TLDRocket
Sign in

Agent Evaluation

53 summarised stories about Agent Evaluation, each linking back to the original source. Browse all topics →

+ Follow this topic

Friday, 22 May 2026

Did Google’s AI agents really build an operating system for $916?

AI as Normal Technology 4 months ago 53

Google claimed its AI agents built an operating system for $916 using Gemini 3.5 Flash and Antigravity 2.0, but the company's disclosure was misleading about key details including that the "single prompt" was thousands of lines long and the system used specialized scaffolding with subagents and anti-cheating measures. The cost reported was $916.92 with a total token budget of 2.6 billion tokens, but Google did not release the lengthy prompt, agent logs, or code, making independent verification impossible. The experiment demonstrates that AI agents can work autonomously on long-horizon tasks but highlights the need for rigorous methodological standards in open-world evaluations rather than relying on vendor-conducted press releases.

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.