README is in my opinion (author here) the most interesting - I wrote it to help others build useful mental model to be able to recreate the project yourself, without need to even read my code
I am not super familiar with C and CUDA, so I read solely for the README and enjoyed it supremely. The blend of cheerful walking through instructive examples and your philosophical takes on how to approach the exercise to get the most out of it put me in a great mood. You captured that special upbeat attitude that comes about when you're doing something as well as you can just because it's so legitimately interesting to you.
Which, to be fair, laid the foundation for the well-educated part.
The Soviets really valued STEM. They also quite valued emancipating women.
Just for context, in the 60s, around 5% of chemistry PhDs in the US were women. In the Soviet Union, it was 40%! [0]
Of course, that doesn't excuse all the other things they did, but the amount of badass female engineers from Eastern Europe I had the honor of working with is a direct result of the pipeline the Soviets built.
They don't if you mean STEM and emancipation, quite the opposite, actually (compared to West Germany).
In addition to the points of sibling comments, their respective starting posititions were drastically different: West Germany got the marshal plan, which benefitted their economy, the East had to pay reparations to the USSR, which meant whole factories, trains, even railroad tracks, all in all amounting to about a third of industrial capacity, were transferred to the USSR.
Without having firm data, I can see a few factors that are different. After the collapse of the GDR, it was easier for eastern Germans to move to west Germany than for Polish to move to a different country in the west. Mostly younger and educated people would have made that move, hampering future generations. With the Reunification also came the whole Treuhand issue which essentially sold off a good chunk of eastern Germany for pennies to western investors, because eastern investors had no capital. That meant the east lost out on the profits from its economy as they would accumulate in the west instead. Even today a large part of east German rentals are owned by western landlords or corporations. Then the industrial base of west Germany was setup far more for competing on the open world market with automotive companies in the NW (VW), SW (Daimler) and SE (BMW) plus the big industrial area Ruhrgebiet. So you naturally got an economic focus even after Reunification on the old BRD with the previous GDR requiring decades to hopefully catch up to the rest of the new country.
Quite a few educated East Germans have become West Germans as soon as they had the opportunity (or moved elsewhere in the world), but East Germany actually has a couple of high-tech 'hotspots' and good universities.
An East German state (Saxony) also consistently has the best education system among German states.
One factor in this may also have been the way the privatization of East Germany was handled. Its often overlooked, but the vehicle for it was called Treuhand[1]. Regardless of whether it was necessary or not or right or wrong, it did basically shift out a large amount of capital assets into West Germany (and still carries this sentiment of "opportunistic theft" today).
The headline figure of the article is purchase power (PPP) adjusted. I couldn't find any numbers for east German states where the purchase power adjustment happens per state.
Since housing is the largest component and housing costs differ between east and west Germany using a nation wide PPP adjustment factor gives wrong results for individual states.
Yes. In most large companies the corporate administration does not have a career in the actual subject the company operates in, but more in finances and economy. This forum is also based in the USA, which maybe has another culture. But it is also routinely pointed out here, how corporations act more in the interests of shareholders, than in the improvement of the actual product and innovation.
The most popular way to own a company is also inheritance, instead of studying an engineering subject.
This comment was also an answer in the context of the peaceful revolution, where a lot of companies where bought by larger companies from the west, both to destroy their better competitor and get funds from the EU. Such actions are seldomly done by engineers.
A less cynical explanation is that it helps decouple product failures from support failures. Last thing you want is for your customer support to break whenever your product breaks.
Your explanation is literally "blame somebody else, shift responsibility". It's a pretty straightforward case of assuming bad intentions.
Obviously, I don't know for a fact what Anthropic's motivations are, but I don't believe I'm being overly optimistic because I know for a fact that "use a third party as a way to de-risk" is a tried and true strategy. E.g. when I was at Facebook, all regular day-to-day comms were on Workplace, but they kept an IRC server hosted by some vendor or other, specifically to coordinate responses for serious SEVs that disrupted the normal channels.
Of course, an even simpler explanation is that they perceive building their own support harness as just low value work for engineers who could be working on their core product instead, and the cost of buying that service from somebody else is probably a drop in the ocean.
Adding a support for new hardware to PyTorch is actually quite convenient. I did that with WebGPU using the same PrivateUse1 mechanism TorchTPU used. Every hardware has its own slot and identifier, and when you want to add a support for a new one without merging it into PyTorch, PrivateUse1 works essentially like plug-in slot
reply