aubanel

You can disagree with Anthropic leadership, but all the points you mentioned can reconcile very well with them thinking in good faith "advanced AI is too dangerous to be left in all hands" Except the "train on everyone else IP" which can be said of all AI companies.

"For god’s sake, it’s Wikipedia page is 19 years old." Rich patronising from someone who didn't bother reading the wikipedia page of Its, the possessive form of the pronoun It.

Please make a benchmark for it, that'd be super interesting! My guess is we'd see models climbing it quickly, but maybe not

Really impressive signal that this 128B model can beat DeepSeek V4 (1.6T) on most coding benchmarks!

Also, I really like Poolside's habit to compare not only to other top models in its weight class (others don't do it, looking at you Mistral), but also to the very top open-weight models, even much bigger ones like the 2.5T Kimi-K3!

Feed algorithms are far from sorting videos only by strict decreasing order of trending factor. One example of that is that feeds are randomized: refresh the page, it changes completely. Another example is how it optimizes for bait content.

Probably a great decision, but why/how can it be decided at a local level by a mayor, instead of a federal level?

This is the absolutely horrific next stage for social media platforms:

- They're already well able to surface the most addictive short video for a specific user out of millions of real videos.

- But these millions of real videos are just darts thrown into the space of "videos that could hook the user", in the end even the best-selected of them is not perfect.

- Now, behold! AI allows to generate the perfect video to surgically hit all the switches in the viewer's brain and turn it into a zombie hooked for days on end.

Let's hope our regulations hit these "social networks" hard enough so that never dare deploy this kind of technology.

Garry Tan's point still stands: he never pretended to be building nice software. But his point was that he can now build AT ALL! Shipping a webpage at all is the firs step ; making it load under 7 Mo is just a refinement, an important one of course (who tf wants bloated webpages) but still only a refinement. Tan is right to be amazed and to be shipping.

There's no question to me, after trying both, that Fable is much better than GLM-5.2 when left alone in front of hard coding tasks Now maybe what plateaus is the human collaboration efficiency, because at some point it will be bottlenecked by the human

Thus companies who still try to have humans perform intertwined work with their AI won't see an improvement, while the ones who fin the right conditions to give their AI more free rein will see it.

Kind of like it's no use having a workhorse pull a combine harvester : at some point, when machines reach sufficient efficiency, you just give wheels to the harvester and let it run.

This is actually a good idea! And still looking forward for aleph alpha to release new models, after the Previa ones a while ago!

> Bigger is not better

The article uses the example of GLM being smaller than DeepSeek, yet better on hallucinations as "smaller can be good too"

But the GLM family itself is scaling up fast: GLM-5.x family is 754B, double the previous generation of GLM-4.x

> comes within just 4 points of GPT-5.5 and 9 points of Fable 5

9 percentage points IS a big difference

Gotta love how Apple cares (or used to care) about making their products a holistically good user experience