Read the piece on rumelt.substack.com ↗
Every claim below was made in this piece, quoted word for word and numbered in the order the piece makes them, so you can read it there rather than take our word for it. The sentence above each quote is our reading of the claim, not their wording. Each quote was checked against a stored copy of the page at build time; where the two differ, the quote is the fact.
Their wordsSo what does Nvidia have that is so special? It has CUDA, the language that most AI runs on. CUDA has two decades of software libraries and developer tools. Graduate schools teach it. It has kernel-level optimizations so that Nvidia’s GPU is the thing you use to test out your new idea at 6 p.m. The Nvidia chip is the castle inside the CUDA moat.
Their wordsGoogle’s open-source moves aim to drain that moat. If Nvidia’s strength comes from having a familiar software stack, then releasing an open-source stack that will work on TPUs lowers the switching costs for developers. Yes, the battle over hardware is important. But the real chokepoint being fought over is the developer stacks.
Keep reading
The other pieces are the ones whose claims are worded most like this one's. That measures language, not agreement: they may argue the opposite.