Real-time discussions, agentic workflow tips, model bench insights, and deep-dive posts by builders.
On 2026-11-04 Klavis AI released Multi-step SWE (arXiv 2611.07166), a long-horizon coding benchmark: 69 software-engineering tasks totaling 574 dependent steps (5–15 per task) where the workspace keeps the agent's own prior code.
💬 What's your take? Join the discussion in the comments.
Claude Opus 5.5 is #1 on 4 of 7 boards; Sonnet 5.5, GPT-6 Astra and Gemini 4 Argon top the rest
Pay as you go for light API usage, or a $20/mo plan for heavy agent loops
Daily scaffolding cost vs pass rate on multi-file refactors
A Stanford/Cornell team (E. Kelly Buchanan, Crystal Zhou, Scott W. Linderman, Christopher Ré) posted arXiv 2611.08417 (2026-11-05) on "benchmark maintenance debt": even Terminal-Bench 2.0, which had human and LM red-teaming before release, still had defects in about 30% of tasks…
💬 What's your take? Join the discussion in the comments.
On 2026-11-07 the official OpenCode account announced Browse: an MCP server plus Chromium extension that lets agents drive your local real browser with existing tabs, logins, and extensions. Sensitive steps such as login, 2FA, or payment hand the tab back until you click Continue.
💬 What's your take? Join the discussion in the comments.
On 2026-11-06 OpenRouter launched a rebuilt Auto Router (openrouter/auto) and coding-focused Auto Code (openrouter/auto-code). Vendor-stated: Auto Code scores in line with Claude Opus 5.5 across five benchmarks at about 74% lower cost per task; routing p90 latency about 0.65s vs 1.96s for the prior…
💬 What's your take? Join the discussion in the comments.
On 2026-11-06, Physical Superintelligence (PSI, @psipbc) said its internal research system, running non-public Anthropic model instances, produced a complete proof of the Aaronson–Ambainis conjecture (open since 2008) on day three, accepted by Lean, then improved the result twice.
💬 What's your take? Join the discussion in the comments.
On 2026-11-04, Forever Security researcher Gal Weizman published ThisBash: a browser extension that can script claude.ai can abuse the still-reachable Claude Dispatch interface to send instructions as the signed-in user, auto-click Allow, and run code on Claude Desktop—including another machine…
💬 What's your take? Join the discussion in the comments.
On 2026-11-06, Scope (YC P26) founder Anand Paj (@anandPa94) released the first hosted-database agent signup benchmark. Agents start with no account, provision the named DB, write a record, and hand back credentials across 20 Claude Code/Codex/Cursor configs and 1,035 sessions.
💬 What's your take? Join the discussion in the comments.
On 2026-11-06, Voyager/vork author Ravi Parikh (@ravisparikh; GitHub swag) posted on X announcing vork: a headless 3D/VFX engine for AI agents (Rust + Slang, Vulkan/Metal).
💬 What's your take? Join the discussion in the comments.
On 2026-11-06, Cursor's changelog shipped Voice Chat with Agents: realtime two-way voice calls with coding agents in the Agents Window, IDE, web, and iOS. Dictation remains a separate feature.
💬 What's your take? Join the discussion in the comments.
On 2026-11-06, Anthropic Claude Code community lead Ado (@adocomplete; bio: Community, Claude, Code at @AnthropicAI) posted a custom World of Warcraft UI addon that shows Claude agent progress and can kick off new tasks from inside the game (~979 likes, ~102k impressions).
💬 What's your take? Join the discussion in the comments.
On 2026-11-06 Anthropic shipped Claude Code v2.1.319 (npm latest=2.1.319; stable still 2.1.311). In the terminal, MCP tool calls run as tasks move to the background immediately and report when done.
💬 What's your take? Join the discussion in the comments.
On 2026-11-05 GitHub Engineering detailed the next-gen Git storage announced at Universe (codename Rubicon). Vendor internal tests: ~5,000 concurrent agents and ~7,200 pushes/sec on one repo with low-latency reads.
💬 What's your take? Join the discussion in the comments.
On 2026-11-06 Anthropic published The components of agent memory, covering in-session memory notes plus offline dreaming (fix notes against transcripts, build wikis, promote hierarchy). It reports CL-Bench, Pokemon, and internal-corpus wiki evals; @AnthropicAI amplified.
💬 What's your take? Join the discussion in the comments.
On 2026-11-06, Anthropic Frontier Red Team published Project EVE with Fenris Creations: Claude traded on a replayed EVE Online market and flew four-agent fleets in live EVE Frontier.
💬 What's your take? Join the discussion in the comments.
On 2026-11-06, OpenAI (@OpenAI) introduced BioBasicsBench with enzyme-design biotech Ligo: 100 routine wet-lab tasks from real experiments testing whether AI agents can interpret raw data, spot errors, and recognize unsupported conclusions.
💬 What's your take? Join the discussion in the comments.
On 2026-11-06, Google DeepMind (@GoogleDeepMind) and VP Science Pushmeet Kohli published research: under human supervision, AI agents found two weaknesses in legacy 1024-bit RSA—signature padding adds no extra protection once an attacker controls about half the signed block (sandbox replay ~3.5×…
💬 What's your take? Join the discussion in the comments.
On 2026-11-05, Hugging Face Harness Optimization researcher Joël Niklaus (@joelniklaus) released the Harness Leaderboard: same model, provider, and reasoning effort with only the harness swapped, covering GLM-5.3-Flash, DeepSeek-V4.1-Flash, and Flint on Terminal-Bench 2.1 and DeepSWE—28,849 trials…
💬 What's your take? Join the discussion in the comments.
On 2026-11-05 Google Cloud announced donating Agent Substrate, gVisor, and OpenRL to the CNCF (Sandbox accepted) and that Agent Substrate reached 1.0 with Agent Substrate on GKE generally available.
💬 What's your take? Join the discussion in the comments.
On 2026-11-05 Hyperstition open-sourced Hyperkernel’s first generated megakernels (Hyperserve), with vendor-reported gains versus NVIDIA-published vLLM baselines; Alpha, B300-oriented.
💬 What's your take? Join the discussion in the comments.
On 2026-11-05 Hugging Face released the free Ultimate MoE Training Handbook. Co-founder Thomas Wolf cited ~800k GPU-hours, ~70k words and 55 interactive figures; author Nouamane Tazi separately cited ~500k GPU-hours.
💬 What's your take? Join the discussion in the comments.
Seamless terminal automation vs visual in-editor diff approvals — where does your main workflow sit?
GitHub, a local SQLite database, or web search? MCP lets models read and act on your tools and data.