I don't think it's there yet. If we knew how the brain's architecture produces consciousness and we could recreate that in software, I would be open minded to have AI replace human interaction, but what Zucc is suggesting is that we build a social (or "para-social"?) relationship with a the output of tensor calculus.
That's what should happen over time within a domain. As you engineer the codebase it should become into its own domain framework. I feel the web industry leans heavily into only using existing frameworks and tools instead of just building it. Frameworks are actually simple, but for whatever reason the industry decided that engineering is more about gluing software together while looking at customer metrics instead of solving computer problems.
This is also what I am doing in my day job - our team own a full-stack solution for a Headless CMS and delivery platform - with that we've had to build our bits of the tech stack that over time now allow us to offer it to more team - but with that we've also had to make parts more specific to the domain we work in.
I think it's silly that the conversation on preserving games revolves around hardware. The hardware is irrelevant. It's about the right to create digital backup for personal use. Whether the game is downloaded or burned on a disc, it's just software.
The main problem used to be about piracy, but I think now it's really about making games as a service (even if they're not online for gameplay) because it allows more forms of monetization. The conversation should be about making games into a digital product that you can download and own the files. Piracy still happens anyway, and maybe this could make companies solve the problem differently, like only allowing digital backup for trusted players.
Wouldn't it be nice if the Library of Congress or the Bodleian or some other prestigous clouty institution could demand that these published artworks be given to them as an unencumbered copy? I know that such a thing might have to include server code and some agreements on runtime environment, but I don't see that as insuperable.
This is culture and it's part of our patrimony. The privilege of getting to publish thinga and having copyright protection ought to include responsibilities to the society too.
The problem is the cost and knowledge base required to keep servers running. A game server is a big proprietary ball of spaghetti with hundreds of API endpoints and only the people who built it really know how it works[0]. It's expensive to keep those folks around and expensive to pay for the cloud services and SaaS tools they need to do their jobs.
All software has a "lifecycle" and has to be turned off at some point because no one is willing to pay the costs of keeping it running (with hosting and client changes as ongoing moving targets). We see this even with games that have sales! So ones that don't have sales are not likely to attract anyone to pay for such staff.
[0] Source: I spent 2 years inside a studio owned by "big gaming."
There is a petition that would require publishers to leave games in a playable state at end-of-life (https://eci.ec.europa.eu/045/public/#/screen/home), but it doesn't look like it will reach the threshold that would require the parliament to respond. It is one of the bigger petitions though, so it might still trigger some action.
What your saying is true BUT. Your are talking about keeping servers (plural ?) running but, for conservancy, you should not need a full fledge cloud. It not about keeping it running for millions of customers. It's about being able to run for research to study games in the long run. One server that can handle a couple or two clients should be enough for conservancy.
People are the harder problem. A game server is not a box you turn on and it just runs. The platforms themselves change under you all the time as do all the SaaS tools you rely on. Clients change all the time, too, and one missed update can make your game unrunnable. Folks don't want to train or pay the humans needed to keep server-based games alive.
It could be paid by taxes, and run by government. Something like the Internet Archive (not sure if that's public but the entire Internet is much larger than all games put together).
Yes that would be nice. It could definitely be a non-profit and get a lot of support. Then they could grant access to past games that are no longer on the market. Ideally anyone can just make their own backups without publishing the copy. I still think there are solutions to allow this but companies want games a service to make more money.
I think React caved in to wider adoption pressure to introduce abstractions that are intuitive on the surface level but are costly in terms of large scale complexity.
> It's "declarative" right up until you're debugging stateful hooks, or resorting to useRef, or trying to reason about when a "component" re-renders
Maybe they should have modularized the core library more and have these things be separate, because the core idea of a uniflow pattern with reactivity is good.
I think what happened, at least in frontend, is that the industry pushed away from having engineers do any design or architecting on the frontend. All of these high level patterns have been "outsourced" to frameworks, and the result usually is something that has trouble scaling and adjusting to whatever domain it's in.
Maybe they should have modularized the core library more and have these things be separate, because the core idea of a uniflow pattern with reactivity is good.
That's what SolidJS does. IE the signal implementation is completely stand alone. I feel like it's better at doing what react purports to do then react is.
* think what happened, at least in frontend, is that the industry pushed away from having engineers do any design or architecting on the frontend. All of these high level patterns have been "outsourced" to frameworks*
I don't think react patterns are particularly high level, or do they save you from architecture. Whether it's vanilla JS or react, you still have to design.
Oh I haven't looked at Solidjs yet, interesting will take a look. And yeah you may still need to design your application, but having hooks be something that is out-of-the-box pushes you into certain patterns and needs to be actively ignored to avoid its design influence. I've worked in large codebases where they make almost everything into hooks, and they start getting ridiculous, breaking composability but at the same time giving the illusion that you are making your code more modular.
If vibe coding turns out to be another hype wave, it seems there is nothing to compare it with in the past. This truly might be biggest hype and letdown in the history of technology and markets... or we will hit AGI by 2030 (because AI will vibe code itself into an AGI).
It just feels like there's no middle ground. It's either a really good autocomplete or we're about to live through the MOST unprecedented time in human history.
Interesting. I see from the video example it took a lot of steps and there is a lot of output for a simple task. I'm thinking this probably doesn't scale very well and more complex tasks might have performance challenges. I do think it's the right direction for AI coding.
Right, but the industry wants bad programmers who are good managers and business thinkers. Unfortunately it's hard to draw a straight line from formal methods to profit. So it's hard to convince people to fund a solution that requires highly specialized knowledge and cannot be easily scaled by adding more people.
The tradeoff is that in the long run codebases accumulate a lot more complexity and become brittle, hard to maintain, and a pain to work with. Companies don't value this long term thinking though, and instead want to see a direct line to profit with every ticket and change to the codebase (which of course isn't really conducive to a well engineered codebase).
This is largely true in my experience. The industry regards software dev as the bottom of the pyramid. It's the job you start doing and you get promoted out of. I believe that's a huge contrast to other professions, in which you continue doing the same practice until you retire. The expectations are a lot more than just coding, as if coding wasn't hard enough. This all piles up and creates an overall miserable experience where the hardest part is NOT the technical part.
Yeah this is how it felt looking at high level specs. I basically wanted a 5080 with 24GB of memory. 5090 seems overkill for even gaming especially with power consumption taken into consideration. At this point though, I want anything I can get at MSRP, and seems it will be a while until supply catches up.
reply