this post was submitted on 23 Feb 2026
15 points (77.8% liked)

a_i

62 readers
59 users here now

Artificial Intelligence

founded 2 years ago
MODERATORS
 

When researchers put AI agents through real-world computer tasks, they failed spectacularly at things any human handles without thinking -- like dismissing a cookie banner. The gap between AI demos and actual work capability is enormous, and this study quantifies exactly how enormous.

top 1 comments
sorted by: hot top controversial new old
[โ€“] slazer2au@lemmy.world 5 points 1 day ago* (last edited 1 day ago)

Can't run a vending machine business without breaking down, not they can't even close a popup window. Fun times for language models.

Gartner's analysts found something else: most "agentic AI" products aren't agentic at all. They estimate only about 130 of the thousands of vendors claiming agentic capabilities are real. The rest are engaged in "agent washing" โ€” rebranding chatbots and robotic process automation tools with the word "agent" bolted on.

Sounds about right.