Skip to content
AI

AI Agents Aren't Ready: The Benchmark That Deflates Automation Hype

The DELEGATE-52 benchmark tested 52 professional domains and found AI agents corrupt outputs in long task chains. Only Python cleared the reliability bar.

Francesco CampisiBy Francesco Campisi 4 min read How we work

Related Articles

AI and Tech's Impact on Jobs and Productivity: What the Data Shows Mattia Mezzetti Mattia Mezzetti 17 Oct 2026 AI and Tech's Impact on Jobs and Productivity: What the Data Shows AI could add up to 1 percentage point yearly to U.S. labor productivity, per Luiss Business School research. Young workers aged 22-25 are already bearing the… Schneider Electric Data Centers Drive a Record 11.5B Quarter Hamza Ahmed Hamza Ahmed 17 Oct 2026 Schneider Electric Data Centers Drive a Record 11.5B Quarter Schneider Electric posted 11.5 billion euros in Q3 2026 revenue, up 17% organically. Data center demand grew at a triple-digit rate, but the headline needs…
Venture Global Plans 10 GW Off-Grid AI Power: What's Real and What's Not Venture Global Plans 10 GW Off-Grid AI Power: What's Real and What's Not Francesco Campisi 16 Oct 2026 Adobe and NVIDIA Speed Up AI Workflows for Creators and Studios Adobe and NVIDIA Speed Up AI Workflows for Creators and Studios Hamza Ahmed 16 Oct 2026