Garry Tan retweets a post arguing that the US faces a worsening strategic scenario, agreeing with claims that China's focus on open source is not altruistic, and urging the US to build power infrastructure and datacenters.
Background
- Paul Graham is a well-known tech investor, co-founder of Y Combinator (the startup accelerator). Garry Tan is a former YC partner and current CEO of venture firm Initialized Capital / publisher of the tech blog "Tan's Newsletter". Their short tweet is a retweet of a user named "Chubby♨️".
- "Build power and datacenters" refers to the massive electricity and computing infrastructure needed to run AI models. Datacenters already consume huge amounts of power, and AI training (especially large language models) dramatically increases demand.
- The retweeted user is arguing the US is falling behind China in building this infrastructure, citing China's massive state-backed investment in AI compute and energy capacity, while the US faces regulatory hurdles (permitting, grid interconnection) that slow construction of new power plants and datacenters.
- This debate is central to US-China tech competition: who can build the physical infrastructure to support the next generation of AI.
The US government is increasing its involvement in controlling AI models, while open-source Chinese models are rapidly improving. According to @jvisserlabs, this trend should push more people toward open-source AI in the coming months.
Anthony Pompliano sat down with J. Visser for a wide-ranging discussion covering weakness among hyperscalers, the rise of open source AI models, the implications of the Mythos/Fable 5 situation for governments and investors, Kevin Warsh's first Fed press conference, the inflation outlook, and why bitcoin remains in a bear market.
Anthony Pompliano and @jvisserlabs discussed the recent weakness from hyperscalers, the rise of open source AI models, the Mythos/Fable 5 situation for governments and investors, Kevin Warsh's first Fed press conference, inflation outlook, why bitcoin remains in a bear market, and related topics in a podcast episode.