AI Scaling Is More Like Nuclear Fusion Than Alchemy
willccbb · x · 2026-07-16
The core idea is that many new model tricks are essentially just "compute multipliers." While valuable, the room for further exploration is shrinking. The author notes that one of the most important data multipliers discovered recently is a 90s RL algorithm, and scaling data is much messier and harder than scaling compute.
In a replied comment, the analogy for ASI was corrected to being "more like nuclear fusion than alchemy": it could fundamentally change the world, but the process will be long, and the laws of physics still apply.
Related event: Researchers cool expectations around 1000x AI R&D speedups(8 posts)→
More from AGI Musings
- The Evolution of LLM Business Models: Selling Outcomes Over Tokens — yacineMTB · 2026-07-22
- Bindu Reddy says GPT-6 is coming soon, with Alibaba, DeepSeek and Kimi close behind — bindureddy · 2026-07-22
- Bindu Reddy says the industry still lacks a way to train 20T models and scale post-training RL — bindureddy · 2026-07-22
- Advanced AI Models Are Becoming Impossible to Plug and Play — emollick · 2026-07-22
- AI suggested a better composition, and that made one user uneasy — Sydde · 2026-07-22
- The Thimble and the Waterfall: AI's Data Bottleneck and Feedback Loops — dyamins · 2026-07-22