On today's chart 594K views since yesterday

Low Level examines what smart TVs may be listening for

His reverse-engineering analysis connects reported TV telemetry with a bigger question: how much control do buyers have over their own devices?

Channel:

Low Level Story of October 30, 2026Trari Newsroom

No. 50 on the daily chart No. 5 on the Tech chart

Why this story? One of the most watched videos since yesterday, with 593,682 views.

Reverse Engineering Proves What We Suspected All Along..

Things we can learn from the video

Press to jump

0:45Reverse engineering investigates how hardware or programs work without access to their source code.
4:41A researcher uses AI and black-box testing to gain privileged access to a Samsung smart TV.
5:31Low Level reviews reported TV telemetry that counts dog-barking and baby-crying events for targeted advertising.
6:56An engineer uses Claude and physical debug contacts to investigate an abandoned smart-home device.
7:44A robot-vacuum investigation reportedly reveals a vulnerability permitting remote access to thousands of devices.
8:30The host argues that locally available AI models will make consumer hardware investigations harder to restrict.

A barking dog becomes an advertising signal in the research Low Level examines in his new video. So does the sound of a baby crying. The host presents the reported smart-TV findings as a particularly unsettling example of what AI-assisted reverse engineering can reveal inside everyday hardware.

"Reverse Engineering Proves What We Suspected All Along.." drew roughly 594,000 views since yesterday. Behind that sweeping title is a concrete argument: AI tools are making it easier to investigate devices whose manufacturers do not publish their source code.

The host starts with a simple explanation of reverse engineering, then turns to a researcher known as log4j, who used AI and black-box testing to gain privileged access to a Samsung smart TV. Reviewing the research and related reporting, he says the television's telemetry counts dog-barking and baby-crying events for advertising purposes. These are findings he is discussing, not tests he performs himself in this episode.

The other examples extend beyond listening. An engineer uses Claude to investigate an abandoned smart-home device through its physical debug contacts. Another researcher's investigation of a DJI Romo robot vacuum reportedly exposes a vulnerability allowing remote access to thousands of devices.

The video includes a paid CodeRabbit segment. Its broader discussion turns toward a tension within AI itself: tools can help owners understand their hardware, while safety restrictions can limit the same investigative work. Low Level argues that locally available models will make those investigations harder for manufacturers to prevent.

Viewers pick up the ownership question, discussing custom firmware and whether a device that cannot run its owner's code feels more like a leased service than a possession. That is where the episode ultimately lands: not just what connected appliances collect, but who gets to examine and change them.

Published on Trari