Latest

Every published item, in order · 9845 items
Nov 7
PapersBuildersMulti-sourceSelected22:02
Google Research Releases SLEEP Framework for Model Memory Consolidation

Unlike Anthropic's Dreaming, which organises external notes and knowledge bases, SLEEP writes learned information directly into model weights — a different bet on how models could retain knowledge permanently. Practitioners running continual learning or fine-tuning pipelines can study the trade-off: stronger retention at 4x the training compute, so far validated only on Qwen3-8B and specific tasks.

Event
shao__meng@shao__mengSource
PapersBuildersMulti-sourceSelected08:57
Harness Leaderboard ranks nine coding-agent harnesses on open models

The harness you wrap around an open model can swing both solve rates and cost by large margins — opencode ran to US$3.35 per solved task, 11.7 times Pi's cost, on identical models. Teams self-hosting GLM, DeepSeek or Kimi models should pick a harness per model rather than a single default. The companion Harness Lens documents prompts, tools and compaction across 34 harnesses with source links.

Event
Thomas Wolf@Thom_WolfSource
PapersBuildersSelected07:03
Stanford Introduces Collaborative Intelligence Benchmark for Agent Teamwork

Individual benchmark scores poorly predict how useful a model is on a team — the gap of up to 11 points between near-equal models gives builders a new selection signal for multi-agent stacks. The persistence of detected cheats through code handoffs is a concrete red flag for anyone running agent handover pipelines. Developers building agent teams can start with the 160 public tasks and the code and data on arXiv (2611.02633).

elvis@omarsar0Source
Tomorrow 08:00, the daily lands in your inboxTelegram channel or email. Same screening as this page, no ads.
PapersBuildersSelected01:13
Exo Agent Runtime Enables In-Harness Self-Editing During Task Execution

The shift from offline harness optimization to runtime self-editing could reduce the dataset and compute burden that makes agent tuning expensive. Builders of long-running agents can inspect Exo's approach for keeping edits auditable through a protected execution-history layer. Compared with offline loop tuning, it trades a small architecture change for continuous adaptation on live tasks.

DAIR.AI@dair_aiSource
PapersBuildersMulti-sourceSelected01:13
EnvOpt paper optimizes agent workspaces for accuracy and cost

Most agent tuning effort goes into prompts and model choice, but EnvOpt shows the surrounding workspace itself is a trainable artifact with measurable returns. Teams building multi-step agents in shared tools can reuse the same optimized environment across stronger models to cut inference costs. The method is documented in the paper, with a chat-with-paper interface linked for quick exploration.

Event
DAIR.AI@dair_aiSource
PapersBuildersMulti-sourceSelected01:01
Lambda sponsors CoRL 2026 workshops and presents I3L paper

Human takeover through touch gives operators a practical fallback when autonomous policies fail, which is one of the friction points holding back real-world robot deployment. Robotics and embodied-AI teams attending CoRL 2026 in Austin can catch both workshops on 12 Nov to see the method demonstrated across three robot embodiments. The results compare directly against imitation-learning baselines, offering a concrete benchmark for labs evaluating interactive learning pipelines.

Event
Lambda@LambdaAPISource
Nov 6
PapersBuildersSelected22:47
Amazon research shows weight memory beats retrieval for agent recall

The practical implication is a hybrid memory design: keep a text store for recent and never-recorded facts, and distill stable user knowledge into model weights. Weight memory alone forgets older facts and cannot flag gaps, so the router matters. Engineers building personalised agents can test this against MemOS-style retrieval pipelines. The paper is available on arXiv (2611.05636) with a Chat with Paper interface from dair.ai.

elvis@omarsar0Source

Only the latest 2000 items are listed; for older ones see the daily archive