That’s right. I didn’t put the question in the search bar for you to tell me the answer. That question was the first shoveling of dirt in a 6 hour rabbit hole that I was fixing to explore.
This is the way
This grew from ‘I don’t want Wikipedia, I’ll pay $150 a month to Encyclopaedia Brittannica for a set of 60 books the way God intended…’
Encarta FTW
Let them waste money on generating an AI overview I’ll happily skip on my way to the wikipedia link.
AI overview uses around 20 times more energy than regular searches.
It’s absolutely batshit crazy that most search engine put it on by default just for it to summarize a wiki page 90% of the time. It’s like we collectively pushing for a mass extinction.
I sure would hope that those AI overviews are cached. Querying an LLMs for every repeated search sounds very wasteful
If it didn’t do the AInsearch repeatedly they couldn’t use the billions of daily searches to artificially boost their AI engagement metrics and further inflate this bubble.
What’s even crazier is for a lot of things we already had a quick result overview. Like if you wanted the definition of a word. For a long time it was the case that you could just “define socks” and you’d get a widget at the top with the definition. Why would you replace that with AI, and have an AI process that request every single time someone wants to define socks instead of just keeping the old widget!?
Indeed.
I’m pretty sure a good percentage of the searches are also just people typing the name of the website they want to access in the search bar. They don’t want any information, just a link. All these searches now consume a massive amout of energy for basically nothing.
“Innovation” lol
I have it disabled through an extension on my phone’s Iceraven browser. Can’t bother reading those. I need the Wikipedia/website info.
Wikipedia has an app. It’s half of why I Google stuff. Usually a good place for quick answers, and an ok place to start digging in.
I’m half amazed at how bad Google has gotten, but that’s probably just me remembering the early days of web searches.
join the elite: kagi.com
A search engine you have to sign up for?
And pay, i’d not mind a sign up but i’d rather not pay, especially if i cannot even have the option to self host
Join the elite: create an account and pay $120/year to a private American company for search results.
That’s still full of AI, the filter quality you can see for yourself without an account for a limited amount of searches. If I search an image and the URLs on slop I see in the results with the slop filter enabled that literally contains keywords related to AI slop and KAGI doesn’t filter them out…well, I draw conclusions.
Don’t worry, they’re also super AI-pilled and known for randomly cancelling projects. But I’m sure the search is great or whatever
That and TVTropes of course
That website was such a timesink. Amazing.
I hate it when i need to google programming stuff because Google Gemini LOVES to make stuff up and takes up half a page trying to answer
I know the rankings change like every month, but isn’t Gemini one of the “dumber” ones? I’ve only tried it a few times and got annoyed, but it’s been a while.
Oh it is, it has the same issues as the first few iterations of Chat GPT if you ever used AI in the early days
It’s certainly no leader for software engineering tasks.
stop googling, start ducking.
is “ducking” an actual term now? Come to think of it, ducking doesn’t sound any sillier than googling
But check out this awesome ai recommendation I got when searching for Halloween 2026… (Fuck everything about this btw, no sarcasm)

Fucking hell, lol
Lmao, I guess my internal ad filter is still working because I had to look at that twice to notice it and wondered if there was a new Halloween movie coming out this year or something. Agreed, fuck that shit.
Ooo, they’re cramming in the ads now eh?
AI has never been 100% accurate to the documents it pretends to reference, so the fact that anybody could even consider reading it like it’s not pure gibberish is mind boggling to me.
I gave Claude a document to proofread today. It came back with an error. Clause said I used “told” instead of “tell.” The error ended up being Claude’s error, as I used “tell,” and it was just confused.
This was the first time I ever used Claude. Just wanted to see what it was like. It did apologize, in fairness. But yeah, it’s clearly fallible. I guess I don’t understand it well enough to get how it read the document incorrectly. I’m baffled.
It’s because it’s a statistical model that guesses the next word. You asked it to read a document and formulate a response. So, based on responses in the dataset to the documents in the dataset, it chose the words and phrases which were most likely to be in a response.
Statistically, the most likely result the machine could come up with, was that you had used the wrong word. Factually, you did not.
When the old new ways become the old ways and i cant keep up with the new new ways

Me after losing 6 days of testing broken theories from an ai search when I could have just spent an hour watching a video that a human made that was fully functional from the start.
That said: No doubt that human spent way longer discovering the technique and then making the video: this should also be considered as a reality that creating something good takes time. And humans are still better at creating scalable content just give them the proper R&D time ffs.
LLMs fundamentally can’t learn. They will continue to fail the same thing over and over until they guess and check into a response that appears coherent and correct.
There’s a site for that…let me remember the name… Wikipedia! There it is.











