this post was submitted on 24 Apr 2024
1063 points (98.9% liked)
People Twitter
5283 readers
1826 users here now
People tweeting stuff. We allow tweets from anyone.
RULES:
- Mark NSFW content.
- No doxxing people.
- Must be a tweet or similar
- No bullying or international politcs
- Be excellent to each other.
founded 1 year ago
MODERATORS
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
This only makes sense for a very limited set of things tho. The things where AI makes a difference are art stuff, news and creative writing stuff. What purpose do news from before 2023 have? Maybe for research purposes it filters out some bullshit but then you also miss out anything relevant that happened after 2022.
I think you've got it backwards. This only excludes new information from the past two years. Literally everything else is majorly improved.
I believe they're arguing that AI is particularly used in news, and when looking up news, you're typically seeking current events, in which case excluding post 2023 content doesn't work.
In my opinion, the place I encounter AI content the most is in list content, not just clickbait lists but also stuff where multiple products are compared. If I'm looking up what laptop to get, AI articles pop up comparing 10 products with inaccurate and messy details, but also I don't want to see old products.
Also IMO, in many cases Google search has been useless for 6+ years now. I think it was around 2018 where I started ending my search terms in 'reddit' because the first few articles were poorly advised clickbait, especially when looking for any advice (Reddit of course went to shit anyway). Google is only useful now for navigating to popular sites that will inevitably float to the top of any search query due to popularity. The only other common use for me is correcting typos when autocorrect is stumped.
Not gonna spend a whole lot time arguing intent; you can see what was posted already.
Per unexpectedhazard above;
I'm saying that news is a very limited set of things that you look up on google, and literally everything else will improve with this trick. Just because a single user only spends their google time looking up news articles does not mean that everybody does that. It might be of limited use for unexpectedhazard, but frankly, unexpectedhazard is an exception, not a representative example of most people.
We all know google search has been shit for longer than two years, it's just that before 2023 you could at least eventually find what you were looking for. These days, you can't even do that.
Oops I misread your message. I interpreted you as saying that they had it backwards, meaning they thought it excluded everything before 2023.
Sadly, most of my searches need more recent answers. Software & hardware (when it comes to coding/tech/IT) and information about local businesses can change quickly and old information usually ends up being just wrong.
Google is really, really letting me down and wasting a lot of my time these days.
I get that people think that but I just googled a really long serial number for a washing machine with no other context and the first result was the repair manual. So at least in my experience it seems to be okay.
I agree that certain types of searches work, and that for many of them limiting your results to past content can be useful. Unfortunately, it doesn't help with the searches I tend to make. I edited my comment to clarify.
It also makes a difference for historical, medical knowledge and basically everything tha you can attach an affiliate link onto (e.g. product reviews), so... My guess is that covers a huge amount of searches.
Y'know, something you'd want up to date results for so you're not potentially buying older worse products.
It seems like every new product is worse than shit that was out 5 years ago anyway. Less functional, shorter lifespan, etc. Especially electronics.
I get what you're saying, but"older doesn't necessary mean "worse".
But it does mean that the reviews you based your decisions on are for old models (that may have sense been upgraded or changed) or discontinued products.
Yes, but if you listen to the 'reviews' now, all you're getting is the regurgitated product description compared to the product description of the other contenders. Except instead of the two sentences that were the original description, you get a five sentence opening about the topic, Somewhere between 12-15 sentences about the product that don't add anything relevant (repeat)^ (for each product), and then a conclusion that's random, arbitrary, and probably influenced by money.
At least with the older models I can see real information if I can find real reviews. If I have to extrapolate from there to today, I'm still better off.
I'm a computer guy, 90% of the things I'm gonna wanna see reviews for are ABSOLUTELY things where older means worse. A review of a year+ old cpu, for example, is absolutely meaningless to me, I wanna see how good the bleeding edge stuff performs.
I'm a "computer guy" as well and I don't nearly have the need for all he bleeding edge stuff. Especially for server, budget/used options, homelabs and open source software (where the drivers aren't optimized yet), older hardware is way enough for the job.
Sure, but I don't need reviews for that stuff. I would've already seen those when they were actually, y'know, new.
The bleeding edge stuff isn't stuff I personally need nor can afford. I just like learning about the newest and greatest tech.
Good for you. Not everyone is as close to the bleeding edge as you, tho.
Not news. Things like recipes have been put through the AI filter to add a lot of useless crap. Try looking up "how to roast a chicken" and getting useable results.
It's always someone's life story when all I want is a discussion of the cooking process. That content is hard to create. It's a lot easier to just have AI make up a fake life story.
That's why I've started using ai with web scraping abilities for my searches nowaday.
Find me how to roast a chicken, and make an instruction list
if I get decent results, double check the source site just to be safe and you're set. Basically fight fire with fire.
It sucks? Yes, adding another layer of complexity for filtering bullshit is not useful at all.