Tech · Import AI
Import AI 455: AI systems are about to start building themselves.
Compiled by KHAO Editorial — aggregated from 1 outlet. See llms.txt for citation guidance.
◌ Single Source
Welcome to Import AI, a newsletter about AI research.
Key facts
- Ajeya Cotra, a longtime AI forecaster who works at METR, thinks it isn’t unreasonable to expect AI systems to do tasks that take ~100 hours by the end of 2026 ( #448 )
- The specific eval scores are derived by a “weighted average is taken across all post-trained LLMs (Qwen 3 1.7B, Qwen 3 4B, SmolLM3-3B, Gemma 3 4B) and benchmarks (AIME 2025, Arena Hard, BFCL, GPQA
- In December 2025 one of the authors of CORE-Bench declared the benchmark ‘solved’, with an Opus 4.5 model achieving 95.5%
- For each run, they ask a CLI agent to maximize the performance of a specific base LLM on a specific benchmark
Summary
AI systems are about to start building themselves. The reporter is writing this post because when the reporter look at all the publicly available information the reporter reluctantly come to the view that there’s a likely chance (60%+) that no-human-involved AI R&D - an AI system powerful enough that it could plausibly autonomously build its own successor - happens by the end of 2028. In terms of timing, the reporter doesn't expect this to happen in 2026. From this data the reporter arrive at the conclusion that all the pieces are in place for automating the production of today’s AI systems - the engineering components of AI development. Two things that exemplify this trend are SWE-Bench and the METR time horizons plot.