Thread Guy

GoodAlexander × David Choi

Chinese open-source models reportedly approach Opus or Sonnet performance on roughly $10,000 GPUs, while Qwen 3.6-class compute costs about $2 per hour.That 40-80x gap supports local inference and crypto-financed compute, but USD.AI’s $1 billion GPU-loan target, reference rate, and CHIP model remain unproven.

Thread Guy × Tulip King

Tulip King frames prediction markets as a general-purpose financial primitive, with the edge shifting from isolated forecasts to cross-market structures such as Lighter’s 91¢ pair that pays whenever either leg resolves true.The broader thesis links stablecoin growth to sovereign-scale Bitcoin demand, while Hyperliquid faces structural margin compression and quantum computing remains a tail risk to monitor.

Thread Guy × Murad

Murad rejects the four-year-cycle bear case, arguing Bitcoin follows steady money-supply growth and should bottom within roughly three weeks, with a monthly close below $69,000 as his falsifier.He sees Pump.fun’s fragmented speculation favoring 50,000-100,000-plus committed holders, concentrates the thesis in SPX6900, and forecasts $100 this cycle and $1,000 next cycle, while community durability and retail’s return remain unresolved.

Good Alexander × Thread Guy

Good Alexander says crypto-AI’s original thesis failed, while AI-linked proof-of-work and multiplayer coordination remain differentiated opportunities.Crypto stays in “PVP hell” until bridges such as possible MicroStrategy S&P 500 inclusion, Robinhood, and regulated tokenized securities unlock outside capital, while stablecoin growth and AI circularity remain key risks.