You're viewing a single thread.
Most Linux users disable Telemetry, so I think it's probably missing around 10%
For the fuck of it, I asked AI to guesstimate the market share
Based on the Mozilla Graphics Telemetry data provided, Linux accounts for 3.4% of the sampled Firefox sessions. However, this is likely an underestimation of the true Linux market share for the following reasons:
- Telemetry Opt-Out: A significant portion of users may disable telemetry for privacy reasons. It's reasonable to assume that Linux users, who often prioritize privacy and are more technically inclined, might be more likely to opt out of telemetry compared to Windows or macOS users.
- Pi-hole and Ad-blockers: Tools like Pi-hole and ad-blockers, which are also more commonly used by technically proficient users (a demographic that overlaps with Linux users), can block telemetry data from being sent.
To guesstimate the true market share, we need to consider a correction factor to account for these missing Linux users in the telemetry data. Unfortunately, precise data on telemetry disablement rates per OS is not readily available in the telemetry data itself or through a quick search.
However, we can make a reasoned guesstimate. Let's assume:
-
Baseline Telemetry Rate: Assume that for Windows and macOS, the telemetry data is relatively representative, and a smaller percentage of users disable it.
-
Higher Opt-Out on Linux: Assume a significantly higher percentage of Linux users disable or block telemetry. Let's arbitrarily assume that twice as many Linux users disable telemetry compared to the average across Windows and macOS. This is a rough guess, but it acknowledges the higher likelihood of Linux users opting out. Applying a Correction Factor (Illustrative Example): This is a simplified example for illustration. To be more accurate, we'd need real data on telemetry disablement rates.
-
Assume a hypothetical "average" telemetry disablement rate across Windows/macOS: Let's say, hypothetically, 10% of Windows/macOS users disable telemetry.
-
Double that for Linux: Based on our assumption, 20% of Linux users disable telemetry.
-
Calculate the "visible" Linux user proportion in telemetry: If 80% of Linux users enable telemetry (100% - 20% disablement), then the 3.4% in telemetry represents 80% of the actual Linux user base within the Firefox user population.
-
Extrapolate to the "true" Linux market share: To find the 100% (true) Linux market share, we can do a simple proportion: If 3.4% represents 80%, then X% represents 100%. X = (3.4% / 80%) * 100% = 4.25%
Therefore, based on this highly simplified and assumption-based guesstimate, the true market share of Linux might be closer to 4.25%. Important Considerations:
- This is a very rough estimate. The actual percentage of users disabling telemetry on each OS is unknown without specific data from Mozilla.
- Firefox User Base vs. Overall Market Share: This telemetry data is specific to Firefox users. The overall operating system market share might be different, as Firefox usage itself might be skewed towards certain operating systems.
- "True" Market Share Definition: "Market share" itself can be defined in various ways (desktop OS, all devices, etc.). This guesstimate is within the context of desktop/laptop users who might use Firefox.
In conclusion, while the telemetry data shows 3.4%, a more realistic guesstimate, considering telemetry disablement, could be in the range of 4% to 5%. To get a more accurate figure, one would need access to data on telemetry opt-out rates per operating system or look at alternative market share data sources that employ different methodologies (though these also have their own limitations).
Why post spam?
You're spam
I asked chatGPT if this was true and it just said that it had sex with your mom
The thing is, I seriously consider you to be spamming. I introduced a variable which is likely unaccounted for and came back later with something which I personally found interesting and insightful that I wanted to share. It doesn't advertise any services at all, not in being mentioned or any links and was posted as a sub comment to not cause issues at the top level. It simply adds further to the concept that I introduced. By very definition, not spam. Now whatever you or anyone feel about AI, asking it to speculate given the vast amounts of data it has access to, is what it's there for. On the other hand, posting just to say you don't like something doesn't add anything to any conversation other than making yourself feel good.
You posted something that is about as wanted or relevant as a random markov chain output, there's literally no world where pasting random LLM output into comments isn't spam.
The only thing that separates you and an actual spam bot is that you prepended "I asked an LLM about this."
spam in the place where i live
I'm not gonna be a dick about it, but you really shouldn't be asking AI (or at least not a language model, use a reasoning model) things like this and accepting what they say as fact.
The information these models have is out of date
The information these models give is inaccurate
You were taught how to research in school, do the math yourself instead of killing your critical thinking skills by letting an AI do the work for you.
I literally asked it to guesstimate and stated as such:
For the fuck of it, I asked AI to guesstimate the market share
It wasn't presented as fact. In fact I even provided the reasoning so if anyone wanted they could rip it apart, because I felt that makes more sense than just posting a conclusion without context
Don't listen to the haters but it would have been nice if you collapsed this because it is very long and generalized to the point that it is pretty much an eyesore. Plus most people can ask their AI of choosing semi-random topics. I don't see what was interesting in the AI response at all. It states some blatantly obvious facts and is rather too wordy. I intentionally include into the system prompt or "personalization" about how I like things to be kept short and to not reiterate what I had posted especially if it just sounds like the "AI" is thinking out loud.