New tutorial traces AI's evolution from LLMs to "thinking systems" via Test-Time Scaling, Loop, Harness, RSI
rsasaki0109 · x · 2026-10-28
Masafumi Oyamada released an expanded version of his IBIS 2026 tutorial, "Modern AI: From Large Language Models to Thinking Systems."
Key framing:
- From GPT (2018) to ChatGPT (2022), o1 (2024), and Claude Code (2025), AI progress appears to be accelerating rather than slowing.
- The gains aren't from any single technique: like processors moving from 2D to 3D integration, the field repeatedly hit walls and pivoted paradigms.
- Modern AI is no longer a standalone language model but a "thinking system" with the model as just one component — capable enough to tackle open math problems.
The tutorial walks through four key recent topics — Test-Time Scaling, Loop, Harness, and RSI — covering each from three angles: phenomena and methods, theory and interpretation, and practice, with contributions from the author's own research group.
More from coding & agent
- Open-source VoiceStudio offers local voice cloning and dubbing in 646 languages, free ElevenLabs rival — thisdudelikesAI · 2026-10-28
- MCP Python SDK maintainer wins Agentic AI Foundation Outstanding Technical Contribution Award — samuelcolvin · 2026-10-28
- Maestro adds GitHub Label Added trigger that spins up agent pipelines — Dave_Maynor · 2026-10-28
- Agents Don't Call Tools: The LLM Only Generates Text You Parse and Execute — ariG23498 · 2026-10-28
- Hillock v1.0: local neuro-symbolic memory engine runs under 1.2GB VRAM with claimed zero hallucinations — Equivalent-Flan-1590 · 2026-10-28
- toolmoat: open-source proxy pins MCP tool descriptions to block mid-session tool-description poisoning — whitul · 2026-10-28