Keep on to blur preview images; turn off to show them clearly
Underrated funny issue: due to the horrors of «involution», they can add GWs fast enough to offset Jensen's efficiency gains purely through energy cost cuts. These subsidies won't be crippling even if they go to 70% or 90% to match Rubin. Only actual NPUs are a limiting factor.
We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1
RT @Jack_Raines: Jacobin not accounting for compounding is so on the nose it hurts.
Author. Coder. CTO. θηριομάχης. Building: https://t.co/otXT4Wy6WR. Writing: https://t.co/dBPBtyCIHw.
阿里郎？😂
🔖 Building · Sharing ✨ Zipic · TimeGo · KeygenGo · HiPixel · Orchard · Prism
Here’s @enesakar from @upstash demoing context7 and @ericciarla demoing @firecrawl_dev at today’s @a16z x Upstash x Firecrawl AI night Two of the *most used* MCP servers on Cursor, Zed and Claude Code 🔥🔥
Also killer demos by @joshtriedcoding @goetzejulian 🔥🔥
RT @phuctm97: Introducing Vibingbase: Lovable for desktop apps Vibingbase is my most ambitious project ever and is my focus at the @Hacker…
Built 8 startups in 12 months • Sold 3/8 startups • Building https://t.co/ns7SH9FiWh, https://t.co/1tAr7flVT2, https://t.co/pL2Vv5UTCK, https://t.co/sYr4JzzOtJ, https://t.co/HvvzdvWg5H, & a secret startup 🤫
China issues 50% electricity subsidies for datacenters. Very clever. Their industrial energy costs are already (generally) below the US and won't spike due to such trifle as datacenters, but their chips are far less efficient. With this they get to > Hopper levels of FLOPs/$.
Underrated funny issue: due to the horrors of «involution», they can add GWs fast enough to offset Jensen's efficiency gains purely through energy cost cuts. These subsidies won't be crippling even if they go to 70% or 90% to match Rubin. Only actual NPUs are a limiting factor.