Hacker News Reader

Comment by whizzter | Hacker News Reader

They should get the best metrics out of update server communications anyhow, 30% or more users getting update downloads than having AI enabled should be obvious.

View on HN

I've read this three times now and still don't understand you. Do you think people don't update their software? Are you just counting every update as people happy about the kill switch?

There'll be so much noise in that signal it'll be almost useless. You can't differentiate it from anything else. For all anyone knows it happened because the word Firefox was in the top story on hacker news

Approximately all people update Firefox so you don't need telemetry to count "AI disabled" installations, instead you can derive it as "updates requested" minus "AI enabled".

But the AI and telemetry toggles aren't coupled. While we can assume that users who disable telemetry are more likely to have AI disabled as well this still isn't translating to user numbers.

Hm. I don't think I follow "isn't translating to user numbers", could you elaborate?

Here's my thinking: There's 100 users getting updates. There's 40 users sending telemetry with AI enabled There's 10 users sending telemetry with AI disabled

So we have 50 people not sending telemetry and using or not using AI. If we assume more likely but not overwhelmingly more it's 30 people.

So we end up with 40+20 with AI, and 10+30 without?

This correlation may be ilegal to do in large parts of the world.

When you disable telemetry you are declaring you don't want to be tracked.

If you attach tracking to get exact set numbers it'd be illegal, but taking an aggregate and doing statistics shouldn't be since it can't be traced to individuals.

I assumed that AI features require server, so the server knows how many users it has. If some features are local then it's not so straightforward.

that's.... really clever

I can think if some mac users I know who deny all updates from being burned by feature creep and breakages in the past.

If my choice is between a single blatant signal of hostility to AI that can’t be misunderstood, and hoping that a pro-AI company’s executives invested in voluntarily correlating two different sets of logs to prove itself wrong, then I’m taking door number one.

Agreed, but we have laws about PII in so many places so this is what we'll probably get.

Still, AI services cost money they probably want to check usage since money talks.

Reply on news.ycombinator.com