Sakana AI launches Frontier Intelligence Group to bet on post-Transformer paradigms
SungjinAhn_ · x · 2026-10-08
Sakana AI announced its Frontier Intelligence Group (FIG), led by CTO Llion Jones, a co-inventor of the Transformer, explicitly questioning whether scaling the Transformer with more data and compute is all you need for AGI.
Key points:
- FIG argues current systems still have fundamental flaws: confident hallucination, collapse on genuinely novel situations, and high energy costs — and that researchers may need new paradigms rather than incremental tweaks.
- The inspiration is biological intelligence: continuous learning, robust generalization from far less data, and self-directed, open-ended curiosity — noting today's artificial neural networks remain far from our understanding of the brain.
- The poster, a university lab researcher in Korea, asks why Korea lacks a frontier lab betting on original ideas, arguing the memory-boom revenue makes this the right moment to fund frontier AI research there.
More from AGI Musings
- Journal editor recruits engineer peer reviewers: no PhD needed, but no AI either — lemire · 2026-10-09
- Proposed 'comprehension audits': independent auditors grill teams on AI training recipes — ronbodkin · 2026-10-09
- The verification tax: why stacking AI tools on overloaded attention stacks adds cognitive load — DrKavner · 2026-10-09
- Stephen Casper weighs in on Dario's essay: embedded evaluators, China, regulatory capture — StephenLCasper · 2026-10-09
- Anthropic staff: researchers no longer understand what they ship as recursive self-improvement kicks in — ronbodkin · 2026-10-08
- Senior i-bank exec: China is furthest ahead in physical AI and LLM-powered robotics — PAstynome · 2026-10-08