WIRED reported that Meta built face recognition into its smart glasses and then deleted the code. The CTO confirmed it. The feature was called Name Tag.
The easy reading is that the company backed off under pressure. The more useful one is different: the technology is already tested, modular, and ready to be switched back on tomorrow.
Picture a camera system in an apartment building. You install the hardware, install the software, but leave face recognition off by default. The code is there, compiled, waiting. All it takes is a toggle on the server. That's what happened with Name Tag, with one difference: the toggle belongs to a company that has already shown where its brakes are when data collection is on the table.
The consequence for anyone operating technology is direct. Deleted code isn't destroyed code. Anyone maintaining a system in production knows that what went into the history comes back. What matters isn't what's in the binary today, it's what was designed to be enabled remotely.
That holds for any surveillance feature. The hardware already shipped the sensor, the software already has the routine. All that's missing is a permission on a server.
The debate about creepy glasses isn't about what they do now. It's about what the manufacturer can switch on tomorrow without telling anyone.