Grok 4.5 vs gpt-5.6, Apple Sues OpenAI, and China Catches up to Elon | #270 — Key Takeaways

China Is Quietly Winning the AI Race

Evaluate Chinese open-weight models (DeepSeek, Qwen) for cost-sensitive inference workloads — they run 36x cheaper than GPT-4.5 and now power the majority of global developer traffic, but route API calls through China-based servers subject to state intelligence law.

Why China's "Thousand Model War" is a Risk to US National Security | BPH Ep 42

Anthropic's discovery of "JSpace" — a subconscious-like processing layer in Claude — may enable real-time detection of deceptive AI behavior, representing a genuine interpretability breakthrough with direct AI safety implications.

The SpaceX IPO Is Wild

SpaceX's IPO at 90x price-to-sales (vs. Google/Meta at ~10x) is priced for perfection, and Morningstar analysts say it's worth less than half its IPO valuation based on current performance — watch for insider selling as nearly $1B in shares unlock within months.

Mira Murati's 975B Open Model, Ramin Hasani on Post-Transformer AI, and Demis' AI FINRA | EP #271

Liquid AI's Ramine Hassani made the strongest case that true recursive self-improvement requires weight-level neural adaptation — not prompt/code engineering loops — and that current "breakthroughs" like Weco AI's system would take 350 years to fine-tune a 2-billion-parameter model under chinchilla scaling laws.