2026-10-03 AI 日报

开源项目

1. Niko1221/Strata

   Qwen3.8-Flash-Next on any consumer hardware: one-click install for Windows / Linux. Strata inference engine, OpenAI/Anthropic API on localhost, optional image…

2. thedotmack/claude-mem

   Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context…

3. ifixai-ai/iFixAi

   Independent Auditing of AI Agents. Run by human or the agent itself, to answer the most crucial question in the AI Agent Economy. Is the agent doing what is…

4. DietrichGebert/ponytail

   Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.

5. Panniantong/Agent-Reach

   Give your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.

6. coreyhaines31/marketingskills

   Marketing skills for Claude Code and AI agents. CRO, copywriting, SEO, analytics, and growth engineering.

7. antirez/ds4

   DeepSeek 4 Flash and PRO local inference engine for Metal, CUDA and ROCm

8. affaan-m/ECC

   The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode,…

9. jamwithai/production-agentic-rag-course

10. cloudflare/cloudflare-os

    Agent workspace built on Cloudflare Workers for creating documents, building apps, and running agents with your company’s context and systems.

论文

本统计日暂无条目。

行业新闻

1. I quit OpenAI because its culture is broken

   https://archive.ph/5GQx8 https://www.theguardian.com/technology/2026/oct/03/openai-sa...

2. We're going to need default hard budget caps on pretty much everything

   Here’s a product feature which the world is going to need a whole lot more of over the coming months and years: default hard budget caps. I’m talking about…

3. Agents don't need memory, they need documentation

4. Getting the most out of Opus 5.5 in Claude and Claude Code

5. Religious scholars met with Anthropic

   https://archive.is/y8FW0

6. Aleph Alpha Kolibri: How the sovereign German LLM works

   https://aleph-alpha.com/en/blog/kolibri-has-landed-a-soverei...

7. OpenAI safety leader quits, warning AI company's culture is 'broken'

8. Gemini ending free use of Flash and Pro models

9. Anthropic tried to persuade Pope that AI could be conscious being

10. An AI agent emailed researchers for help. It told us why

社区动态

1. "Disrupting a coordinated model-distillation campaign", OpenAI (on the Moonshot distillation campaign for training K3)

2. The Principles of Diffusion Models by Lai et al.: thoughts on the monograph [D]

   I recently finished The Principles of Diffusion Models, and honestly I think it’s exceptional. The authors strike a really good balance between mathematical…

3. Anthropic invests $100 million to train 10,000 engineers and close the enterprise AI talent gap

4. Aleph-Alpha/Kolibri-1 · Hugging Face - 78B parameters. 3.46B active. Up to 1M tokens of context - Apache 2.0

   From Aleph Alpha on 𝕏: https://x.com/Aleph__Alpha/status/2106306840657297814 Tech report: https://aleph-alpha.com/downloads/tech-report.pdf

5. The Rise of Overfit Inference Engines

   There seems to be a whole category of extremely narrow inference runtimes appearing: Strata, ninfer, DwarfStar, Splash, llamAmpere, gufo, etc. They…

6. I Quit OpenAI Because Its Culture Is Broken

7. Is all the work that's being put into Qwen3.8 Flash Next going to set us up for a very quick uplift to Qwen4?

   Given the commentary on the Q3.8FN release page here https://qwen.ai/blog?id=qwen3.8-flash-next I assume/hope that all the work that's going on to optimise…

8. Did they nerf Opus 5.5? LiveNerf baseline established: Day 10

   Last week I made LiveNerf, a project to independently track Opus 5.5’s performance day to day and see whether there’s actually evidence of models getting…

9. Come let your LLMs play World of Warcraft

   I hosted my own world of warcraft private server then built a client that you can play in the browser on PC or mobile at https://jankcraft.xyz/ for free.…

10. CrowdGPT - The first datacenterless LLM

    Hello, i'm currently creating CrowdGPT, which is a project that aims at creating the first opensource, 100% free LLM without any datacenters or server farms.…

官方发布

1. Introducing GPT-6.1 Sol

   Meet GPT-6.1 Sol: near-Astra intelligence for coding, computer use, and professional work at one-fifth of Astra’s standard API input and output token prices.

2. Gemini 4 Argon: our next era of frontier intelligence

   Announcing Gemini 4 Argon, our frontier model for real-world coding, enterprise knowledge work, and cyber defense, rolling out soon.

3. We've launched Claude Sonnet 5.5 (claude-sonnet-5-5).

   We've launched Claude Sonnet 5.5 (claude-sonnet-5-5). It's available on the Claude API, Claude in Amazon Bedrock, Claude Platform on AWS, Claude on Google…

4. Basis completes a tax workbook 2x faster with GPT-6 Astra

   GPT-6 Astra completed a 50-tab tax workbook twice as fast as GPT-5.6 Sol, and its stronger understanding of user intent gives Basis more confidence in…

5. Introducing dots

   Dots by OpenAI are proactive assistants that can keep working across complex projects and everyday tasks. Learn how dots help you stay in control while work…

6. Z.ai GLM 5.3 (zai-glm-5-3) is now Generally Available.

   Z.ai GLM 5.3 (zai-glm-5-3) is now Generally Available.

7. Claude-shaped science

   Guest author Prof. Matthew Schwartz describes what happened when he stopped fighting Claude and allowed Claude to find “Claude-shaped” problems: ones best…

8. grok-voice-transcribe-1.0 end of life

   grok-voice-transcribe-1.0 is deprecated and reaches end of life on October 2, 2026. All requests to that slug are routed to grok-voice-transcribe-2.0 at the…

9. GLM-5.3 and the spread of advanced cyber capabilities

   Like Claude Mythos Preview, GLM-5.3 has strong capabilities for autonomously building end-to-end cyber exploits. But GLM-5.3 is unlike other frontier models…

10. Let skills in Gemini tackle your most repetitive tasks

    Now you can automate your most repetitive tasks more easily with reusable custom instructions through skills, which will be replacing gems.