Live data from Hacker News

As AI eats the web, the internet’s collective memory is disappearing

thewalrus.ca

931–940 of 989 posts

Re: As AI eats the web, the internet’s collective memory is disappearing

#931
post #285

Earlier quoted context omitted.

I, too, look back fondly on the early web. Lately, though, I have been thinking that one reason for its decline wasn't just corporate interests like we often talk about here. A significant number of early bloggers were middle-aged and elderly people. After decades, they simply aged out. The younger generation that replaced them (albeit not so much among OS nerds like yourself) was less likely to use a real computer a…

While we're here exchanging old man stories, I remember the web before blogs existed! I guess we can organize the internet into these eras, each of which was in some sense harder and more expensive to find available info than in the previous: 1. Pre-web. Internet is mostly about messages sent to individuals or groups. USENET organizes group discussion into browseable topic-oriented hierarchies, IRC does the same but…

Social media sites actually want search engines to index them. It's just much more computationally intensive and thus hard to cheaply to index a photo on Instagram or a video on TikTok or a podcast on Spotify than plain text. So search engines decide not to do it. Only google indexes YouTube transcriptions

This shift to hard to cheaply index content is maybe why search engines died. People don't use hyperlinks anymore

There's also a psychological aspect to relating faces to information. It's much more engaging and attractive. Content creation is still alive and more popular than ever but text based content creation brings no clicks and creates no interest. There's something about video that makes it very appealing like feeling you are outside and thus needing to be alert

Re: As AI eats the web, the internet’s collective memory is disappearing

#932

Earlier quoted context omitted.

Most hobby communities in my spheres of interest have been subsumed into Discord. It is yet another silo, but it does feel lively. I don't think I'm contradicting you here, I just feel less negatively about it.

I think Discord is one of the worst things to happen to the web. So much useful, interesting content gets walled off and locked away. It's not public, not searchable, not linkable, not indexable. I'm never going to stumble across an interesting tidbit of information on Discord while browsing the web. And even when you do have access to a particular server, you'll often struggle to find something posted a while back,…

And if you try to index it, people get mad at you saying you are publishing their private conversations online.

Re: As AI eats the web, the internet’s collective memory is disappearing

#933
post #892

Earlier quoted context omitted.

I think Discord is one of the worst things to happen to the web. So much useful, interesting content gets walled off and locked away. It's not public, not searchable, not linkable, not indexable. I'm never going to stumble across an interesting tidbit of information on Discord while browsing the web. And even when you do have access to a particular server, you'll often struggle to find something posted a while back,…

Yes, I share your hate. Unfortunately, the open web is apparently too hostile now for any sizeable open community.

Discord sold their product like a YC backed startup. Forums don't do that.

Re: As AI eats the web, the internet’s collective memory is disappearing

#934

Earlier quoted context omitted.

If ChatGPT goes bust, they'll find another ChatGPT, not another Google. I wouldn't be surprised if we'll see a popular competitor from China in the years to come. TikTok already took social media by storm, something nobody thought that was possible either.

> I wouldn't be surprised if we'll see a popular competitor from China in the years to come. TikTok already took social media by storm, something nobody thought that was possible either. Tiktok is cheap to run. AI however, it needs absurd amounts of power, RAM and GPU compute capacity to run... and there's serious constraints everywhere.

TikTok and every large social media site needs absurd amounts of AI to run the recommendation algorithm for every user, moderate content, TikTok also uses it to recommend music and their companion app, capcut, also uses it for video effects which are widely used. Bytedance, their owner, has over 10000 Nvidia GPUs for this reason.

Re: As AI eats the web, the internet’s collective memory is disappearing

#935
post #909

Earlier quoted context omitted.

Worker newspapers in the US had circulations as high as 10s of thousands to almost a million during the height of national strikes during the new deal (nearly several every month). The point is you are letting a system designed by elites to uphold elites set the boundaries. There are many forms of alternative media out there, some more popular than any billionaire owned rag. Just expand your definition and don't let…

*bangs head on desk* You're reading a description and hearing endorsement.

I really don't know what you are saying then because US labor movements did not distribute just pamphlets, there were truly alternative media sources out there. You're forgetting that there was a real communist, socialist, and anarchist movements for almost entire lifetimes before they were all systematically destroyed after WW2 with the help of powerful organizations like the CIO.

Add in the red scare + knee capping more leftists + academics during the 60s, 70s and you have a once massively influential independent confederate type of worker lead organizations that now amount to some moderate pundit's substack where they traded advocacy for ad dollars.

But before all this there were dual power institutions and mutual aid even going as far as providing free public education to all (who do you think taught cattle ranchers how to read in the 1880s? Then look at how quickly they took up leftist literature then writing their own).

What are you truly saying here because it just sounds like you don't know the history of labor in the US, which no shit. They don't teach this stuff anywhere, you have to purposely seek it out because neoliberals consider this type of stuff heretical.

Re: As AI eats the web, the internet’s collective memory is disappearing

#936
post #179
post #160

Earlier quoted context omitted.

So there's a new Internet Archive, it's just split across 3,000 AI labs.

I wonder if the AI labs will throw out their own copies of the Internet Archive after it's been sued out of existence. Probably not.

Yes, yes they will. When money runs out, just hit delete on their aws account to immediately stop all billing.

Re: As AI eats the web, the internet’s collective memory is disappearing

#937
post #666

My sister, a journalist, mentioned to me that she only uses google search because she had learned how to get information typically only Google indexed in the country she lives in, in a way it was not exposed on chat bots. She often has to search for information like Old govt forms released as public record with a fixed a certain format photo scanned into a pdf and indexed by Google were often on the second page of th…

This is a conundrum I always found interesting. If you know how to use a search engine (i.e. knowing how to use operators and structure a search query), you're almost always able to find what you're looking for very quickly, and in most cases (well, before SEO), the results are high-quality. You'll spend the same amount of time trying to fact-check an LLM (since you'll likely skim the articles it used in generating i…

What are these ways of using search that yield nom-garbage?

Re: As AI eats the web, the internet’s collective memory is disappearing

#938

There's too much going on in this article and while I think some of the points are valid, others go too far. A lot of the "cultural record" the author refers to is just digital junk. Random digital content that very few people care about, if we're being honest. Trying to hoard every bit of digital information ever produced is not the same thing as preserving "culture". Case in point: > Even the increasing use of ephe…

The mayor of Panama city, Mayer Mizrachi, almost exclusively only posts city updates as Instagram stories on his personal account. Many other politicians worldwide do the same. Many neighbor groups worldwide do the same. Many Town Halls and businesses worldwide do the same and that's because stories are actually seen by your followers unlike Instagram posts most of the time. People like to see a curated feed of the people they follow and stories are exactly that.

I think that's the kind of cultural record the author refers to. The average person has no way to post a screenshot let alone have it indexed by a search engine.

Re: As AI eats the web, the internet’s collective memory is disappearing

#939

Earlier quoted context omitted.

Personally it's because i used it when it was brand new and work paid for it and i had no idea what model it was using, my boss just turned it on for automatic PR summaries and code reviews and it was universally dogshit, and the auto complete in my IDE was awful as well.

Well it's much better now.

Well the PR summaries are still dogshit lol

Re: As AI eats the web, the internet’s collective memory is disappearing

#940
post #616

Earlier quoted context omitted.

I'm glad it went well for you. But I run a not-for-profit pinball museum, and I have specifically had to ban volunteer repairpeople from using the chatbots because it will confidently propose idiotically wrong solutions. Which will then get proposed to me, or just directly applied, with equal confidence. At this point I'd rather hear, "My horoscope said..." than "ChatGPT said..." I expect that it's good for common us…

Two years ago it would have been insane to say that you got help from ChatGPT to fix your outdoor plumbing, people here would have been frothing at the mouth for merely suggesting it. Four years ago it wasn't even on the radar of future possibilities. The quality it is today, is the worst it's ever going to be.

It's still insane to do that today. Using any LLM to get info for things is a terrible idea, because they will slip inaccurate info into the mix and you have no way of knowing where it is.
Post reply on HN