AI News
AI community posts, news and release updates
Latest updates
A writing fingerprint analysis of model responses reveals that Kimi K3 closely resembles Anthropic's Claude in writing style, sparking discussion about model origins and potential copying.
- Claude
- Kimi
- 指纹分析
- 模型相似性
- Launch
Etched Raises Funding: AI Inference Chip by Three Harvard Dropouts, Surpassing $1B in Contracts
Published Jul 23, 2026Etched, an AI chip startup founded by three Harvard dropouts, raised major funding (including SK Hynix). Its first chip is purpose-built for modern model inference, optimized from materials to packaging, and has already secured over $1B in customer contracts.
- AI芯片
- 推理
- 投资
- Etched
- Idea
Salesforce executive praises Anthropic's hat trick: Claude, Claude Code, and Cowork
Published Jul 23, 2026A Salesforce executive lauds Anthropic's revenue growth and its enterprise hat trick: best model Claude, best coding agent Claude Code, and best productivity tool Cowork, with Salesforce proud to be an early investor and partner.
- Claude
- Claude Code
- Anthropic
- Cowork
- News
Global Open Model Adoption Dashboard: US Slowly Growing, Still Behind China/Qwen
Published Jul 23, 2026A daily-updated dashboard shows open model adoption data for US, China, and globally. The US share is slowly growing but still far behind China and Qwen.
- 开源模型
- 采用率
- 仪表板
A user praises Opus 5 for the best 3D output they've seen from any model, but complains it's still token-hungry and slow as hell.
- Anthropic
- Kimi
- 吐槽
- 3D
- Opus 5
- News
Anthropic's Marketing Strategy for Fable 5 Fails to Meet Expectations, Overshadowed by GLM, GPT, etc.
Published Jul 23, 2026Analysis shows Anthropic's hype around Fable 5 backfired. Fable 5 was suspended globally on June 12, GLM 5.2 launched the next day, and by the time Fable 5 relaunched on July 1, it was already surpassed by GLM 5.2, with GPT 5.6 and Kimi K3 catching up, making the marketing strategy ineffective.
- Anthropic
- GPT
- GLM
- Fable 5
- Kimi K3
- 营销策略
An innovative approach that performs LLM inference computation directly within memory, drastically reducing data movement and potentially breaking performance bottlenecks, with profound implications for AI hardware architecture.
- LLM推理
- 架构创新
- 内存墙
- 存内计算
- News
Anthropic revamps free Claude Code course with 10 new lessons and final exam
Published Jul 23, 2026Anthropic has completely renewed its official free course 'Claude Code in Action', now featuring 10 new video lessons and a final exam.
- Claude Code
- Anthropic
- 免费课程
- 教程
- Launch
Google announces Gemini 3.6 Flash and cybersecurity AI, teases 3.5 Pro and Gemini 4
Published Jul 23, 2026Google unveils Gemini 3.6 Flash and a cybersecurity AI, with hints at upcoming Gemini 3.5 Pro and Gemini 4.
- AI
- Gemini
- cybersecurity
- Flash
- Discussion
Kimi K3 is amazing! Recreated famous motion design team's video, proving them wrong
Published Jul 23, 2026HyperFrame team used Kimi K3 to recreate a famous motion design team's iconic video with high fidelity. The twist: that team had previously claimed Kimi K3 couldn't produce such videos.
- 视频生成
- AI创作
- Kimi K3
- 动效