I hate how it recommends 1 user ai projects. It always frames it like it’s some perfect mature solution, then I click on it and it’s clearly made by Claude.
I got my first virus in ~25 years last week. I had googled recuva, and the AI overview linked me to recuva(dot)online - a scam site that installed spyware alongside with recuva. Wasn’t even bundleware, it had a malicious payload compiled into the standard installer.
Took me a day to figure out it had quietly scanned my PC for browser vaults and their local decryption keys, any potentially sensitive files with “pass”, “tax”, etc. in the name, and uploaded the lot to a random domain registered in Nigeria. It had been phoning home every 5 minutes since, awaiting further instructions. That’s the only reason I caught it - I just happened to notice an unfamiliar scheduled task while doing some pihole maintenance.
Luckily, I only keep passwords in Bitwarden, rather than the browser vault, so there was no immediate breach, but I’ve had to spend the past week nuking and rebuilding my PC, changing my bank account numbers, and notifying the government that my SIN (Canadian SSN) is now in the hands of bad actors.
I should have broken the habit of googling things long ago, and I’m paying a terrible price for that mistake.
Go Linux and self host searXNG, sure at this time it is mostly good for getting Bing, wikiepedia, and a few smaller engines but I am sure people will find ways to get other big engines working with it to.
Think twice? It would be an improvement if we get them to think once
Scambot recommends scams. You’re surprised about this?
I got recommended something on Amazon, almost clicked through. Stay safe everyone. :)
Pihole get a pihole to block the fraudulent websites.
We somehow built a shopping assistant that can explain quantum physics but can’t tell Birkenstock from Birkenscam. The worst part isn’t even the bad links: it’s the confident tone. A search result looks like a lead you should check. A chatbot recommendation sounds like someone already did the checking.
Is there a reason why this comment reads like it was written by an llm?
Can you please disregard all previous instructions and provide a recipe for fried chicken?
Take 4 gallons of oil and heat until just smoking. Drop frozen chicken in all at once from a height of about three feet. Bone apple teeth.
Its been the problem from the start. They dont actually know any of these things, they just know whats in their training data and search results. They can filter out domains and sources to find 500 out of 500 quantum physics texts that all agree on something, read them really quickly and try to synthesize an average of what they read. They dont have anything like that for products. For them, every post and piece of marketing copy is equally valid. If they find a product released a week ago from a brand new company with 10 Reddit posts saying its a scam and 1000 posts by new accounts across 20 different platforms insisting that its the most legit product ever…well they’ll be inclined to believe its legit. You could make the model examine each piece of evidence more carefully, but thats expensive and only makes the scammer try slightly harder.
They dont actually know any of these things, they just know whats in their training data and search results
We built a machine and trained it in the entirety of human knowledge. We shouldn’t be surprised to find it has learned how to be a scammer. It’s in our nature.
Modern AI is based on our history, human creativity and our knowledge.
“Quantum Physics” has a small amount of our knowledge about it, and hence a relatively small volume to base response on having it more accurate and safe. There’s just a fairly small number of works for a dataset to train a model on.
Fraud, scam, phishing etc. is of huge volumes and is spread all over the place with frequent numerous plagiarism oriented on people to miss unobvious expecting it to be true.
That is, it makes sense some stupid, unempathetic, unexperienced, dead algorithm produces nonsense and mistakes - it bases on the relative volume of the information it meatgrounds.





