Hacker Newsnew | past | comments | ask | show | jobs | submit | tsumnia's commentslogin

Here's one from my old AOL days. We originally just used email, AOL message boards, and a scheduled weekly chatroom, but once the Web took over it merged into Starmen.net [2]

Since we're on a "reminsence about legacy Internet" trend right now, here's the opening to [1]:

"What most people forget to remember is that it’s not just about the game. It’s about the people, it’s about the newsletters, it’s about the discussions, the trivia, the polls, the websites, and the meetings. Everything that was a part of the club was a part of the community, and there was so much involved that it was almost too much to handle. Who had the time to be a member of some 15 Online clubs? I can distinctly remember sending out invitations to join Moonside and receiving replies along the lines of “Sorry, I’m already in like 5 of these things.” Now, I wish there were more clubs and to any of you who have one: I will readily join. The only last great, recently active club I can think of now is the EarthBound Gang, arguably the greatest Online EarthBound Club ever. In early 1999, a lot of the clubs started dying out. I know that mine began to slow down, only to be restarted in the fall of 99’, and again in the summer of 00’. But as a whole, the EB clubs were never restarted, which is a shame, because some of them were downright fun."

[1] https://www.angelfire.com/ga3/ebhistory/intro.html

[2] https://starmen.net/


The problem is "fandoms" as a whole have now become such toxic hellscapes I'd rather just enjoy the game/movie/TV series myself and completely ignore what anyone else's opinion of it is.

I don't need things that bring me joy to be ruined by the most obsessive weirdos in the world.


That's easily solved by turning off comments or forcing every comment to be approved by an admin before it was displayed; much easier back then as there simply weren't that many people online. The "Web 2.0 read/write web" of XBL lobbies, Battle.net and Discord perverted gaming culture to a point of no return.

Admins are likely to also be the "obsessive weirdos".

Exhibit A: Reddit.

Sites back then had benevolent dictators that curated an experience for fans. I think in many ways it worked better than the democratized communities we have today.

It wasn’t so much “benevolent dictator” as it was “you’re in my house, so quit being a dick”. Toxic fans certainly existed but this approach usually led to them splintering off to create their own thing that’d inevitably wither away.

(You can perhaps substitute “wasn’t so much” for some form of “in addition to”)


All fun and games until the one person running the site crashes out and rage quits, taking the site down with it.

The thing was you had so many sites that losing one or two was sad, but not hugely destructive.

Honestly that's just another thing that's down to algorithmic social media. Echo chambers + rewarding ragebait is an explosive combination.

While I agree with current "Top-Level Internet" as I call it, the blurb is referring to the '99-'00s era when we were still very much disconnected. Those clubs had a forum on AOL and that was it.

I think Discord is where the "New Internet" is forming, because that's where this generation of kids are hanging out. We were in the Nintendo chat rooms, and they're in the modern day equivalents. We just think they are on Twitch and Kick because that's where the grown ups are playing games, but remember there's a reason Roblox is popular.

Kids don't want to hang out with grown ups.


I think <map> should make a comeback [1]. Then we could go back to having sick looking art pages

[1] https://www.w3schools.com/tags/tag_map.asp


I remember having to do an image map as part of a web design project in school, circa 2005. True to 2005, the image was one simple rectangular banner at the top of the page, with text for different pages like About and Portfolio etc. placed along it. It felt like a (cool) hack that you could define regions for the text to route to different pages on click.

Of course we have modern solutions [0] nowadays but that sure seemed cool 20 years ago!

0: https://droughtmonitor.unl.edu/CurrentMap/StateDroughtMonito...


Map wasn’t really the thing that enabled all of those designs, even though it was certainly often used to build them.

We just didn’t have a million and one layouts and device sizes to handle back then and so you could get really creative with available space. Even CSS Zen Garden later on had designs that worked much better on the limited screen sizes of that era - which don’t work well today.

Flat design trends killed off the rest of it I think.


I mean, its still in the spec. Go for it!

Likewise, we've always liked to "name things". My personal desktop is named "Foundation" (first built PC) and my car is named "Big Boi" (first adult purchase). Generic names are fine for operational equipment, but no one wants to refer to natural disasters as "HURR-2026-EC02". That's why COVID is "COVID" instead of "severe acute respiratory syndrome coronavirus 2".

May I interest you in my doctorsensei.com/how-to-get-off-the-internet.html page?

tl;dr Use dark mode and set f.lux (or the equivalent) to Cave Painting. Helped me out a lot


Now that's a template I haven't seen in a long time! Thanks for the fun trip down memory lane that "started it all".

Fly some drones, maybe thermal image wildlife


> You're just arguing about semantics. It doesn't matter in any substantial way.

While I agree for many general aspects of LLMs, I do disagree in terms of some of the meta-terms used when describing LLM behavior. For example, the idea that AI has "bias" is problematic because neural networks literally have a variable called "bias", thus of course AI will always have "bias". Plus, a biases AI is literally the purpose behind classification algorithms.

But these terms, "bias" and "hallucinations", are co-opted to spin a narrative of no longer trusting AI.

How in the world did creating an overly confident chatbot completely 180 years of AI progress and sentiment?


Terminology sucks. There is an ML technique called "hallucinating", that can really improve results. It works, for example, on Alphafold, and allows you to reverse the function of Alphafold (instead of finding the fold that matches a given protein or protein complex, find a protein complex that has a specific shape, or fits on a specific shape).

It's called hallucination because it works by imagining you have the solution and then learning what the input needs to be to get that solution. Treat the input or the output as weights and learn an input that fits an output or vice-versa instead of the network. Fix what the network sees as the "real world" to match what "what you already knew", just like a hallucinating human does.

You can imagine how hard it is to find papers on this technique nowadays.


Now that we've reached the end of the Internet, enjoy instructions on how to get OFF the Internet.

https://doctorsensei.com/how-to-get-off-the-internet.html


"An information flow model for conflict and fission in small groups (1977)" by Wayne W. Zachary [https://www.jstor.org/stable/pdf/3629752.pdf]

I know this paper isn't about social networks, but we know this, we knew it in the 70s. The only difference is that we continue to ignore and forget it.


Unfortunately "ethics" can also be a sneaky way to insert "politics" into the curriculum.


I've never seen this. Is this some weird right wing talking point?

I have seen a disconnect between what is covered in ethics classes and the types of scenarios students will encounter in the working world. My (one) ethics class was useless. But not political even with the redrawn ethical map of the Trump era.


Building systems that don't bake bias into code or worrying about privacy in a dating app is probably the kind of politics the parent is talking about.

I'm not really sure how you could totally separate politics (forming laws) from ethics anyway.


I linked my lecture about how I see "ethics" as another term for "politics" in the academic sense.

Other points to note: All AI will always have bias; a) because neural networks literally have a constant called 'bias' in them; and b) training a prediction algorithm means it is being trained to be bias and running a clustering algorithm means lumping commonly themed attributes together.

If you mean "BIAS" as in "RACE/ETHNICITY/SOCIOECONOMIC_STATUS", then most groups already do this, state they do it, and still deal with the general public not believing them.


Reread my last sentence - allow me to rephrase it:

If you mean "BIAS" as in avoiding "RACE/ETHNICITY/SOCIOECONOMIC_STATUS", most groups state they do not use them for decision making and still deal with the "general public" not believing them because results do not support their opinions.


Feel free to review my Ethics in AI lecture: https://www.youtube.com/watch?v=XBJRDxwbI9Y

This isn't a "Trump Era"/"Right Wing Talking Point" amigo, its the truth; also it's always been this way, listen to some punk rock. Is AI allowed to generate child pornography? Because that is currently being argued as protected by Free Speech in courts because no child was actually involved [1]. Sorry your ethics class was useless, I don't believe mine is.

[1] https://www.usatoday.com/story/news/nation/2024/11/21/pennsy...


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: