Do browser agents actually help with tasks?

As was long predicted, OpenAI has released a browser: ChatGPT Atlas. It's a pretty clear attempt to attack Google's $100B search moat and establish a new surface for all the stuff OpenAI wants to do with agentic AI, personal assistants and all the rest of it.

But is it any good? We find out. Also, another conclusion to TIKTOK BANWATCH.

See omnystudio.com/listener for privacy information.