Has anyone tried making a p2p search engine? Naively it seems like one of the p2p networks from the late 90s or early 2000s should be able to handle it. Searches propagate between nodes, nodes specialize on specific sections of the index, which can be random + user preference (maximizing redundance and availability for the things people actually care about). Soulseek already gets like halfway there and it's not even…
What we lost when search stopped making us think
41–50 of 64 posts
Re: What we lost when search stopped making us think
#42I do worry about Google AI overview's long-term impact, but man, it has been such a great experience of late. It surely does feel like the old google but better. There was a point, perhaps about five years ago, where using Google had become a genuinely frustrating hurdle I needed to overcome to do the actual work. Now the frustration is completely gone.
Interesting given this is the polar opposite of my experience. I find Googles AI overview to give wrong answers constantly. Occasionally, even for questions I believe Google circa 2018 was able to answer confidently. Ironically, Googles AI overview often links to references that state the polar opposite of what the AI overview claims.
Re: What we lost when search stopped making us think
#43The AI thing is poor as well, because it makes things up and changes context all the time.
Re: What we lost when search stopped making us think
#44Re: What we lost when search stopped making us think
#45> This isn't really anyone's fault in a simple sense Or, in a more simple sense it’s entirely Google’s fault
Google's fault that SEO exists? The incentive isn't even there, they want people to pay for ads rather than gaming their results. But SEO finds a way.
I bet a large number of them run Google ads.
As such, it is not in Google's best interests to cut them out of search, regardless of how low quality they are.
It is well known that Google and SEO have a wink wink / nudge nudge relationship where Google "hides" its search algorythm, and somehow the websites with the most ads get closer to the top every year.
Re: What we lost when search stopped making us think
#46I spent an hour debugging my networking setup and I just turned on voice mode and used ChatGPT to help me figure out why everything was working except my networking equipment wasn't changing IP addresses when I set it. As it answered my questions, it gave thoughtful answers and explained everything along the way to the point where I started to guess that I needed to upgrade my control plane, and lo and behold, that was the answer. If I had to peruse through links just for the false idea that I would be getting a cognitive benefit, I might have spent 4x the amount of time or given up.
I'm very happy that we are living in an where we can get reliable answers immediately and move on to what we want to do.
Re: What we lost when search stopped making us think
#47I was still digesting the throat-clearing intro when my eye noticed "quietly" and "honest" and I decided I'd not bother. But I scrolled through out of politeness, just in case the Basilisk will be very thorough. > I want to be clear that I'm not against these tools generally. I use them myself No essay is made worse by being stripped of filler. If the reader has got this far, this sentence is entirely redundant, I th…
Re: What we lost when search stopped making us think
#48Re: What we lost when search stopped making us think
#49There is nothing noble about search over and above AI answers. It's just an outdated way of delivering answers but massively impaired because the technology didn't fully exist. Imagine you went to your doctor to ask a medical question and she gave you a list of books to research and to formulate your answer that way. I spent an hour debugging my networking setup and I just turned on voice mode and used ChatGPT to hel…
So for now I'm doing both. A bit search and a bit AI.
Re: What we lost when search stopped making us think
#50There is nothing noble about search over and above AI answers. It's just an outdated way of delivering answers but massively impaired because the technology didn't fully exist. Imagine you went to your doctor to ask a medical question and she gave you a list of books to research and to formulate your answer that way. I spent an hour debugging my networking setup and I just turned on voice mode and used ChatGPT to hel…
An extreme example, but what happens if Bob Nottechnical follows GPT instructions which happen to include `sudo rm -rf /`? How would Bob know to stop trusting the instructions when every time he's followed things before it worked as expected?