? Or you simply didn't like the truth? Ok, I got it. No problem.

> Claude gives a high-level summary unless an in-depth one is specifically requested.

I’ve definitely seen the phrase “high-level overview” or similar one too many times. Perhaps that’s from the prompt.

Neat. Curious to see if RL pans out. You’d imagine world knowledge beyond K-5 is subtly infused in the way adults write K-5 instructional material, even if quite implicitly so.

Yeah, the filtering process probably wasn't particularly robust. The Schrodinger's cat example ("It's a cat that has been misbehavin'!") sounds like... a joke?

Looking at the paper, it looks like they started with FineWeb-Edu, then filtered it based on an "age of word acquisition" dataset, with word frequency used as a proxy for values not in the dataset. They "only discard samples in which more than 5% of the words exceed the target age of 12." Maybe 5% was too high? They also filtered out beyond K-5 math symbols, like sigma. Then they trained a classifier to do more filtering.

And they tested it on two grade-level benchmarks, and it only got 0-3% correct on the beyond k-5 boundary, while also decreasing in performance on the k-5 boundary (which they say is an acceptable tradeoff, since they were trying to get a sharp cutoff). So presumably since they got good results from the benchmark they stopped.

> Motor Power 250W (Rated) / 500W (Peak)

A 4090's TDP!