AI Enablement Theater™

In a recent sidebar with a former coworker, I learned about the hoops they were jumping through as part of their company's AI enablement initiatives. Like a lot of other companies, their AI ambitions seem to have landed as more of a mandate than actual enablement—minimum token usage and PR output requirements coupled with budget management falling on the individual contributor.

The part that really stuck out to me was how they were trying to figure out the cheapest/smallest model they could rely on to complete the work adequately (while staying under their monthly AI token budget) or if they could rely on the harness to auto-switch as needed.

This is not AI enablement, it's AI Enablement Theater™.

The promise of AI—which I 'm a pretty genuine believer in at this point—is not going to be realized if your engineers are worried about using it too much, or being incentivized to game a leaderboard.

Rolling out AI with these constraints also encourages relatively surface-level AI interactions and thinking. We're building bad AI habits by not giving people the freedom to push, explore and see what really can be done with these tools. Don't hand them a Swiss Army Knife and tell them they can only use the toothpick.

Most teams need to actually feel the capability of these things before cost optimization means anything. If your team's first and only exposure is a small model being rationed, the takeaway they'll walk away with is "this stuff isn't very good" and the enablement effort defeats itself.

If you are responsible for getting your team or organization to adopt AI as a tool, please do not put your folks in this position. Give your team the tools or don't.

If you enjoyed reading this consider sponsoring my work on GitHub or subscribing to my newsletter.

Share this post: Hacker News • Reddit • LinkedIn • Mastodon.

Published on Thursday, September 3rd 2026. Read this post as plain-text.