Ryan Greenblatt says today’s AI still cannot automate AI R&D or a typical SWE job
RyanGreenblatt · x · 2026-07-27
Ryan Greenblatt argues that the linked piece is wrong on three fronts: AI capabilities, recursive self-improvement, and the near-term trajectory of the field. In his view, today’s systems still are not capable enough to greatly accelerate AI R&D, let alone automate AI R&D or the compute supply chain end to end.
He also pushes back on the idea that current models can fully automate a typical software engineer’s job, saying this is nowhere near AGI or ASI yet.
More from AGI Musings
- Gary Marcus says ARC-AGI’s name makes people think it tests AGI itself — GaryMarcus · 2026-07-27
- Fchollet says personality is partly innate, but can be reshaped from 15 to 25 — fchollet · 2026-07-27
- Creator says AGI is basically here and we are already in the singularity — yacineMTB · 2026-07-27
- Fake AGI Club launch pitches “GoalOS Singularity Navigator Ω” as a singularity dashboard — Ghost_Pilot_MD · 2026-07-27
- OpenAI researchers urged to probe whether models can become long-term misaligned — deanwball · 2026-07-27
- A proposal calls for LLM prompts to replace peer review’s first pass — ChenhaoTan · 2026-07-27