Business, Technology, Internet and network concept. Young busine
© Marina Putilova

PLATFORMER‘s Casey Newton writes:

For the past few months, the big artificial intelligence labs have signaled that 2025 will be a breakout year for agents: tools that use computers on your behalf, working across different apps to perform multi-step operations just like humans do. 

Agents have rapidly become a top priority at AI labs, as they represent a key milestone on the way toward building full-featured virtual coworkers. Starting last year, a steady drip of announcements signaled that researchers are rapidly making progress toward that goal.

Anthropic introduced its first take on agents, which it calls “computer use,” into its API in October. Two months later, Google said that its Gemini 2.0 models were designed for “the agentic era.” 

Despite those announcements, though, until now most agents have only been available to developers. As of this week, though, that’s starting to change….

To read the full post, please click here. This comes just as the debate is open about OpenAI’s future value, pencilled in at $60bn

Comment on this article


You must be logged in to post a comment.