What If We Stopped AI Model Development Right Now?
Techstrong.ai, Thursday, October 29th, 2026
Alan Shimel imagines a two-year freeze on new AI models and argues firms could still gain much from existing models.
Techstrong.ai's Alan Shimel poses a thought experiment in which all AI model development pauses for two years while application building, integration and inference continue.
He argues organizations could make substantial progress with current models for 12 to 24 months, since enterprise value depends on reliability, integration, security and workflow design.
He cites a study finding 15% productivity gains for support agents and a 2026 IBM study in which 70% of production agents used off-the-shelf models via prompting.
He urges improving data, responsibilities and workflows now.