Anyone who spends $200 million to develop a completely new high-end processor microarchitecture every 3 to 5 years is not going to complain about an ARM license too much. You get freedom with $$.
2,776 karma · joined January 16, 2026
Anyone who spends $200 million to develop a completely new high-end processor microarchitecture every 3 to 5 years is not going to complain about an ARM license too much. You get freedom with $$.
If Nvidia sells hardware for $100B with 75% cross margin, and provides $50 billion in backstop for that same hardware, it would be still be nicely profitable deal ($25B) if the backstop capacity would be a total write-off recovering $0. Reselling that capacity in some large discount below already low backstop price would increase the profits.
It's all those pension funds, sovereign wealth funds and Softbank getting into that $500 billion deal that will be hurt.
Find a bicycle without mudguards and ride very slowly, or just ride on clean roads on a sunny days.
The RISC/MIPS concepts date back over 40 years. The base instruction set is intentionally designed with unencumbered, expired, or public-domain architectural concepts.
RICV-V microarchitectures and implementations are at much high risk of violating patents. Especially anything that is even slightly high performance. SiFive, Andes , and Alibaba’s T-Head are filing thousands of patents on microarchitectural optimizations and extensions. China's RISC-V patent-sharing alliances and other industry groups are building defensive patent cross-license around their RISC-V-related patents.
I see losing Archimedean property as more of a win than loss. It's a one simple thing that turns calculus operations into single algebraic evaluation, where you can avoid dynamic limits. Things become simpler and easier to prove.
Author suggests using them. Resistance to Hyperreals seems to come from era before there were rigorous definitions for them.
Typical use cases are for doing biometric authentication without giving your biometric information, or sensitive queries using medical information. Apple has homomorphic encryption in image search. You can use your own photos encrypted into the cloud to search for landmarks in the image without revealing photos.
People can also coordinate and compare information without sharing sensitive data.
I think the most immediately useful is checking if a username/password hash appears in a leaked database without revealing which account or password hash is being checked.
It narrows the 10^3 - 10^6 penalty to 10x - 100x.
The basic idea of of the project is to remove the need for trust.
As a business Google might want to focus on fundamental research 2-3 years from now and not compete on who acquires more money losing customers. Just stay little behind and invest money better.
Coin flip game:
You invest $100. Heads you win 50%, tails you lose 40%. At every round you must bet all you have acquired so far. We play this 200 rounds. Would you take the game?
(btw. it has positive expected value)
rot13: Lbh zbfg yvxryl ybfr. Rkcrpgrq inyhr vf zrna erghea naq cbfvgvir, ohg zrqvna erghea vf artngvir. Zrna erghea vf vasyngrq ol ener bhgpbzrf.
Gur ibyngvyvgl gnk vf gur qvssrerapr orgjrra gur nevguzrgvp naq trbzrgevp nirentr (be “rafrzoyr nirentr” naq “gvzr nirentr”) ergheaf. Anffvz Avpubynf Gnyro naq Znex Fcvgmantry ohvyg gurve ragver pnerref nebhaq gur rssrpg bs gur qvssrerapr orgjrra rafrzoyr nirentr naq gvzr nirentr.
ps. Being influential in Silicon Valley just means you are influential, it does not mean substantial. Leopold is still influential and gets money thrown at him at $100s of million despite having no substance.
Nvidia chips last just 3-5 years in data center use.
ps. The author seems to be Simon's problem Simon https://en.wikipedia.org/wiki/Simon%27s_problem
At least half of that $1.5T is an AI business they don't yet have and have failed to get into. They are renting data centers to those who can. Expect the price to drop 10–20 percent this month as they keep releasing stock to the market.
The leverage was 4x or 5x of that and highly concentrated. Citadel bought $16 billion public portfolio with 10% discount.
If a Tesla-SpaceX merger is in the plans, China can prevent it unless Tesla gives up the Chinese market, half its size, and becomes a regional car manufacturer.
More than 90% of "in mice" studies do not transfer to humans.
China is already in their third and fourth generation superallloys with DD6 and DD9 (nickel-based single-crystal superalloys). They are deployed in three different engines.
It's the same with DUV. Just 5 years ago they were 20 years behind. Now they are maybe 5 years behind. Maybe next 2-3 years they are still not getting the yield par with ASML but just watch.