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.

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.

Comments URL: https://news.ycombinator.com/item?id=48607346

Points: 3

# Comments: 0

* 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.

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

Comments URL: https://news.ycombinator.com/item?id=48448386

Points: 16

# Comments: 2

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.

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. :-)

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

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

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.