Oh come on. It’s called AI, as in artificial intelligence. None of these companies have ever called it a text generator, even though that’s what it is.
I get that it’s cool to hate on how AI is being shoved in our faces everywhere and I agree with that sentiment, but the technology is better than what you’re giving it credit for.
You don’t have to diminish the accomplishments of the actual people who studied and built these impressive things to point out that business are bandwagoning and rushing to get to market to satisfy investors. like with most technologies it’s capitalism that’s the problem.
LLMs emulate neural structures and have incredible natural language parsing capabilities that we’ve never even come close to accomplishing before. The prompt hacks alone are an incredibly interesting glance at how close these things come to “understanding.” They’re more like social engineering than any other kind of hack.
The trouble with phrases like ‘neural structures’ and ‘language parsing’ is that these descriptions still play into the “AI” narrative that’s been used to oversell large language models.
Fundamentally, these are statistical weights randomly wired up to other statistical weights, tested and pruned against a huge database. That isn’t language parsing, it’s still just brute-force calculation. The understanding comes from us, from people assigning linguistic meaning to patterns in binary.
Brain structures aren’t so dissimilar, unless you believe there’s some metaphysical quantity to consciousness this kind of technology will be how we do achieve general AI
Living, growing, changing cells are pretty damn dissimilar to static circuitry. Neural networks are based on an oversimplified model of neuron cells. The model ignores the fact neurons are constantly growing, shifting, and breaking connections with one another, and flat out does not consider structures and interactions within the cells.
Metaphysics is not required to make the observation that computer programmes are magnitudes less complex than a brain.
Neural networks are based on an oversimplified model of neuron cells.
As a programmer who has studied neuroanatomy and the structure/function of neurons themselves, I remain astonished at how not like real biological nervous systems computer neural networks still are. It’s like the whole field is based on one person’s poor understanding of the state of biological knowledge in the late 1970s. That doesn’t mean it’s not effective in some ways as it is, but you’d think there’d be more experimentation in neural networks based on current biological knowledge.
The one thing that stands out to me the most is that programmatic “neurons” are basically passive units that weigh inputs and decide to fire or not. The whole net is exposed to the input, the firing decisions are worked through the net, and then whatever output is triggered. In biological neural nets, most neurons are always firing at some rate and the inputs from pre-synaptic neurons affect that rate, so in a sense the passed information is coded as a change in rate rather than as an all-or-nothing decision to fire or not fire as is the case with (most) programmatic neurons. Implementing something like this in code would be more complicated, but it could produce something much more like a living organism which is always doing something rather than passively waiting for an input to produce some output.
And TBF there probably are a lot of people doing this kind of thing, but if so they don’t get much press.
Pretty much all artificial neural nets I have seen don’t do all or nothing activation. They all seem to have activation states encoded as some kind of binary number. I think this is to mimic the effects of variable firing rates.
The idea of a neural network doing stuff in the background is interesting though.
The fact that you believe software based neural networks are, as you put it, “static circuitry” betrays your apparent knowledge on the subject. I agree that many people overblow LLM tech, but many people like yourself grossly underestimate it as well.
Language parsing is a routine process that doesn’t require AI and it’s something we have been doing for decades. That phrase in no way plays into the hype of AI. Also, the weights may be random initially (though not uniformly random), but the way they are connected and relate to each other is not random. And after training, the weights are no longer random at all, so I don’t see the point in bringing that up. Finally, machine learning models are not brute-force calculators. If they were, they would take billions of years to respond to even the simplest prompt because they would have to evaluate every possible response (even the nonsensical ones) before returning the best answer. They’re better described as a greedy algorithm than a brute force algorithm.
I’m not going to get into an argument about whether these AIs understand anything, largely because I don’t have a strong opinion on the matter, but also because that would require a definition of understanding which is an unsolved problem in philosophy. You can wax poetic about how humans are the only ones with true understanding and that LLMs are encoded in binary (which is somehow related to the point you’re making in some unspecified way); however, your comment reveals how little you know about LLMs, machine learning, computer science, and the relevant philosophy in general. Your understanding of these AIs is just as shallow as those who claim that LLMs are intelligent agents of free will complete with conscious experience - you just happen to land closer to the mark.
it’s like how techbros constantly want to reinvent transportation, if they assign an AI to give them an answer it would just say “build more railways and trains” and they’d throw it out a window in anger
They re-invent everything for no reason. Every mundane device has been “re-invented” using big data, blockchain, VR, now AI and in a few years probably quantum-something.
The entire tech world fundamentally ran out of ideas. The usual pipeline is basic research > applied research > products, but since money only gets thrown at products, there’s nothing left to do research. So the tech bros have to re-iterate on the same concepts again and again.
Watch, hold on, I’ll prove it! I’ll perform a feat of brute strength in a blind rage that will end up hurting me in the long run! Then later when I find out that massive fall didn’t actually kill you and you fought your way back up through 2km worth of test chambers powered by sheer spite to come and confront me, I’ll act like nothing happened and beg you for your help because I have no idea how to run this place and it’s falling apart and the robot test subjects I built don’t work at all!
The perpetual chicken egg problem of IPv6: many users don’t have IPv6 because it’s not worth it because everything is reachable via IPv4 anyways because IPv6 only service don’t make sense because they will only reach a subset of users because many users don’t have IPv6…
I mean, yes and no. For an individual or individual systems? No, it's not hard. But I used to oversee a WAN with multiple large sites each with their own complex border, core, and campus plant infrastructure. When you have an environment like that with complex peerings, and onsite and cloud networks it's a bit trickier to introduce dual stack addressing down to the edge. You need a bunch of additional tooling to extend your BGP monitoring, ability to track asynchronous route issues, add route advertisements etc. when you have a large production network to avoid breaking, it's more of a nail biter, because it's not like we have a dev network that is a 1-1 of our physical environment. We have lab equipment, and a virtual implementation of our prod network, but you can only simulate so much.
That being said, we did implement it before most of the rest of the world, in part because I wanted to sell most of our very large IPv4 networks while prices are rising. But it was a real engineering challenge and I was lucky to have the team and resources and time to get it done when it wasn't driving an urgent, short timeline need.
How about “Let me selfhost my own repos, so other people working with my stuff can use IPv6, as well as be sure no large corporation known for being cancer stands behind it and monitors every thing I do.”?
Honestly this isn’t even true anymore. Most major ISPs have implemented dual stack now. The customer doesn’t know or care because it’s done at the CPE for them.
I use a browser extension which tells me if the site I’m at is 6 or 4 or mixed. In 2024 most major sites support V6. A lot of this is due to CDN supporting it natively.
The fact that GitHub doesn’t is quickly becoming the exception.
Globally it’s at about 47% and growing at about 4% per year. If the rate remains unchanged it’ll be about a decade for >95%.
But the reality of it is, you don’t need global adoption out of the box. You just need majority adoption in the countries you visit, which for me are western countries (north America and Europe) which now have a majority adoption.
In the USA they charge extra for IPv6? I'm in the UK and while there are some ISPs that don't provide IPv6 at all, and some that do shitty things like dynamic prefixes on IPv6, I've not seen anyone charging for it.
Likewise, server providers generally don't charge for it. In fact, they will often charge less if you don't need IPv4.
No don’t take shitposts literally. I’ve been using ipv6 for a decade at home now in the USA and I don’t pay extra for it ever. Also why are you assuming this post refers to the us?
There's been other posts about IPv6 and the TL;DR is that while there are shitty implementations everywhere, the USA seems to be ahead of the game of doing it badly, if at all.
The USA is ahead of most nations at about 50% so not sure how you’re coming to that conclusion based off of evidence. Outside of maybe Brazil in the americas on both continents our ipv6 adoption is better than the rest, Canada included.
I reckon I see most IPv6 complainers are from the US though…
In my country, turning on IPv6 is not really something ceremonial, it’s just literally clicking on the IPv6 checkbox. The default configurations set in the router are good enough for an average home user, firewalls and all that security jazz are enabled by default.
The DNS didn’t break just because I enabled IPv6, nor did my phone apps stop working. Life goes on, and I have gotten rid of that terrible CGNAT. Somehow this is not the case for many US users across multiple ISPs, I have heard IPv6 horror stories from Verizon, Comcast, and AT&T. Like how did you manage to do that?
I mean I’ve been using native dual stack for over a decade and I’m most definitely American. A fun anecdote was I was having issues with clicking on links from Google once and turned out ipv4 was busted but 6 worked fine for half a day. And there really isn’t any turning on ipv6 I get it by default and it’s with the most hated isp Comcast. They’re actually really good about v6 support I’ve not moved off them because of it. It’s literally 10ms faster than 4 lilely due to cgnat.
I’ve tried multiple times to go IP6 only. I mostly thought, despite my reasonable understanding of IP4, that I was the problem in trying to set it up. I found my dns host was being forgotten multiple times a day, set to something invalid, then it would time out and revert back to the working one. I couldn’t figure out how to connect two computers together for Minecraft.
Now I hear it was just garbage consumer hardware and software? Fuck me. So much wasted time and effort to say nothing of believing I had turned into a tech idiot.
Do you have an example? Because it works great on openwrt, dd-wrt, pfsense, opnsense, unifi, mikrotik…and then if you’re using the isp equipment it works out of the box.
TP-link can’t open ports in the v6 firewall neither can Linksys and it doesn’t support DHCP forward so literally was incompatible with my ISP implementation. Some current TP Link router sold at Walmart don’t even have an IPv6 firewall.
Open source works great. Can’t speak to unifi never seen it for sale here.
You’re using open source third party firmware and higher end networking gear as an example. Of course they work. Shitty consumer grade brands aren’t in the same class
Interesting that India has such a high percentage. I’m guessing it’s because most of their network infrastructure is probably relatively new and so they can include support right off the bat, instead of having to retrofit stuff?
Not much choise i guess. Usa and europe grabbed the majority of available ipv4 space. Asia got a bit. And only scraps and leftovers for africa and latin america.
China block much of the internet so who knows with china. Do not know if anyone have real china numbers of IPv6 deployment. They also had their own “IPv9” that was rumored some years ago that may or may not have been used internally.
CEOs are obsessed with value derived free of all that messy human labor. It would make sense if they didn’t still want the people they fired to pay money to talk to the robots.
github.com doesn't have a AAAA DNS entry. So it's not serving anything directly over IPv6. Likewise, ping -6 github.com fails. So, what are you seeing that is supporting ipv6?
That is interesting. I figured they would be something like cloudflare/other redirection for github pages. But the IPv6 address space is github registered.
So, really not sure why they don't have the rest of their site enabled.
programmer_humor
Oldest
This magazine is from a federated server and may be incomplete. Browse more on the original instance.