AI代理通过新的编码工具和推理能力取得进展

报道来源 [9]

1. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] · 2026-05-20 09:04

   OpenAI GPT-Realtime-2：GPT-5 级别推理对语音代理的实际改变 OpenAI 的 GPT-Realtime-2 是首个具备 GPT-5 级别推理能力的语音模型

   OpenAI GPT-Realtime-2: What GPT-5-Class Reasoning Actually Changes for Voice Agents OpenAI's GPT-Realtime-2 is the first speech model with GPT-5-class reasoning. Here's what genuinely changes for voice agents — and what to test before you migrate. https:// pickuma.com/posts/opena…

   链接 pickuma.com/…/openai-gpt-realtime-2-voice…
2. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] · 2026-05-20 09:04

   如何构建一个能连夜提交 GitHub PR 的自主 AI 编码代理：自主 AI 编码背后的计划-执行-验证循环的实用解析

   How to Build an Autonomous AI Coding Agent That Opens GitHub PRs Overnight A practical breakdown of the plan-execute-verify loop behind an autonomous AI coding agent, and how to wire it to GitHub so an issue becomes a reviewable pull request overnight. https:// pickuma.com/posts/…

   链接 pickuma.com/…/build-autonomous-ai-coding-…
3. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] · 2026-05-20 09:03

   持续利用：Gemini Pokémon Agent 重写其自身循环 Continual Harness 模式，来自 Gemini Plays Pokémon 和 PokeAgent 团队，让

   Continual Harness: The Gemini Pokémon Agent That Rewrites Its Own Loop How the Continual Harness pattern, from the Gemini Plays Pokémon and PokeAgent teams, lets an agent rewrite its own harness mid-run — plus how to apply that online-adaptation idea to autonomous agents you buil…

   链接 pickuma.com/…/continual-harness-gemini-se…
4. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] · 2026-05-20 09:03

   oh-my-agent v2：九项新技能、一流的 Cursor 和 80/100 基准 oh-my-agent v2 增加了九项新技能，将 Cursor 提升为一流供应商，并

   oh-my-agent v2: Nine New Skills, First-Class Cursor, and an 80/100 Benchmark oh-my-agent v2 adds nine new skills, promotes Cursor to a first-class vendor, and ships a benchmark scoring 80/100. A measured look at whether it fixes the agent failures developers actually hit. https:/…

   链接 pickuma.com/…/oh-my-agent-v2-nine-skills-…
5. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] · 2026-05-20 09:03

   Judea Pearl 的因果关系阶梯与 LLM 推理的局限性 Judea Pearl 的三层因果层级——关联、干预、反事实——解释

   Judea Pearl's Ladder of Causation and the Limits of LLM Reasoning Judea Pearl's three-rung causal hierarchy — association, intervention, counterfactual — explains why data-driven ML and LLMs hit a structural wall at causal reasoning, and what that means for agents and RAG. https:…

   链接 pickuma.com/…/judea-pearl-causal-hierarch…
6. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] · 2026-05-20 09:03

   Apify Fingerprint Suite：用于隐形抓取器的开源浏览器指纹识别 Apify 的 fingerprint-suite 可生成统计上一致的浏览器指纹

   Apify Fingerprint Suite: Open-Source Browser Fingerprinting for Stealth Scrapers Apify's fingerprint-suite generates statistically consistent browser fingerprints and injects them into Playwright or Puppeteer. How it works, how to wire it in, and when a scraper actually needs it.…

   链接 pickuma.com/…/apify-fingerprint-suite-ste…
7. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] · 2026-05-20 09:02

   Codex Auto Review Loop：一个在您提交代码前进行审查的 MCP 工具 codex-mcp-code-review 是一个开源的 MCP 服务器，可自动执行 Codex 的 /review 流程 f

   Codex Auto Review Loop: An MCP Tool That Reviews Code Before You Commit codex-mcp-code-review is an open-source MCP server that automates Codex's /review flow for uncommitted changes by spawning background Codex instances. Here is how the review loop fits an agentic coding workfl…

   链接 pickuma.com/…/codex-auto-review-loop-mcp-…
8. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] · 2026-05-20 09:02

   Conductor 加入云端编码代理热潮：远程 AI 开发者离开笔记本电脑 Conductor 与 C 的后台代理一起进入云端编码代理类别

   Conductor Joins the Cloud Coding Agent Rush: Remote AI Devs Leave the Laptop Conductor enters the cloud coding agent category alongside background agents from Cursor, GitHub, OpenAI, and Google. What changes when your AI coding agent runs on remote infrastructure instead of your …

   链接 pickuma.com/…/conductor-cloud-coding-agen…
9. Mastodon — fosstodon.org TIER_1 English(EN) · [email protected] · 2026-05-20 09:01

   Optuna教程：在Python中自动化机器学习模型的超参数调优 Optuna的define-by-run API、TPE采样器和pruners如何自动化超参数调优

   Optuna Tutorial: Automate Hyperparameter Tuning for ML Models in Python How Optuna's define-by-run API, TPE sampler, and pruners automate hyperparameter tuning for scikit-learn, PyTorch, and TensorFlow models, with runnable Python code. https:// pickuma.com/posts/optuna-tutor ial…

   链接 pickuma.com/…/optuna-tutorial-hyperparame…