Post #3942304
2026-06-13 18:51 UTC
That's why https://en.wikipedia.org/wiki/5_nm_process says:
"The 5 nm node is expected to have a gate length of 18 nm, a contacted gate pitch of 51 nm, and a tightest metal pitch of 30 nm. In real world commercial practice, '5 nm' is used primarily as a marketing term by individual microchip manufacturers..."
Moore's Law has been dead for 15 years. Number go up because number MUST go up, capitalism demands it as a sacrament, so they changed what number is measuring.
Replies (1)
-
@landley@mstdn.jp 2026-06-21 19:38
An LLM is basically lossy compression for the google cache. "Training" is the data compression. "Tokens" are decompressing a zoomed in fractal slice of morphed together training data, with lots of brownian motion jiggling to make it work and a ton of autocorrect applied to the jpeg artifacts. "Agents" feed the output back into the input a bunch of times (adding zillions of prompts to the decompressor's starting dictionary) to take lots of slices until the metrics add up to something sellable.