Linux didn't just eat 10% of Windows market share, AI bots are inflating the numbers
https://www.windowslatest.com/2026/08/03/linux-didnt-just-eat-10-of-windows-market-share-ai-bots-are-inflating-the-numbers/?=0
43 points · 15 comments · view on lemmy.world
15 Comments
user_6282638282@sopuli.xyz · 85 pts · 32d
Finally some unbiased Linux reporting from... windowslatest dot com.
Rhaedas@fedia.io · 24 pts · 32d
It's a fair article. Mentions that while it may not be the year of Linux (whatever that ever means every time it's claimed), it does suggest the bot activity validates the death of the internet as we knew it. It also says that Linux is a great platform and they use it for certain things. If the adjusted amount it right, 4.7% is still pretty impressive, as I can remember people celebrating like 1% and 2% once. And for some reason I thought WIndows would be higher... they are slipping.
ViatorOmnium@piefed.social · 18 pts · 32d
Are AI crawlers really reporting as Linux though? It would make more sense for them to report Windows or Mac because sites will be less willing to start blocking them based on user agent.
halcyoncmdr@piefed.social · 10 pts · 32d
It's very likely default User Agent strings pull from the system they run on.
1984@lemmy.today · 2 pts · 31d
Yep. Almost all cloud machines are Linux. Does not mean that desktop market is.
vanillama@programming.dev · 7 pts · 32d
Might be a vibecoded docker container or something, but I agree with your skepticism here
Dadifer@lemmy.world · 32 pts · 32d
This article is based on this guys opinion and very little objective evidence
mesamunefire@piefed.social · 13 pts · 32d
Most of it is. Theres a good portion on cloudflare. Linux its over 4% on desktop marketshare when filtering. Which sounds correct. But the rest of the article is just as you say.
I wonder if cloudflare includes chrome os, specialty linux distros, ect... are pulling the numbers one way or another?
Seimhe@lemmy.world · -1 pts · 32d
He seems to have better evidence than the original claim did, though. He quotes statcounter’s own method of producing their figures, and what they actually mean.
He points out a very odd anomaly with OSX vs macOS numbers too that highlight the problem even more.
halcyoncmdr@piefed.social · 14 pts · 32d
Is it an anomaly? His point is that "OS X" refers to macs pre-2016 and "macOS" as the newer devices. But that's not accurate, at all.
Someone in another thread elsewhere said:
I can't confirm those results, in fact, I can't confirm the "macOS" bit at all. On my Macbook Neo (obviously a current modern Mac, running macOS Tahoe 26.5.2), I get these results:
Notably The Macbook Neo is not an Intel Mac, and it is not running version 15.7 which was the previous version macOS Sequoia, but that is what it reports to websites. It should report as "macOS" according to that one dude, yet it doesn't in either Librewolf or Safari (I don't have Chrome installed, and I'm not going to just for this shit). The User Agent is actively reporting to websites that it is an older device, even in Safari which does also mention the actual current OS version at near the end.
So if they are relying solely on User Agent reporting, it's obviously not accurate at a granular level. But at a general-OS level it could be depending on exactly how they're handling it.
Either way, simply saying "macOS" and "OS X" reporting refers to old and new devices... is plain wrong. And if he's wrong on such a basic assumption right at the top of his opinion piece... why would I trust any of the rest?
Seimhe@lemmy.world · 4 pts · 32d
Excellent points. I hadn’t thought of that.
randamumaki@lemmy.blahaj.zone · 9 pts · 32d
M$ fanboys say what?
ianhclark510@lemmy.blahaj.zone · 3 pts · 32d
Just think of how awesome it’s going to be when we actually make it
Dookieman12@piefed.social · 2 pts · 32d
Called it
naught101@lemmy.world · 0 pts · 32d
Wow, doubly depressing