'They want to build a moat': Anthropic's scary warnings about rapid AI 'self-improvement' and 'temporarily' pausing development aren’t convincing the cynics
- Anthropic suggests a "slowdown or pause" on AI
- AI could soon start developing itself, the company warns
- There have been mostly negative responses online
Even the giant companies at the forefront of AI are suggesting it might be time to take a step back and think about where the tech is heading: Claude developer Anthropic has posted a lengthy new blog post in which it suggests "a meaningful slowdown or pause" while we all take a breath.
Authored by Anthropic executives Marina Favaro and Jack Clark, the post centers on the idea of AI developing itself, known as 'recursive self-improvement' — at which point it might get out of our hands very quickly, as AI takes over the business of designing and developing its own models and interfaces.
"We are not there yet, and recursive self-improvement is not inevitable," explains the post. "But it could come sooner than most institutions...
Copyright of this story solely belongs to techradar.com. To see the full text click HERE
