The air in Kraków is fine once you give it a good chew. I don't know why people are complaining.

LOL, Slavic humor at its best :)

I live in Ostrava, some 160 km away. Entire Upper Silesia is a bad place for air quality in winter, it can often be seen on continental maps as a sore red spot.

Fortunately most of the coal burning is gone, but individual people still burn all sorts of shit in their homes. PET bottles etc.

Is it in a valley?

The official term is a Basin:

https://en.wikipedia.org/wiki/Ostrava_Basin

We tend to suffer from "inversions" here, and way back in the coal times, the air quality used to be comparable to London during the Great Smog of 1953. Nowadays it is better, but still quite bad compared to rest of Europe.

In December 2024, I traveled from Ostrava to Warsaw and back via train, so through both Czech and Polish Upper Silesia. The Silesian part of the journey was a pea-souper, like riding through a yellowish cloud. (Warsaw itself had crisp chilly air.)

I don't understand why even put that HBM on top of the core.

From what I understand, in a typical gpu core you put logic and connectors on one side and innert silicon on the other. So unless you drill through silicon you don't get shorter routing.

Why not put GPU one one side and HBM on the other side of the PCB? This would fix the cooling problem?

Routing signals through PCB vias requires greater voltage and has lower available bandwidth than silicon-to-silicon bridges. AMD's first generation of cache dies bonded to the top of the CPU, but their second generation bonded to the CPU's bottom which improved cooling for the fast logic on top. Similarly, HBM under logic would be ideal.

They lose the competitive advantage. They have nothing more to offer than what Google has in-house.

This means that current tokenisers are bad, and something better is needed if text rendering + image input is a better tokeniser.

Humans recognise two vastly different types of language input (auditory and visual). I doubt that one type of tokeniser is inherently superior.

Compute has been getting cheaper and models more optimised. So if models can do something it will not be long till they can do this cheap.

GPU compute per watt has grown by a factor of 2 in last 5 years

Polish is quite good, expected based on the founders' background

And the Hubble tension is around 4%, within a factor of 2 of that 10%. This sounds reasonable.

Also, with renewable energy sources, the problem is not manufacturing energy per se but getting the energy where and when you want.

For example, you could run a gas turbine in Germany in winter at night during a spike of the consumption and run carbon capture on a desert during the day

In a sense you use atmosphere as battery/transportation system.

I am not saying that this must make sense but this is far more complex than napkin math.

Because in a classical theory, where there are no particles, there is still the same short range potential.

This arises from a parameter in the elementary field equation. If that parameter is non-zero than it is both true that the field is stiff and it must be mediated by a particle with non-zero rest mass. This says nothing about causality.

Correct. The author explains how both limited range and particles ability to have energy even at rest comes from "stiffness" (which is how he decided to call this parameter you talk about, in order to convey physics to laymen).

What about using multiple open source projects with already existing tests? You check if results of execution are the same between compiler versions? This should also provide you with a better coverage of various features

Sure that counts as differential testing. The issue is there are limited number of them compared to the amount of code you can generate mechanically, and especially in the case of C, it is not straightforward to write a standardized script to build and run a bunch of random projects.