With 68% of consumers reporting using AI to support their decision making, voice is making this easier. [1]
Does anybody actually believe that 68% of consumers use or even want Copilot? But they included a source for this very generous assertion at the bottom of the page:
[1] Based on Microsoft-commissioned online study of U.S. consumers ages 13 years of age or older conducted by Edelman DXI and Assembly, 1,000 participants, July 2025.
Oh yeah, that’s compelling: US consumers, 13 years old and older. An entire thousand of them!
So the only question I have left is which junior high principal Microsoft “compensated” for this survey, and what happened to the 320 summer school attendees who said fuck you, no anyway.
Yeah like we all use chatGPT for the most part now but that still does not mean copilot
Fun fact though out of topic: I once searched for 2 girls one cup in copilot, and though it said I cant talk about it, it provided sources and one of them was a link to the video
When google shoves their ai to the top of search results, its hard not to read it. I’ve been spoiled by ublock and I am no longer used to ignoring the first few things that come up.
I’ve been using Duckduckgo with uBlock for years, so I had no real problems with anything like the hell of Google “sponsored content” until Duckduckgo started putting up their own AI search assistant. Since then I’ve gone from start.duckduckgo.com to noai.duckduckgo.com because I got tired of turning their search assist off and couldn’t reliably block it with uBlock because they kept changing it. (I delete all cookies after every browser session and do not maintain individual app accounts, so their AI settings options were never gonna work for me.)
Because of the way my brain works, I literally don’t even want to see what AI says until I’ve done my own looking. Yet I never failed to turn it off, because I just can’t rely on it.
Usually when I’m looking for something I’m in a hurry, so it’s less trouble for me to just pick my own sources, preferably older than 2023 if possible, and read a bit myself than to spend time getting blithely lied to, or even just suspect hallucination/omission to the point that I think I need to verify it before I can rely on it.
It’s not an exaggeration to say that for me, it is literally faster to skim three or four completely different primary sources than it is to try to verify the assertions in a single search assist paragraph: one is just light reading, the other is point by point comparison of the AI offering against multiple independent sources. So I read.
I’ve never regretted summarizing a topic myself, but I’ve definitely gotten some rotten eggs from AI, both in blatant non-truths AND in holes of omission you could drive a truck through. I won’t make that mistake again. So for me, AI summaries are well worth staying wary of for now.
Maybe Marginalia could work for you? I’ve tried using it, but it’s a lot more focused on academic stuff (rather than figuring out song lyrics or which episode some TV quote came from). It’s an “old school” search engine, though, so a bit less convenient than google, duckduckgo, etc. if you weren’t around in 90s/early 00s for that.
Also just because you have used AI doesn’t mean its overly useful. Gone to ChatGPT multiple times to try getting information that Google now is too shit to provide, and ChatGPT ends up providing some stupid response that is clearly wrong.
Occasionally used ChatGPT to find a website to use as an actual source, but now those sources are also AI written bullshit that is clearly wrong. Which is increasingly concerning because while I know some things are wrong, I don’t know everything. How many other things that it points to are wrong? Its not too bad if you are able to verify it through non LLM sources, but what if you can’t?
It’s the newspaper (news site or app now) problem. You read the news from your venue of choice, taking it all in, sorting out how you feel about it, maybe pick a side on an issue. Then you turn the page and there it is. An article about your career field in A.
Wow, you might think, they got this so wrong. They clearly have no understanding of A. You might even get angry about it.
Does anybody actually believe that 68% of consumers use or even want Copilot? But they included a source for this very generous assertion at the bottom of the page:
Oh yeah, that’s compelling: US consumers, 13 years old and older. An entire thousand of them!
So the only question I have left is which junior high principal Microsoft “compensated” for this survey, and what happened to the 320 summer school attendees who said fuck you, no anyway.
Yeah like we all use chatGPT for the most part now but that still does not mean copilot
Fun fact though out of topic: I once searched for 2 girls one cup in copilot, and though it said I cant talk about it, it provided sources and one of them was a link to the video
I can’t judge you for that, I feel like you probably got the very best Copilot has to offer, lol
When google shoves their ai to the top of search results, its hard not to read it. I’ve been spoiled by ublock and I am no longer used to ignoring the first few things that come up.
I’ve been using Duckduckgo with uBlock for years, so I had no real problems with anything like the hell of Google “sponsored content” until Duckduckgo started putting up their own AI search assistant. Since then I’ve gone from start.duckduckgo.com to noai.duckduckgo.com because I got tired of turning their search assist off and couldn’t reliably block it with uBlock because they kept changing it. (I delete all cookies after every browser session and do not maintain individual app accounts, so their AI settings options were never gonna work for me.)
Because of the way my brain works, I literally don’t even want to see what AI says until I’ve done my own looking. Yet I never failed to turn it off, because I just can’t rely on it.
Usually when I’m looking for something I’m in a hurry, so it’s less trouble for me to just pick my own sources, preferably older than 2023 if possible, and read a bit myself than to spend time getting blithely lied to, or even just suspect hallucination/omission to the point that I think I need to verify it before I can rely on it.
It’s not an exaggeration to say that for me, it is literally faster to skim three or four completely different primary sources than it is to try to verify the assertions in a single search assist paragraph: one is just light reading, the other is point by point comparison of the AI offering against multiple independent sources. So I read.
I’ve never regretted summarizing a topic myself, but I’ve definitely gotten some rotten eggs from AI, both in blatant non-truths AND in holes of omission you could drive a truck through. I won’t make that mistake again. So for me, AI summaries are well worth staying wary of for now.
My favorite is when AI summary answers a question, then the links from the search below contradict that answer. It’s shit for biomedical research.
Maybe Marginalia could work for you? I’ve tried using it, but it’s a lot more focused on academic stuff (rather than figuring out song lyrics or which episode some TV quote came from). It’s an “old school” search engine, though, so a bit less convenient than google, duckduckgo, etc. if you weren’t around in 90s/early 00s for that.
They got that 68% usage number likely by counting everyone accidentally using it after a search swap or similar trick.
Also just because you have used AI doesn’t mean its overly useful. Gone to ChatGPT multiple times to try getting information that Google now is too shit to provide, and ChatGPT ends up providing some stupid response that is clearly wrong.
Occasionally used ChatGPT to find a website to use as an actual source, but now those sources are also AI written bullshit that is clearly wrong. Which is increasingly concerning because while I know some things are wrong, I don’t know everything. How many other things that it points to are wrong? Its not too bad if you are able to verify it through non LLM sources, but what if you can’t?
It’s the newspaper (news site or app now) problem. You read the news from your venue of choice, taking it all in, sorting out how you feel about it, maybe pick a side on an issue. Then you turn the page and there it is. An article about your career field in A.
Wow, you might think, they got this so wrong. They clearly have no understanding of A. You might even get angry about it.
And then you read the next article.
they are equating “AI support” with “I want AI copilot integrated into my OS”
and that’s a big leap