Stop Vibe-Checking Models: A Repeatable Comparison Harness You Can Run on Free Compute

Liquid syntax error: Unknown tag ‘endraw’

Top comments (1)

Subscribe

Collapse Expand

Alex Shev

Building AI-powered tools. Creator of seog.ai & terminalskills.io — curated terminal skills and CLI tools for modern devs.

• Location

  Dallas-Fort Worth, Texas

• Work

  Founder at SEOG & Terminal Skills

• Joined

  Mar 7, 2026

• Aug 13

Repeatability is the useful part of the harness. A model comparison should produce artifacts another developer can rerun, not just screenshots of a good answer. The more boring the harness, the more credible the result.

Code of Conduct • Report abuse

For further actions, you may consider blocking this person and/or reporting abuse