Three years ago, as OpenAI’s ChatGPT was making its splashy debut, a Pew Research center survey found that nearly one in five Americans saw AI as a benefit rather than a threat. But by 2025, 43 percent of U.S. adults now believe AI is more likely to harm them than help them in the future, according to Pew.
1 in 5 people seeing something as positive is not a high approval rating in the beginning.
I mean 4 out of 5 Americans probably held the opinion:
What began in 2022 as broad optimism about the power of generative AI to make peoples’ lives easier has instead shifted toward a sense of deep cynicism that the technology being heralded as a game changer is, in fact, only changing the game for the richest technologists in Silicon Valley who are benefiting from what appears to be an almost endless supply of money to build their various AI projects — many of which don’t appear to solve any actual problems.
many of which don’t appear to solve any actual problems.
That’s putting it lightly. If only the issue was merely not having sufficient use cases, rather than actively making lives worse through environmental strains, supply chain hoarding, and misinformation.
It’s good to see the sentiment growing. Anecdotally, there are non-technical people in my circle that use LLMs frequently as search engine replacements or to do stupid shit like generate pictures and emojis. I hope that begins to decline with the general sentiment called out in this article.
The sheer number of useless LLM integrations in every website, every mobile app, and hell, even smart TVs is insane. I feel like it’s causing people very real feature fatigue. And all of the Internet content and advertising slop is making the takeover seem so much worse.
Edit: Grammar, formatting
I react with either neutral apathy, disgust or surprise when somebody tries to show me their latest AI generated blob. Repeat twice and they stop using it. Our fear of social embarassment is higher than our desire to use AI.
“Look at this picture of me in a Ghibli style I generated”
“Oh… It’s kinda bad isn’t it? I’d avoid sharing it”
“Oh remember what we were debating earlier? Gemini said that…”
“Oh I know what you’re going to say, it said something totally dumb, right? I know, one must be very stupid to trust it haha so anyway what were you saying?”
Search engine replacement is probably the only use case of AI for me - for the times when I don’t know exactly what I’m searching for so the conversation style is helpful.
There’s times when I want to find “exact matches and nothing but” - searching for error messages, for instance - and that’s made much harder than it should be by AI bullshit search engines that don’t want you to switch off their “helpful” features. Considering moving to Kagi instead.
Definitely, there is using AI when you want and then there is having AI forced down your throat
For horribly inaccurate results that sound like they were written by a $5 SEO article writer.
Ill stick with key word search and skipping over all the SEO crap for.real results.
With the amount of AI generated slop everywhere, I’m afraid that’s becoming less and less effective. For what it’s worth, LLMs work fairly well in filtering it out in a first pass. I don’t take what it spits out directly but uses the links it cites as sources and find my answers there.
Is it the conversational style? or that search engines have been designed to be actively worse to keep your eyeballs spending more time looking at advertisements now?
I used to be good at googling information I needed. Then Google changed what googling does.
Don’t use Google. DDG works with keyword searches and you get exactly what you expect.
I do use DDG as my default. I was using “googling” for the sake of making a witty remark. Even so, the results are usually comparable - the first couple dozen results are mostly AI/SEO slop.
Oh, go to your DDG options and turn the AI slop off.
When I was in grade school we had multiple lessons on using Boolean search terms to find useful information. Google-fu used to mean something.
When I was in grade school we loaded up Logo and wrote scripts to make the turtle do abstract art.
is old
>_>
Try to find a big rig truck wash near your location on Google maps. I cannot. Only 57 million car washes and no way to filter them out.
Edit: lo and behold: “big rig truck wash” is the magic phrase it turns out. “Truck wash”, “Semi truck wash”, “semi truck wash -car” don’t work but “big rig truck wash” does.
Fuck you Google, you waste my time.
That’s because pickups and bigrigs are the same things now.
Not OP but - It’s both. Sometimes I’m searching for something very specific and Google just refuses to give me what I need, while serving tonnes of promoted bullshit. Copilot will then (sometimes) give me the results I need much faster.
Other times I have a vague idea of what I’m looking for and an AI like Copilot can narrow it down for me or sometimes just flat out give me the result I need.
Of course, that only works if it doesn’t hallucinate bullshit answers, which it does regularly. Still, with how far Google has fallen, sometimes it’s just faster to go through Copilot and sift out hallucinations anyway.
When I read this crap, all I can think is that yeah backlash is growing because the forced implementation is growing. Another useless sentiment-based article.
Lets use LLMs for things LLMs are useful for. It is not a panacea, and it is not appropriate for every use case
Yeah, LLMs are interesting tech products to play with and find some niche uses for.
But for the love of god they are not “prop up the entire stock market and numerous multi-trillion-dollar companies indefinitely” good!
