Mousepower: agents that can’t be measured, can’t be managed. — Maximillian Piras, Yutori
Maximillian Piras from Yutori argues that AI agents — programs that control a computer like a human would — have a measurement problem.
We count tokens (small text units that agents consume) but don't know if that actually delivered value. His comparison is James Watt's "horsepower" from the 1700s: an imprecise but useful metric that got people to try steam engines. Piras suggests we instead focus on concrete results — bugs fixed, support tickets resolved. He also presents an "entropy matrix" for choosing which tasks suit agents: the best tasks are neither too simple (a script suffices) nor too complex (too hard to verify), but fall in a middle ground where it's cheaper to check a result than to solve the problem from scratch.
Vibekollen prepared this summary with AI from the original publication. The content belongs to AI Engineer.
More to read
Lumaril, which develops AI software for optimizing industrial production by connecting to existing control systems, raised a $66M Series A at a $400M valuation
Techmeme 3 h ago
DeepSeek open sources Super-Agent as a research prototype, a proactive coding agent it says can monitor group chats for coding issues and suggest solutions
Techmeme 3 h ago
Insurance software maker Guidewire agrees to acquire Hyperexponential, which makes AI-powered insurance underwriting and pricing software, for $1.7B in cash
Techmeme 3 h ago
One word for AI agents in 2027
Nate Herk 7 h ago