mindcrime

Not sure if you're familiar with this or not, but that concept appears as part of the story in the novel[1] and movie[2] "Johnny Got His Gun". The premise is based on a badly injured soldier during WWI who has lost his ability to speak, see, hear, etc. He communicates by banging his head on his pillow in Morse Code. Many scenes from that movie are interspersed throughout the video to the song "One"[3] by Metallica.

[1] https://en.wikipedia.org/wiki/Johnny_Got_His_Gun

[2] https://en.wikipedia.org/wiki/Johnny_Got_His_Gun_(film)

[3] https://www.youtube.com/watch?v=WM8bTdBs-cw

See, to me, you're just leaning into the outdated and archaic aspect of this. Sure, that's the way encyclopedia's have been for a long time. But none of that means that they need to continue to work the same way. Technologies change, times change, standards change.

And that said... this whole issue may become moot anyway. Perhaps LLM powered AI systems (or whatever comes after them) will take the place of Wikipedia for certain classes of queries, that one might want to conduct. Eg, if a person wants to know about Odin and ChatGPT or Gemini can tell them all about Odin, maybe it's less important that Odin have a dedicated Wikipedia page (note: just an example, I'm not trying to make any particular commentary on Odin per-se).

Not really relevant in this case (that the article talks about), but I don't think that it is so cut and dry as "someone spent time on this so we have to keep it".

No, and I'm not arguing for that. Just pointing out that that insidious nature makes contributing to Wikipedia a questionable activity for (a few|some|many|???) people.

To be clear, I'm not arguing for having NO standards at all for what's included, but I think there's room to rethink the nature of some of those standards.

Wikipedia seems stuck in an antiquated worldview where things like traditionally-published books with second- or third-hand reports of what happened, and which are frequently incomplete or wholly inaccurate, are nonetheless considered more authoritative than primary sources you can find with a ten-second Google search.

So much this. Wikipedia's processes and policies are - in ways - an outdated and archaic relic of a bygone time. OTOH, I don't have a definitive answer ready "off the cuff" on what the standard should be. But I think everybody involved needs to acknowledge that the current setup is wrong, and needs serious thought and revision.

And the really insidious thing about this, is the fundamental asymmetry of effort between creation and deletion. Creating a Wikipedia article can take hours, days, or longer, of effort. Tagging an article as AfD takes a few seconds. The actual deletion (once whatever discussion happens) probably takes even less time.

It's amazing that anybody creates Wikpedia articles at all, TBH. I mean, you can spend hours on top of hours working on something and have it all mooted in a few seconds.

This is why you clone a copy to a machine you have full control of!

You're right, I was looking at submissions for userplane.io[1] and over-fixated on it only being submitted twice. My mistake. Nonetheless, the basic principle is the same. Looking at their submission history it appears to be almost all self-promotional and correspondingly many of the submissions are [dead]. As far as I can tell, that's what happens here. (aside: I'm just commenting based on observation of the years, I'm not anybody "official" here or anything).

[1] https://news.ycombinator.com/from?site=userplane.io

Nothing weird about it. The HN guidelines[1] make it clear, IMO, that while some self-promotion is fine, the intent here is NOT to use the site primarily (or exclusively) for self-promotion. The account that submitted the userplane.io link had only ever submitted two links, and both were to userplane.io. What do you expect to happen?

Please don't use HN primarily for promotion. It's ok to post your own stuff part of the time, but the primary use of the site should be for curiosity.

After a few more experiences like that, I've mostly stopped posting because it doesn't feel like the effort is worth it

Ironically this is exactly the wrong response. You should post more, but more stuff that's simply intellectually interesting, and not just to promote your own stuff. But if your only reason for being here is self-promotion, then you're right... not worth it. shrug

[1] https://news.ycombinator.com/newsguidelines.html

Call me nostalgic or whatever, but my laptop to this very day...

https://fogbeam.com/free-kevin.jpg

Depends on how they address the side fumbling between the Spurving bearing and the Dingle arm, vis-a-vis the lotuso-toroidal damping force.

I'd be working on AI.

Interesting. I think I see what you're getting at, and I can kinda buy that.

Speaking as Libertarian, I generally agree that the two major parties have been more alike than different over most of my 50+ years on Earth. But right this minute, I think the Dark Enlightenment dorks and their neo-fascist bullshit have completely captured the Republican Party to the extent that the Democrats are clearly much more favorable. For my part, I'm at a place I never thought I'd be: actively donating money to Democratic candidates and overtly supporting their campaigns, and planning to vote for some of them in the upcoming election. It leaves a bad taste in my mouth for sure, but to me it feels like what has to be done in the moment.

The surge of authoritarianism is a response from people who, much like you, have tried to unplug but who realized closing their eyes does not stop the changing world around them.

I'm not really following your logic there. If the argument is that people choosing to disengage allows the rise of authoritarianism, I could buy some of that. And I'll accept whatever guilt I deserve. But I don't see how that could be the proximate cause of the surge of enthusiasm for authoritarianism. I believe it has more to do with macro-scale world events. And I'm sure one could construct a "just so story" to run the trail back as far as one wants, but I think an awful lot of it can be traced back to the 2008 financial crisis. That along with the continued deterioration of the middle class, rising wealth inequality, etc. And, as much as I kinda hate to say it, I think there's a hint of lingering White angst over having had a Black POTUS for the first time not so terribly long ago.

I’m not saying you’re doing anything wrong,

No worries. And for what it's worth, I'm not - for various reasons - going to elaborate on everything that's on my mind, or every action I still take, or hope to take, in the name of trying to support the aforementioned "freedom and Enlightenment ideals". I'm not trying to write an essay here on HN or anything. :-) My reasoning is in part that if I stop sending time engaging with people on Facebook / Twitter / etc. then I can spend that time on higher value activities.

That said, I'm just a regular Joe (er, "Phil") and I'm no hero. I figure I do what I can in the end I can live with that.

Do you think humans in the future in general will be more or less happy than they are now?

I honestly have no idea. I strongly want to be optimistic and I am generally an optimist by nature. But recent events have made it hard to justify a lot of optimism. I have a fear that we are rapidly sliding into a straight-up cyberpunk dystopia to rival anything from sci-fi.

And sadly it's not just what's going on with technology that fosters my doubts. It's the apparent surge of enthusiasm worldwide - but particularly in the United States - for various brands of authoritarianism and fascism, the wholesale abandonment of Enlightenment ideals about individual freedom and the nature of justice, the loss of respect for science, logic, and reason, and the rising preference for various brands of mysticism, superstition, and magical thinking over rational thinking. Not to mention more overt nationalism, jingoism, bigotry, attacks on minorities of all flavors... Yeah, I gotta be honest, optimism is hard to come by right now.

All in all, this has led to me choosing to drop Facebook (which was the main place I was exposed to a lot of public discussion on this stuff) and to aggressively tune my Twitter feed to eliminate most political stuff. I had to do it to protect my own mental health. And while I feel guilty about the sense of abdicating my responsibility to be involved and to play a role in trying to improve things, I have had to accept that there's not a lot I can do as an individual - at least in terms of influencing other people. I still vote, donate to campaigns, donate to charities I support (EFF, for example) and so on though. I'm not completely checking out, but I was getting way too stressed before to keep dealing with this shit.

Anyway, sorry for the rant. But despite all that, I still hold out some hope for a better future. Time will tell, I guess.

That said, I'm kinda hoping somebody does create a "GNU deadline" project now. I'm curious to see what kind of project it would be.

My first intuition is that we don't have a (generally accepted) definition of AGI that's rigorous enough to really have this conversation. I've heard people make convincing arguments in favor of:

* We won't have AGI, ever.

* We won't have AGI for another 50 years

* We'll have AGI in 2029

* We already have AGI

* We've had AGI since 2021

and so on. So clearly there's a broad spectrum of definitions at work in all of this.

My personal take? I wouldn't call anything we have today "AGI" just yet. And I generally think that just making bigger and bigger LLM's probably isn't the path. OR, I might say that "making bigger and bigger LLM's" could theoretically get to AGI, but if that does happen, the result will be so computationally inefficient that it would still be worth researching other (better) ways to get there. As an aside, I guess it's implicit in what I just said, but to say it explicitly: I believe there are almost certainly multiple different ways to get to AGI.

With all of that said, I think there's probably a lot of value in leveraging well known and understood algorithms that run efficiently (on CPU, maybe GPU, maybe on analog computer, whatever) in combination with neural networks (whether using Transformer architecture or something different). I find myself drawn to looking into neuro-symbolic techniques, and looking at ways to reuse some older ideas like stuff inspired by Minsky's "Society of Mind" or the old "Blackboard Architecture" approach. Throw in the idea of using (possibly) ensembles of SLM's, LLM's, maybe some "MLM's (Medium Language Models, if we can talk about such a thing), in conjunction with ideas from classical planning, FOL inference, etc., and collaboration the aforementioned SoM / Blackboard kind of approaches and I think you might get somewhere. I expect there's a place for some evolutionary techniques in there as well.

Over the years, I've developed a belief that the single biggest problem with all of this is the lack of shared representation (of "knowledge" or whatever you want to call it) that can easily be shared across these different modalities. Like, interfacing FOL inference and an LLM isn't obviously straightforward, at least not in an efficient way. You could hard force "natural language" to be the closest thing to a shared representation and translate in and out of NL and things like N3, Common Logic, Ontolingua, KQML, UNL, etc., but that's still leaning awfully hard into making the LLM do a lot of the heavy lifting.

Anyway, sorry for the long-winded ramble there.

Not Ai, Trusted Citizen Indicated Or Suggested?

Welcome to Vietnam 2.0. Or I guess 3.0 if you count the quagmire in Afghanistan.

Huh. Interesting. I gotta admit, I don't known what this will ultimately lead to for SO and their business model, but give them props for at least conducting a pretty interesting experiment.

These guys (the publishers) are fighting last year's war. Nobody (to a first approximation) gives a shit about going to the NY Times website, or The Guardian website, or the BBC website, etc. to find information. They expect to use search engines and AI services to find stuff, and then maybe click through to the source site(s) for more details or whatever.

The publishers need to rethink their entire take on how the Internet works or any "victory" they earn is going to be extremely Pyrrhic.

"The first 80% is easy... it's the second 80% that gets you."

... and only until the whole thing can be fully formalized with an approved PEP. I'm no Python insider, but that doesn't sound horribly controversial to me.

And yet, I have a hunch it will piss off a lot of people nonetheless and lead to much outrage and wailing and gnashing of teeth. Hopefully it all works out in the long run.

I don't remember one specific moment, but I was fairly impressed with ChatGPT from the first time I started interacting with it. Was I ready to call it "AGI"? No, absolutely not. But it was clear that it was something new, and it was also intuitively obvious to me that "this AI is as bad today as it will ever be" and that predicting the rate of change would be difficult.

The more I use these things, the more I'm 100% convinced that it makes sense to say they are "intelligent" (for some meaning of "intelligent"). AGI or "human level intelligence"? Still no[1]. But some kind of intelligence. And I'm quite happy to allow that there can be "intelligence" that doesn't work anything at all like human intelligence, so arguments of the form "this isn't real intelligence", etc, etc. carry very (very) little weight with me. I've actually been sitting on a half written blog post on this very topic for a while, titled "The Marquee Sign Says 'Artificial' Intelligence"[2]. Finding time to finish it has been the challenge.

And before somebody says "Use AI to write it for you". Nah. I am generally what you might call "pro AI" and / or an "AI enthusiast" but I still draw lines. I'll use AI for research, for outlining, for brainstorming, etc. sure. But I have a hard-line stance against letting AI fundamentally write for me. I want anything that goes out with my name associated with it to have my genuine voice.

[1]: I like the term "jagged intelligence" that Demis Hassabis has been using. That is to say, the bounds of the intelligence are jagged or spiky: very intelligent in certain areas, much less so in others.

[2]: for any old-skool pro-wrestling fans, yes, that is an intentional nod to "Double A" Arn Anderson and his "The marquee sign says 'wrestling'" catchphrase. :-)

C'mon man. I know everybody has seemingly given up on trying to have non-ambiguous / redundant names for things, but that's a little on the nose, considering that there is an actual programming language called Arc already.

https://en.wikipedia.org/wiki/Arc_(programming_language)

It's amazing that after all these years, a famous criticism of Neuromancer seems to have been mooted. That is, the bit about Case having that stash of stolen RAM that was his "big score" for the moment, and how Linda Lee stole it from him, yadda, yadda. For years people have read that and said something like "WTF? RAM isn't valuable enough to be a black market commodity".

Well... I guess William Gibson laughs last, after all!

I haven't always been the biggest fan of systemd in some regards, but I will say that I mostly agree with this sentiment. I've almost completely quit using cron, and now favor systemd timers for scheduled jobs - at the "system" level anyway. I might still embed Quartz for scheduling that's scoped to a particular application or something.

Why? It's one of those fuzzy and somewhat hard to explain things. The systemd approach just maps more cleanly to my mental model of "how things should work" I guess. And maybe some of it is that I did indeed experience plenty of " Ambiguous $PATH settings make cron script execution difficult to predict" in the past, although it's not just that.

I won't sit here and claim that systemd timers are necessarily better than cron in any universal / objective sense. But they've won me over, for what it's worth.

https://www.youtube.com/playlist?list=PLoROMvodv4rMqXOcazWaT...

And there's nothing "wrong" with the GoF Patterns per-se. The issue was always people applying them blindly without understanding why (or more to the point, "if") they were needed. Once writing code filled with patterns became "the thing you do"... all bets were off. :-(

I don't quite get it why they can't take another LLM and vet the output of the first with the second one.

Yes, this technique and its variations[1][2] "work" but it's still not 100% perfect. And it's not as widely used it might be because, among other reason:

a. it takes longer to implement

b. it costs more (more tokens spread across multiple llm calls)

c. higher latency (getting an answer takes longer due to multiple llm calls involved)

d. the final answer is probabilistically more likely to be correct, but is still not guaranteed to be error free, so you can never fully escape the need for Human in the Loop.

[1] https://en.wikipedia.org/wiki/LLM-as-a-Judge

[2] https://github.com/karpathy/llm-council

As far as I can tell, Lattice products are the best if you're interested in using F/OSS tools. Yosys / Icestorm supports the Lattice ICE40 and ECP5 devices. QuickLogic also apparently do some officially supported F/OSS stuff for their devices, but I don't know much about them. Think I might have ordered one of their dev boards a while back, and then promptly put it on a shelf and still haven't gotten around to touching it. :-(

The problem is, none of these devices can match the raw performance of the devices from Xilinx / Altera / Microchip / etc. So you're kinda limited if you are trying to do something that needs serious horsepower.