I think I’m too cynical to be involved in this discussion I would be happy to live without image recognition (which I think is a positive) if it meant not having deep fakes and AI art. Better assistive technologies and cancer cures, yes please—if those were at the forefront of this whole thing I wouldn’t be complaining. I do understand that this is how the world works and I’m too much of an idealist.
Otherwise, I have motivations for not using AI for programming that I don’t want to get into in a public forum so I should bow out here, heh. Again, just nothing else to talk about 'round here without me spending a few months diving into database theory
Did I say that? Some will some won’t. How much of them can tell the difference? I don’t know. I am neither a doomer or enthusiast for LLM. All I was doing was to point out the weakness in your arguments.
Of course, please pardon my lack of imaginations. For the sake of discussions, I prefer to anchor somewhere with a least amount of surprise. If I may, you position seems to be LLM will turn tech stacks into something like fast fashion. It is a big if, and even it turns out this way, I don’t think it is a good thing for the new framework creators. Fast fashions tend to build the wrong kind of user base.
This has not in the least bit been unexpected - I posted about the impact of LLMs in the mod forum towards the beginning of last year, and even that was after having thought about it for a couple of years beforehand (I had a lot going on so didn’t get time to put my thoughts down on paper till then).
So if you’re worried about the forum, don’t be if anything, the impact of LLMs is actually somewhat freeing and exciting - because it means we don’t have to be as strict on some of the things we have been strict on in the past (like SEO) and so it heralds a new direction and chapter in the forum’s life We intend to cover some of this in our annual update (which we are holding off posting for now - it’s our tenth birthday next month so we’ll post it closer to that) so definitely keep your eyes peeled for that!
You also probably won’t be aware as you only registered here in 2020, but the forum has already gone through several changes as and when our needs and goals have changed. For instance, in the early days (arguably what many might consider when we were setting ‘the gold standard’ for forums like this) we had things like a general dev section. José was (understandably) actually apprehensive when I first mentioned it but I promised him it would be fine - and it was - it was more than fine, it played a significant role in helping with adoption as it helped create the kind of buzz and activity that was mesmerising and unmatched at the time. You could literally feel that something special in the air! It was infectious! (It was also hard work, hence when it did its job we adapted once again.)
So if you’re worried about the forum please don’t be - it’s not the first time we’ve had to traverse the impact of external forces… some of you may remember the (catastrophic) impact on some forums with the advent of social networks, yet many adapted and are still going strong today. So we’ve certainly been around the block and you’re in good hands - if people continue to enjoy and use the forum it will continue its path… even if it’s a little different to what you may have witnessed yourself during the time you’ve been here
Be sure to check out our annual announcement when it drops for more on this!
Going back on topic…
What new tech stacks
While we may see an acceleration in new languages and frameworks, largely put together with the help of LLMs, LLMs (or rather their overlords) will ultimately get to decide which ones live or die.
We’ve covered some of this in our thread:
And we recently saw a real-life indication of this with Tailwind:
If a huge project like Tailwind struggled to survive, what hope does that leave everyone else?
I think people aren’t considering the fact that LLMs are replacing those sources - because whereas previously people were asking questions on forums and SO, now they are asking LLMs and so LLMs are becoming those resources. And this not only ties back in to that thread where LLM overlords will dictate which languages live or die, but one of the greatest human inventions and one of the best things about the internet - of freely and openly sharing knowledge - will start to die. This could be devastating for humanity as the LLM overloads end up controlling, and in some cases, especially when you consider the evil-ness of billionaires, gate-keeping or nefariously skewing and manipulating that knowledge.
We may well be living in the beginning of Star Wars - where you’re either part of the Federation or the Resistance. Will people choose the side they chose in those movies? I sincerely hope so. I still have faith in y’all!
You said “people like to mess with new stuff when $CONDITION”.
The latter does not on its own negate the former (it supports it!), but because you led with “I would like to challenge this” I assumed that was your intent lol.
I don’t think it’s a good thing. I also don’t feel comfortable asserting that it’s a bad thing.
The only position I am willing to commit to this time is that I don’t care, which is why I have been avoiding these threads, a policy I will be reinstating as of this reply.
Never too late. Start with Andy Pavlo’s courses, he’s the best.
We need more people diving into databases around here. I think that will be Elixir’s biggest strength going forwards.
The next generation of open social networks (ATProto most likely) will consume forums completely this time, but it won’t be a bad thing. Instead the forums will live inside the network and interoperate. There is a beautiful future to be had.
I’ve not looked into ATPro (What is it? What makes it different to all the other decentralized social networks?) but if I had a penny every time I heard something like that
Ok, so we are on the same page. You said: People just like to …, I said no, it depends. My challenge was to the word just, and that you did not prove the $CONDITION is a given..
It is a bad thing for creators of “frameworks” as they exist today. But as you said, “frameworks” may become entirely different things from today’s. So I don’t know.
What makes atproto different is that the list of engineers who consulted on the design was essentially a who’s who of past failed attempts, and they all tried very hard not to screw it up this time. They succeeded, and it works.
As I’ve mentioned a couple of times now I think the Elixir community needs to start taking a serious look at atproto. It’s exactly the sort of infra that we’re really good at powering, and now is the time to get in at the ground floor of what IMO has a solid chance of being a substantial chunk of the future internet.
Of course I will have to put my code where my mouth is. Soon
Speaking of Star Wars. were you referring to the original trilogy or the prequel trilogy? In the original trilogy, it was the Empire vs the Resistance. In the prequel, it is the Republic vs the Trade Federation. The ethical choice is much less obvious in the latter case. I also think the choice we faced today is a lot closer to the latter case.
I’m not sure if I’ve seen any of the new ones tbh, but:
In Star Wars, the Galactic Empire is a totalitarian regime built on military might and fear, controlling vast resources and planets through oppression, while the Trade Federation (part of the Separatists) was a corporate entity focused on economic control, which later became integrated into the Empire
In a real world context they are one and the same to me. The Empire (such as like the British Empire) just basically became (owns/integrated with) capitalism as we see it today.
I mean I could link you the docs but it’s complicated stuff. There is no easy answer to your question that doesn’t involve “trust me bro” unless we get into the details.
They fixed a lot of problems from other attempts at decentralized socials.
One big one is the ability to move accounts between providers, which ActivityPub (Mastodon) categorically fails at.
Another big one is to split up the “network view” providers from the “personal data hosting” providers, which is important because self-hosting your own posts is tractable but self-hosting a view of the entire internet is a serious job. The design enables user choice in both cases, but there are likely to be far fewer “view” providers. This is a natural consequence of full interoperability, and the alternative is to give up completely and silo the network like ActivityPub, which is a design that cannot compete with centralized social media.
They also took great care to come up with a design that allows information to propagate quickly through firehoses while remaining cryptographically signed by the user that submitted the post, meaning notifications can be delivered at the speed users of centralized socials expect.
There is more, but again all of this is in the weeds. What matters is that it works.
One of the reasons I think atproto is so promising, which you may appreciate, is that it’s the first time I’ve ever seen a plausible answer to Joe Armstrong’s “How do we host data forever” question. Unlike the web, data hosted in the ATProto network can be redundantly stored and served by any host/view. There are designs that have achieved this for individual files (torrents, IPFS) but atproto achieves it for schematized, meaningful data.
This is why I am quite confident the network is going to eat forums. I do not mean forums will disappear, I mean that they will be stored within the network logically rather than stored on the “elixir forum servers” and then (maybe) backed up.
Instead of being stored on this server, my posts would be stored on my server and then logically syndicated into a view of the “elixir forum”.
Importantly, this means that the “elixir forum” can never disappear. It would live forever. Not just backed up. Live.
(As a side note, obviously this implies you need moderation tools built into the protocol, and they took care of that too.)
So basically it’s the same as the others (just ‘better’). Maybe it will gain some traction but it seems that it’s basically the same sort of thing as the others (and so my guess at this stage is it will head in the same sort of direction)…
If you mean other attempts at decentralization, like Mastodon, then no. You cannot move a Mastodon account between providers, and this is actually disastrously bad.
If you mean other centralized socials, then no. Centralized socials only have one provider! There is no user choice at all.
As far as traction, Bluesky has tens of millions of users (including, as of recently, me).
So it seems it’s the same sort of thing (but better in your opinion). I’ve also never heard of it other than in your posts (seems like you’re the only person who has brought it up on the forum) so I’m not really picking up any hype - that even platforms like Mastodon has had (but I’ve not been actively looking either as it’s not something that currently interests me).
Star Wars prequel trilogy (1999-2005) are bad movies but deep in philosophy. Star Wars sequels (2015-2019) are garbage.
I agree there is striking resemblance between Star Wars and how the East India Company being absorbed into the British Empire. However, there is never being a time of Federation vs Resistance. Resistance came into being after the Republic and the Federation fused and became the Empire. The badness of the Empire did not come from the trade federation alone; the Republic was flawed too.
To be perfectly honest, I was more lamenting what the form “once was” (like a year ago, lol) since, as I’ve overstated in the past, this is the main place I interact with strangers online. It had been at least two decades since I regularly participated in a forum before joining this one and I just really enjoyed talking about Elixir and LiveView and syntax and conventions etc. I really enjoy programming languages themselves and it’s sad that the bulk of the industry is can’t wait to consider them “lower level details,” but progress is progress and it it is what is it.
Cool that the forum is now 10 years old, congrats!
I’m sure there’ll still be plenty of things for you to discuss Andrew keep an eye out for our announcement as we’ll be making some changes to the forum this year to continue down our new path
I would be happy to live without image recognition (which I think is a positive) if it meant not having deep fakes and AI art. Better assistive technologies and cancer cures, yes please—if those were at the forefront of this whole thing I wouldn’t be complaining. I do understand that this is how the world works and I’m too much of an idealist.
if anything, the impact of LLMs is actually somewhat freeing and exciting - because it means we don’t have to be as strict on some of the things we have been strict on in the past (like SEO) and so it heralds a new direction and chapter in the forum’s life 
We intend to cover some of this in our annual update (which we are holding off posting for now - it’s our tenth birthday next month so we’ll post it closer to that) so definitely keep your eyes peeled for that!
(It was also hard work, hence when it did its job we adapted once again.)
