The twilight of the chatbot
Original title:The twilight of the chatbots
AI summary
Frontier AI model capabilities are accelerating, and US labs are releasing faster, but government intervention has restricted access to Claude Fable and GPT-5.6. Epoch testing found that Opus 4.7 running autonomously for 14 hours can complete software engineering that would take a human 2-17 weeks, at a token cost of $251. Chinese open-source models lag the frontier by 6-12 months, but their performance is also improving rapidly, showing an independent exponential curve on the AA-Briefcase test. Usage is shifting from chatbots to agents; a quarter of OpenAI employees run at least four agents simultaneously each week. Claude Code user data shows that domain experience, more than occupation, determines effectiveness, and experts are using agents to replace the chatbot usage patterns of non-experts.
Key Takeaways
- Frontier AI model capabilities are accelerating,and US labs are releasing faster,but government…
- Epoch testing found that Opus 4.
- 7 running autonomously for 14 hours can complete software engineering that would take a human 2-17…
- Chinese open-source models lag the frontier by 6-12 months,but their performance is also improving…
- Usage is shifting from chatbots to agents
Quick Facts
- Published
- Updated
- Category
- Opinion
- AI score
- 73
Ethan Mollick 用数据和亲身实验把话说得很直白,AI 不再是聊天工具而是能独自干上 9 小时活的 agent。做非技术岗的也在用,工作变成管理 AI,这个转变才刚开始,但已经让机构跟不上。
Related entities
Source & copyright
This article is based on curated titles and AI summaries from the AIHOT public API, with attribution and links back. Original copyright remains with the original publishers and authors; this site does not mirror full texts.
How to Cite
When citing this article, keep the title, date, and permalink, and credit Orange News as the source.
Orange News: The twilight of the chatbot. July 1, 2026. https://new.nat6.net/en/news/cmr19cqj100s1slnlze0lucb0/