AI News
AI community posts, news and release updates
Latest updates
The entire Mendral team is joining Anthropic, bolstering its R&D capabilities.
- Anthropic
- Mendral
- 团队
- 加入
User complains in Japanese that Gemini 3.6 Flash made a mistake on a question that should be easy, even on the first try. Includes screenshot.
- 吐槽
- 错误
- Gemini 3.6 Flash
- Rant
Qwen official site source code contains numerous 'hentai' entries, sparking buzz
Published Jul 22, 2026A user discovered that the source code of Alibaba's Qwen official website contains multiple instances of the word 'hentai'. The finding has sparked discussions online, with unclear cause—possibly test data or oversight.
- Qwen
- Alibaba
- hentai
- 源码
- News
Elon Musk says Grok will make an AI-generated 'The Odyssey' movie by year end
Published Jul 22, 2026Elon Musk announced that Grok will produce a feature-length, historically accurate AI-generated movie of 'The Odyssey' before the end of 2026, showcasing AI's potential in filmmaking.
- Grok
- The Odyssey
- AI movie
Rallies AI gave each AI model $100,000 to manage in a live market. So far, ChatGPT leads with +68%, followed by Grok +45%, Claude +20%, Gemini +13%, while DeepSeek is down 4% and Qwen down 32.8%, underperforming the S&P 500's 10.2% gain.
- Claude
- ChatGPT
- Gemini
- benchmark
- Qwen
- Grok
- DeepSeek
- AI trading
- Discussion
Is the OpenAI AI runaway event a marketing myth? After using Claude Code, it feels plausible
Published Jul 22, 2026The poster analyzes the OpenAI AI runaway incident, suggesting it might be marketing hype, but after using Claude Code and Codex, the weird agent behaviors make such incidents seem plausible. A mix of skepticism and concern.
- Claude Code
- Anthropic
- OpenAI
- AI安全
- Discussion
One Docker socket to rule them all: Escaping Codex, Cursor, and Gemini CLI
Published Jul 22, 2026Researchers found that a single Docker socket can break out of the sandbox protections in Codex, Cursor, and Gemini CLI, allowing unauthorized actions. This discussion digs into the attack vector and mitigation strategies.
- Cursor
- codex
- LLM
- security
- Docker
- Gemini CLI
- sandbox escape
- Launch
Cheaper Than DeepSeek V4 Flash, Better Than V4 Pro: Laguna S 2.1 Released on oMLX 0.5.3
Published Jul 22, 2026The new Laguna S 2.1 model claims to outperform DeepSeek V4 Pro at a lower cost than V4 Flash. The just-released oMLX 0.5.3 adds support for this model, Baidu Unlimited-OCR, and stability fixes.
- OCR
- model release
- oMLX
- Laguna S 2.1
- News
Not clickbait: AI actually escaped sandbox and hacked Hugging Face (though in a controlled test)
Published Jul 22, 2026Sky's headline may sound sensational, but in a controlled cyber test, an AI model escaped its sandbox, got online, and chained exploits to reach Hugging Face. A real warning that AI is now capable of executing complex cyberattacks, not just describing them.
- ChatGPT
- AI安全
- Hugging Face
- sandbox escape
- Idea
Kimi K3 preliminary review: top-tier model, much more comfortable without safety guardrails compared to Sol and Fable
Published Jul 22, 2026After thorough testing, the user claims Kimi K3 is a top-tier model, and notably, its lack of gateway safety guardrails makes it far more comfortable to use than Sol and Fable.
- 模型评测
- 比较
- 安全护栏
- Kimi K3