Oleg Ciubotaru - Latest News and Information
Tech commentator who engages with the scale of resource extraction (e.g., copper demand vs recycling limits) as it relates to electrification, data centers, and AI.
We've had 1M token context windows for 2 years [with many caveats]. By now you'd expect 10M+ of coherent, high quality context window to be available everywhere. Turns out unlocking that is a much harder [expensive] problem than anyone wanted to admit.
Dylan Patel, SemiAnalysis CEO: Late last year we were spending $50k on Claude Code. By Super Bowl weekend, our run rate hit ~$6M/year... Our data center team lead, not a developer, now spends $5k/day on Claude Code building his own tools.
