Key Takeaways
- AI has reached the 90th percentile of mathematicians, with open-source models surging and unresolved personality control issues.
- U.S. manufacturing surged in July (Philly Fed highest since 2021), but housing weakened sharply, and labor market remains tight.
- Tech stocks have seen major corrections (Netflix -45%, Oracle -60%) even as S&P 500 nears all-time highs, indicating extreme market dispersion.
- Raoul Pal predicts crypto market cap reaching $100 trillion by 2032–34, with AI, robotics, and blockchain converging before 2030.
- Declining trust in official economic statistics could reduce GDP by ~$20 billion, far exceeding the BLS budget.
1. AI Advances and Open Source Momentum
- Marc Andreessen highlights that open-source models are enjoying a golden moment due to multiple factors: ability to catch up with state-of-the-art, many teams building open weights, organizational desire for data control, freedom of choice, token cost concerns, and geopolitical considerations. He concludes this leads to more options. — via 1
- AI has already reached the 90th percentile of practicing mathematicians as of April/May 2026, with the next benchmark showing superhuman performance expected by mid-2027. However, large labs' reluctance to use formal methods in reinforcement learning loops is hindering further breakthroughs. — via 1
- GPT models claimed solutions on 19 Erdős problems, with GPT-5.5 Pro solving the most and GPT-5.6 Sol Ultra adding three new claims. These are solution claims, not verified proofs. — via 1
- No AI company can reliably control model personality; OpenAI, Anthropic, and others have all been surprised by post-training personas. The personality problem remains unsolved. — via 1
- Jason praises Inkling's open-source weights, arguing that open-source enthusiasm combined with falling intelligence costs creates an "AI sovereignty stack" that OpenAI should have pursued. — via 1
2. U.S. Economic Divergence: Manufacturing Boom, Housing Slump, and Market Dispersion
- The July Philadelphia Fed manufacturing index surged to +41.4, the highest since November 2021, with new orders and shipments spiking sharply. — via 1
- Housing market continues to weaken: July NAHB index fell to 34, June existing home sales dropped 5.4% month-over-month (far below expectations), and mortgage effective rates climbed to the highest since late 2011. — via 1 2 3
- Initial jobless claims came in lower than expected at 208k, and continuing claims also below forecasts, indicating a still-tight labor market. — via 1
- June retail sales rose 0.2% month-over-month (in line), with the control group's 3-month annualized rate still at a strong 9.2%. — via 1 2
- PPI for goods fell sharply month-over-month, while services edged up; the portfolio management component was unexpectedly weak, lowering expectations for the PCE price index. — via 1 2
- The U.S. stock market has rallied 11% year-to-date alongside 2-3% GDP growth, 3.5% inflation, and a 4.5% 10-year yield, leading Ben Carlson to question if this is the most "normal" year. — via 1
- Despite the S&P 500 near all-time highs, major tech stocks have experienced severe corrections: Micron -25%, Microsoft -27%, SanDisk -31%, SpaceX -33%, Netflix -45%, Oracle -60%, illustrating extreme divergence. — via 1
- Aswath Damodaran released his annual country risk update, covering all his knowledge on country risk used in corporate finance and valuation. — via 1
- Clifford Asness highlights that declining trust in U.S. official economic statistics could reduce GDP by about $20 billion, while the BLS FY2025 budget is only ~$704 million, a striking contrast. — via 1
3. Crypto and the Economic Singularity
- Raoul Pal predicts the total crypto market cap will reach $100 trillion by 2032 or 2034, creating nearly $97 trillion in new wealth, calling it the way to change the future. — via 1
- He argues that the "economic singularity" is not a metaphor: AI, robotics, and blockchain will converge before 2030, fundamentally reshaping work, growth, and investment. He advises using tools and sticking to compounding. — via 1
- Lowering energy costs to near zero through AI and robotics is critical to addressing depopulation and debt problems, as AI and robots will replace much of the labor force. — via 1
