This is so wasteful it boggles the mind.

Now we get to see how OpenAI will do similar harm, more effectively and at much greater environmental cost.

I feel like we have an opportunity to break a feedback loop. SEO worked because of links. No links in my chatgpt discussions. Ah, but what about all the junk ai gen content you may counter, but that stuff only works because of SEO links. As more people abandon searching for links for discussions, the SEO usefulness diminishes. Maybe to the point where many parasites stop making SEO shite in the first place.

But to your points. I think the problem with your analysis is that it forgets that the real driver of the junk is the advertising environment. The SEO links were profitable because the advertisers were willing to pay a few cents for space on those pages. Yes, the incentives are changing for the teenagers who are churning out text and adding seo links to their stable of cheap websites, but the advertisers are going to find a way to manipulate consumers that's compatible with the new order. I don't know what that will be, but whatever it is will depend on information pollution just as much as the current one.

I wasn't even talking about the information pollution -- I was talking about boiling the ocean. Not that I'm not concerned about the former.

The costs are too great. We can't go on like this.

That's not the web

Why not?

I'll take a stab at it. What is the web, really? Gotta be stuff you see and interact with in web browsers right? Sure, you can get to HN, YouTube and Instagram in a browser. But by traffic for example, how much of activity on HN, Instagram and YouTube combined is through a browser? I mean, gotta be pretty low...like 5%? Just a guess, but remember all the app usage and TV usage for those sites is pretty big.

So if 95% of traffic/users/whatever metric are not using a web browser for those activities, is it really the web? It can't be called the web just 'cause they use HTTPS. It's gotta be a 'world wide web' experience, which I think a good proxy for would be using a web browser.

I got no horse in this race, just thinking out loud about it.

“The web” is, by definition, a collection of things that are loosely connected and accessible (searchable, etc). While the current internet is still “a web”, it’s mostly a web of 2-3 massive properties, entirely operated by the same 2-3 companies, completely devoided of public apis (and sometimes even web accessible content). The fact that sometimes they have an html version makes them “websites”, I guess, but not really a “web of nodes” the same way it used to be

Another common phenomenon these days is that lots of businesses don’t even bother having a web presence - it’s all instagram, WhatsApp and tiktok accounts, mostly only accessible via apps (or worse, chat platforms like discord)

What other way is there to access HN but the website?

Agree YouTube and Instagram are probably mostly apps which puts them in the “Internet” category but not “world wide web”.

Technically there is an API and there are some client apps for HN. No way they make up 95% of traffic though.

WWW != Internet, yes. I would count mobile apps as part of the web too, they're simply another sort of "browser."

I think the idea is that those are websites on the web, which are distinct from the web itself.

There are several meaningful difference between surfing Youtube and surfing the web. These include ownership, access, review, exposure, and more.

the web was the clicking of links from site to site. the interconnectedness of information. Searching the web was the start of browsing but it was not the only means of traversing.

Honestly the web died long ago imo. Wikipedia and other wikis are the only places that feel like the old web to me now.

Something for nonequilibrium statistical mechanics

More to the point, photoemission spectroscopy has been a workhorse tool for understanding the electronic properties of materials for quite a long time now (though perhaps not yet in 1921).

I would add to this that it had the advantage of something like 40 years of history as a field that was the basis for some of the biggest advances in instrumentation of that era.

ndiswrapper was a big learning moment for me as well

I'm on a framework 13 AMD edition since early this year, running arch. There were issues early on but seemed to be mostly on the firmware side and would not have been distro specific. After then April(?) firmware updates I have not had any issues.

Is .8 or .9 considered good enough accuracy for something as simple as this?

I'd say how much is good enough highly depends on your use case. For something that still has to be reviewed by a human, I think even .7 is great; if you're planning to automate processes end-to-end, I'd aim for higher than .95

Well, when "simply" extracting the core text of an article is a task where most solutions (rule-based, visual, traditional classifiers and LLMs) rarely score above 0.8 in precision on datasets with a variety of websites and / or multilingual pages, I would consider that not too bad.

Yes, because the prompt is simple as well.

Chain of thought or some similar strategies (I hate that they have their own name and like a paper and authors, lol) can help you push that 0.9 to a 0.95-0.99.

I found that strange. Normally 2 significant digits are used, e.g. 74%