Tap a claim on the ring to put it at the centre.
← Falling back to older 7-nanometre chips would not rescue AI compute,…
8 connected korrents · 6 moments on record from 3 Feb 2025 to 30 Jul 2026.
Drag to move · pinch or scroll to zoom
same subject or similar wordingshaded: claims sharing a subject
At the centre
Falling back to older 7-nanometre chips would not rescue AI compute, because the real gap between chip generations is twentyfold, not the threefold the flops suggest.
Holds DPDylan Patel
Read this korrent →
-
Similar wording
Inference, not training, is now the constraint, and specialized low-energy hardware will beat general-purpose GPUs and TPUs on latency.
Holds JDJeff Dean
-
Similar wording
Semiconductor infrastructure has not kept up with AI's growth, and closing that gap needs capacity built in unconventional ways.
Holds LTLip-Bu Tan
-
Similar wording
The faster AI progresses, the better NVIDIA does, so a shock like DeepSeek should expand its market rather than shrink it.
Holds NLNathan Lambert
-
Similar wording
The world can make roughly 200 gigawatts of AI chips a year by 2030, which makes a gigawatt a week a plausible quarter of the market rather than a fantasy.
Holds DPDylan Patel
-
Similar wording
There will be no overnight intelligence explosion, because a model's best work takes so much test-time compute that time itself is the bottleneck.
Holds NBNoam Brown
-
Similar wording
Nobody wants a slow model, which is why labs will not trade inference speed for cheaper memory even though they easily could.
Holds DPDylan Patel
-
Similar wording
Building capacity at massive scale on a single hardware generation is the trap; pacing matters more than gigawatts.
Holds SNSatya Nadella
-
Similar wording
The memory crunch will make ordinary phones and computers worse year on year, and that is why the public will come to hate AI more.
Holds DPDylan Patel