AI News
AI community posts, news and release updates
Latest updates
OpenAI announces ChatGPT Voice is coming to the desktop app, using the GPT Live model, now able to control your entire computer. Works with Codex and ChatGPT Work.
- codex
- desktop app
- GPT-Live
- ChatGPT Work
- ChatGPT Voice
- Launch
ChatGPT Voice Comes to Desktop: Control Your Computer and Direct Multiple Agents with Your Voice
Published Jul 23, 2026OpenAI launched ChatGPT Voice for desktop, powered by GPT-Live, allowing users to control their computer and direct multiple agents in ChatGPT Work or Codex using just their voice. Rolling out globally on macOS and Windows to Plus/Pro/Business/Edu/Enterprise plans.
- ChatGPT
- Windows
- desktop
- voice
- macOS
- GPT-Live
An in-depth analysis of DeepSeek's company philosophy and strategy, exploring how it competes in AI with a unique approach.
- DeepSeek
- 行业分析
- AI策略
A researcher used ChatGPT to find a counterexample to the Dinitz–Garg–Goemans conjecture, highlighting AI's role in mathematical discovery.
- ChatGPT
- AI Research
- mathematics
- conjecture
- Launch
Claude Voice Mode Now Runs on More Capable Models, Supports Tools and More Languages
Published Jul 23, 2026Anthropic announces that Voice mode now runs on Claude's more capable models, reaches tools mid-conversation, and supports many more languages.
- Claude
- Anthropic
- voice mode
- Launch
Codex app 26.715 released: ChatGPT Voice, screen context, multi-folder projects
Published Jul 23, 2026Codex app version 26.715 released, adding ChatGPT Voice via GPT-Live, macOS screen context for frontmost window, and multi-folder projects.
- codex
- macOS
- GPT-Live
- ChatGPT Voice
- multi-folder projects
A team announces hosting DeepSeek V4 in the US at a significantly lower cost than competitors, rolling out soon.
- DeepSeek V4
- cost reduction
- US hosting
- Launch
Show HN: Echo – Fable-level results at 1/3 the cost using open-weight models
Published Jul 23, 2026Echo achieves Fable-level performance using open-weight models at one-third the cost—a cost-effective breakthrough.
- 开源模型
- 成本优化
- Echo
- News
US & UK gov tests: China's Kimi K3 'significantly below' US frontier AI models
Published Jul 23, 2026Joint US-UK government tests find China's Kimi K3 significantly underperforms compared to US frontier AI models, highlighting America's continued lead in AI.
- frontier models
- AI evaluation
- Kimi K3
- US-UK testing
- Idea
Elon Musk: Competitors should review each other's AI models before release, like movie ratings
Published Jul 23, 2026Elon Musk proposes informal mutual review among AI competitors before model release, arguing that competitors are incentivized to flag security risks. He cites the Mythos incident where Amazon highlighted risks to the White House.
- AI safety
- Mythos
- Elon Musk
- Model Review