深度研究｜递归自我改进（RSI）全景 2026：AI 正在加速 AI，但「验证瓶颈」决定它能走多远
梳理 RSI 从 Good 1965 到 2026 的思想史、技术图谱与一手实证：Anthropic 承认其代码库 >80% 合并代码由 Claude 撰写、METR 测得 AI 可完成任务时长约每 4 个月翻倍、AlphaEvolve 优化了支撑自身的计算栈。核心判断——有界自我精炼已工程化，开放式 RSI 尚未发生；而进步与安全共享同一个「验证瓶颈」。
阅读全文 →
最新文章
LATEST 共 515 条What LLM Trading Agents Actually Do in Production: A Six-Month, Population-Scale Record from Two Fleets
We present a continuous, population-scale measurement record of autonomous language-model trading agents operating in pr…
Refuse without Refusal: A Structural Analysis of Safety-Tuning Responses for Reducing False Refusals in Language Models
Striking a balance between helpfulness and safety remains a fundamental challenge in aligning large language models. To …
What Did I Just Say? Self-Listening for Full-Duplex Speech Models
Full-duplex spoken language models can listen and speak simultaneously, enabling them to handle interruptions and backch…
RISE: Recursive Improvement via Self-Extrapolating Policy Distillation
On-policy distillation (OPD) provides dense, per-token supervision for language model post-training, but its effectivene…
Scal3R: Learning Efficient Multi-Relative Pose Query for Scalable Online 3D Reconstruction
Online 3D reconstruction models perform poorly on long videos. This happens because regressing poses relative to a fixed…
EVOHARNESSBENCH: Can Your Agents Keep Pace with an Evolving Harness?
Modern LLM-based agents operate through a harness of tools, reusable skills, and specialist agents that shapes what they…
Iris: Climbing to the Search Frontier
We present Iris-mini and Iris-pro, two search agents trained at the 35B-A3B and 397B-A17B scales, together with the data…
DRACO: Fine-Grained Credit Assignment with Dynamic Rubrics for Long-Horizon Agent Training
Reinforcement Learning from Verifiable Rewards works well when a task has a programmatic checker, but most long-horizon …
When Quantization Breaks Memory: Recurrent-State Write-Back in Low-Precision Temporal Inference
Quantization is widely used to reduce the computational and memory demands of neural-network inference. In recurrent net…
FlowBalance: Verifier-Grounded Self-Improvement from On-Policy Reasoning Experience
A reasoning model can improve from its own on-policy experience, but this inner loop is fragile: terminal verifiers prov…
The Attention Triangle in Audio-Video Models
Audio-video diffusion models rely on cross-modal attention to coordinate text, sound, and visual content, yet this same …
When Models Edit Too Much: On the Fidelity of Minimal Code Edits
Large language models (LLMs) are increasingly used to edit existing code, but correctness alone is not enough: useful re…