Both raytracing and NPUs use a lot of bandwidth and that is scaling the least with time. Time will tell if just going for more programmable compute would be better
Banning apps that linked sales out of app is what they already did, but that was ruled to be anti-competitive. Then they allowed it, but with an added 27% fee for whatever was purchased through the link (banning if you didn't pay the fee), which was also found to be anti-competitive and a attempt to avoid having to comply with the first ruling.
Charging for bandwidth of app downloads and updates would equate to pennies (relatively). It's not worth it. And would probably be deemed anti-competitive in the future anyway if they didn't also allow 3rd party app stores / installs (like in the EU).
These companies didn't even exist 100 years ago and most of the success they have had, are big innovations in their respective areas in last 20 years. Yet now they are designated as some sort of immortal "gatekeepers".
They're not designated as "immortal" gatekeepers. The commission can revise the list at any time, as market conditions change.
Every monopolist was once not a monopolist. Every monopolist came into existence at some point in the past and did not exist before then. So what? Should we abolish all antitrust law for this reason? It's a very strange criticism.
> Under the DMA, the European Commission can designate digital platforms as ‘gatekeepers' if they provide an important gateway between businesses and consumers in relation to core platform services.
Microsoft and Apple are in their late 40s, Amazon and Google are in their 20s, Facebook is 19 and ByteDance is a teenager with 11 years old, yet managed to be classified as gatekeepers.
I think no other company, with a young age, in any other domain but tech could ever qualify as a gatekeeper
Sometimes it makes sense to pay even failed things because your overall cost is lower.
Say hypothetically that TSMC said $30 per working chip or $8 per chip and $2 to test. Taking the later choice would result in lower overall costs at current yields.
Typically for this kind of work the manufacturer wants their yield improvements to lead to reduced cost to them so only paying for what works is standard to my knowledge.
In contrast it is bespoke work that sometimes shifts, where there is no possibility for things like down binning to recover failures (disabling parts of a chip and selling for less)