Yeah, you can make more stuff , but no one cares. Plastic used to be so rare and luxury , now most of them are crap. Software is the same. Creating software is becoming a practice that produces a tiny bit convenience at the price of sustainability.

Music generation, image generation all suffer the same way. You can make some beautiful stuff sometimes, but it (almost) always comes out 'hollow'.

The music industry has already been through this gutting through hyper-abundance.

Huge numbers of people are going to spend billions of hours playing GTA6 though, and that won’t be scarce once it finally arrives.

I remember setting a "4 minute World 1 in Mario 3" as my personal "4 minute mile" in elementary school and the thrill of actually achieving it.

The real treasure is the friends we make along the way... Only: We would we even embark on a journey.

"Art is born of constraint, thrives on struggle, dies of freedom."

That's not really true. Music is still valuable even though the value of music people write themselves is already practically 0 and has been for a very long time. Same with paintings. People are always looking for a connection. Just because some exists now doesn't mean it automatically earns it. It has to have something else going for it. I think that promises a bright future for creaters and consumers of that creativity.

>The other night I carelessly made an infinite runner game while watching TV and the feeling of sheer abundance quickly faded into a bit of depression. Without scarcity, nothing is really valuable.

This is just a lack of ambition. LLMs allow one to make far more sophisticated games than ever before, even as a single developer, and if doing so development still takes many months even if running an agent 24/7.

I don't know about that. My son is about 13. He's got into coding using godot and initially built a few games in Javascript (via. a book which I worked through with him). Then, he picked up Godot from tutorials online and recently has built something multiplayer. He asked for a Claude subscription for his last birthday and since then has been doing all the parts he enjoys by hand (combat mechanics etc.) but using Claude to do things like trailers, website aesthetics, sound effects etc. It's been nice and the results are impressive. e.g. one of his first games was a simple hangman. Now, he added a keyboard to that via Claude so that it can be played using a browser.

He's been publishing his work at https://waitinggodot.in/

My younger one has been creating little things on a microbit using the block based environment online. He's enjoying it too.

I am still imoressed by hand-painted signs in Nepal and India - enough to take many pictures of them - even though AI-generated signs are now everywhere.

My watercolours still impress people enough to get requested even though they're fairly mediocre.

I can't imagine not being impressed by something made by _the person that I created_ because a machine can spit out a neater version of it.

I agree. But I think the problems have now just moved. What is 'hard' is different. I don't think we've collectively grasped that yet. We're going to need to use our imagination and show others when we find something. I'm looking forward to what we can solve, I just hope the opportunity and benefit is well distributed.

It’s a different medium altogether

That’s like saying they will never impress anyone by discovering a continent

That ship has sailed

It's truly awful. Best case, they go on and start tinkering and fixing and learning. What is more likely and awful is they will sit there pulling the dopamine lever and get hooked on using AI to do everything difficult.

Try to go back before our era though. In the early age of computers, computers were basically really big one-way calculators that could execute an algorithm and output a result. In this context a kid creating a program to output the fibonacci sequence would have been quite impressive. But by our era, we needed to be able to produce basic text games to impress. Did we really lose anything in the process? I don't think so.

And now too the impressive demo isn't going to be just making an interactive or multiplayer game which anybody can do with a single prompt, but creating something that's actually interesting/fun. Interestingly enough this now overlaps quite well with the sort of stuff that could be monetized today. That's why I think LLMs are going to throw a curveball in the economy in ways many aren't yet appreciating.

Meh, now they can impress with something cooler.

I spent so much of my youth building things that were far beneath what I wanted to be building because it took all my time/energy/attention/expertise to build the lesser thing.

It's a far better deal to saturate your time on the high-level stuff building the ambitious thing you really want to exist. e.g. https://www.gamedeveloper.com/design/-quot-the-door-problem-...

> I feel for kids growing up with LLMs like this. They will never impress a parent, teacher or peer with a simple “hello world” type text game, when the peer could generate a full multiplayer video game with a couple of prompts.

Adults are also desensitized to impressive deliverables. We see it even among peers who barely bat an eye when someone goes above and beyond to deliver a feature. The toxic ones even try to downplay your achievements by throwing around claims that your deliverables were vibecoded.

An old tech friend and I were talking a few years ago about hobbies, and how our old CTO was one of the most impressive hobbyists we knew of. Not only was he an incredibly accomplished exec but he also produced amazing YouTube content about his model railway collection and model rocket building all in a custom home where everything had custom smart interfaces with a growing family who enjoyed home cooked meals together on furniture he built himself.

Built himself? Well, designed by him and built in his home workshop. Some of it done by him at least with his assistants. They had a home chef too of course who gave the kids a bowl of stuff to mix during prep. The model railway involved design work that was more like commissioning 3D prints and model assembly. Input on the design really, as a lot of it was handled by assistants. Assistants assisting assistants based on prompts and review and feedback. Everything was custom and bespoke and made in his vision. Almost the entirety of the labor and craftwork were outsourced.

This person was a great leader — or “orchestrator” in the language of agents — but a lot of his accomplishments boiled down to having the physical world equivalent of a massive token budget: money and staff.

The great scientists of the 19th century, the ones whom we named fundamental units after, had skilled craftsmen who cut precision screw threads and blew perfect glass for them. Major scientific leaps are written up on papers authored by department leads et al. Les chefs célebres of the world front a kitchen filled with skilled staff supporting the name over the door.

If your work can be done on a computer though, AI can provide you with a Python line cook or a C++ glassblower of your own, so you can’t just show us your perfectly julienned potatoes or evaporation flask and expect us to praise anything other than your wallet, and your ability to motivate robots to do your bidding.

Not once did my boy ever impress me with his ability to program.

An AI companion that acts impressed will be easy to provide if the winners see a benefit of doing so.

They will just have to compete to see who can make the best GAME, I suppose.

Seriously. I remember coding something in Macromedia Flash to impress my computer-illiterate parents. It was silly but it was all mine. I am sad that I will never see that for the future generation, not this way at least. The bar is too high now.

over-achievement society in the words of Byung-chul Han.

Not sure whar you are mourning. It is just that the bar has gotten stratosphericly higher, which is awesome.

I along with most of HN probably had a new idea every month for a game or software back in the day. 12 year old me could only dream to maybe work on something tangential to them in the next twenty years or even a lifetime. Now i could prototype most of those ideas as i thought them up and really get into the promising ones.

I feel like i was born too late and missed out, straight up jealous of the possibilities in front of the kids willing to take it

But for how long?

I have a personal benchmark suite that runs a bunch of difficult algorithmic problems in 3 tiers: vibe coded prompt (just a few sentences), vibe coded prompt plus a description on what the solution will be evaluated on, and finally a fully specc'd prompt that leaves little space for interpretation.

The current Fable models are approaching a level where the -eval runs perform systemicallty better than the "clear spec" runs. (And right now its just Fable 5.1 and Opus 5.5 that behave so well, all the other models I tested - all the OpenAI and the leading Chinese models - work better with clear specs). I'd encourage everyone to experiment a bit; I am running my serious workflows with labourious force-me-through-all-the-design-questions mode, whereas my hobby projects, some of them hard, I am trying to see how much I can get away with with vibe-grade prompts. My conclusion is, with Fable its starting to look like what Vibe coding was promissed to be a year ago - a magical box that generally does things very well.

Btw, congrats to the OP -- the game is enjoyable and I played a few levels. I dont remember when I last played any games for more than 5 minutes. Good stuff.

But one shot prompting does not mean one shot testing. I see continually that Fable can design and test prototypes before committing to a design goal and testing multiple layers until the goal is reached. Many of the tests are code unit or integration tests that can be re-used without token spend.

> "one-shotting" is fun for a low (human) effort toy, but it's a mistake to get overly hung up on LLMs getting everything right on the first try, since a lot of that is the harness and the internal prompting that makes it burn extra tokens to test everything to get it in that one "shot" (ui turn).

I'd be more specific and state that these models are now way far from being LLMs and instead they are full blown agent orchestration systems capable of running multi-layer ReAct loops and tool-calling, and with your whole session history as context. It doesn't make any sense to refer to them as glorified chatbots as the LLM part is nowadays far from being the component that delivers the highest value.

> Fable 5.2 changed some game mechanics despite my prompt asking for them to be the same as original game.

I couldn't care less about AI-generated games, but such restrictions are just destroying the value of proprietary models. Even more frustrating is that it so far hasn't been considered illegal to steal all the information for training the models, but now it's becoming illegal to ask for these same models to output this information back.

In my 2 minutes of testing I liked the gameplay elements from Fable 5.2 much better. I think Opus 5.5 sound effects may have been a bit better but that's it.

I am kinda new to this -- how is this deployed in claude website?

opus 5.5 version seems to never time out powerups like fire ball. this is wrong and obvious.

I think there is an important difference here.

It can create something like DX ball easy, because its already there, the AI already knows what it is

It can not create a game on the same level as DX ball, something new i mean, not a mechanic added on top, a real new game that feels like a game that would have come out the same year as DX ball

AI is great at executing the task, but it can not for its life (or for the anthropic reputation, it seems to care more about that) create something new from scratch, it can not do game design thats not just change the speed of the ball or the paddle color. Probably a limitation thats there from the start

> Fable 5.2 changed some game mechanics despite my prompt asking for them to be the same as original game.

Couldn't you have been downgraded to 4.8, since direct reproduction of existing games classifies as unsafe?

I use Sonnet 5.5 and Opus 5.5 both on Medium to produce apps in Rust/Tauri all day every day on a $20 Pro sub, and absolutely these two models do a great job with that stack. I rarely run out of weekly credits, unless I use Opus 5.5 on high or ultra code. Medium does the job 99% of the time. Fable is good, but it'll never be as good as Opus when well prompted and with some iteration. I see Fable as a preview of what the next model will be able to do, but it's really not a good production model unless you have a lot of cash to burn.

What model did you use for the rust rewrite? Opus 5.5? I mean it seems pretty crazy that you had such different results.

I don't know, when every kid has had 3 custom made games offered for their birthday, won't they just get bored? Obviously through our lenses this seems awesome because that's something we would have dreamed of, but we would have dreamed of it because it seemed impossible.

Agreed. Could you imagine? Those kids must have been so stoked!

Though, now they will never stop asking for more and more. OOP may have opened Pandora's box. :)

its insane, to actually get a custom game made for you for your birthday, that you enjoyed? wow... never thought id see this happen.

Kids? Forget the kids. This is awesome for everybody. I'm busy making 'adult games' (not that sort) that I always wanted to make, but lacked the artistic/sound/etc competence to do so. And while I hope to make a buck off them, I don't even really care. Even if somehow they failed, in the end I'd have stuff that I wanted but always lacked certain skills to complete.

I don't think people are yet appreciating how much LLMs are going to change the world, especially not the doomers. People are just jamming the present into the future, with greater efficiencies. But it's nonsense. The cost of cognitive work will approach $0 which is going to send the raw value of cognitive output to something close to $0.

So the 'winners' will be whoever fills the resultant vacuum in the economy. That will probably be a mixture of social, physical, entertainment, and other such domains. But it's all so hard to predict. We've lived our entire lives in a completely agrarian society, and we're seeing the advent of the Industrial Revolution. It's easy to predict super-farms, but those aren't going to be the 'real' winners.

Not really, one of the joys of being a kid was forming your own "studio" and making games for game jams and stuff. Being able to one shot stuff really destroys the fun.

Positive little kiddo gets more screentime!!!! Yaaaaaayyy!!!!! Negatives: we either are annihilated by AI or enslaved by technocratic feodalism living in a worse conditions than 1800s mine workers. Uau

Yeah, it would have been expensive via the API; the tokens were free when I tested it, but it would have been something like $300 USD for the Claude tokens at Fable 5.2's API prices.

Fable is $8/$40, with cache reads at $0.2 and cache writes at $10, so you simply calculate based on the info from the blog:

> about 320M tokens read from cache, 19M written to cache, 1M output

320 * $0.2 + 19 * $10 + 1 * $40 = $294

The article clearly explains what the author/their nephew did. There is no ambiguity. One-shot never means 'so perfect there is literally nothing to be improved.' It means no back-and-forth iterations between the human in the loop and the AI.

One shot: One prompt; no need to interfere or make further modifications after it is done. You are totally inspecting other parts of the game. Not the same meaning.

Good call. I haven't really tested it much on the keyboard - we mostly used the PS4 controllers, but I'll add these fixes.

I’m going to go out on a limb and predict the opposite.

I think “the youth” are going to grow up with a Cambrian explosion of styles, mashups, callbacks, revivals, etc.

Right now a vibecoded game just uses the generic style. But you could easily imagine the kid saying to Claude “rebuild this game but in the style of that old show [Samurai Jack | Ren and Stimpy | … ]. We are so early in the tech curve, market penetration is still very low, and the “creator-years” in the medium is tiny; it’ll be 10x in a year if not more.

AI is amazing at style transfer, and so there’s a huge payoff for exploring and learning references to existing artists.

(I also think copyright will die pretty soon for similar reasons, that’s a separate take though.)

That was already happening before AI, with in-house game engines dying off and all converging on the same quasi-realistic style and often the exact same assets.

Games used to have very unique, noticeable styles that you could instantly place from a screenshot. Now everything just looks like Unreal, Unity, or generic 2D.

Hopefully AI might actually make it easier to be more ambitious with unique art styles, should the developer choose to do so.

I'm kind of hoping that the youth will recognize that stuff that can be made with AI is just nothing special anymore, and find something that is (or still is) special. I don't know what that is, maybe something that AI can't do on a fundamental level (maybe something physical?).

(Of course, at this point it might be necessary to point out that "the youth" is a diverse bunch, and some will be like this and others like that.)

You can change the stylization with LLMs. I think what we're seeing is simply any barriers to code/design/art/music/etc completely going away.

But those barriers going away now means that the starting point is much closer to where the finish line used to be. Like a few years ago this would have been an upper tier beat-em-up. I think we're still early enough in the window that I'd still call this upper tier, but the reality is that it's also now something that an uncle can vibe code for his nephew over a weekend.

In general I think this is going to be very good for everybody, but it's going to completely destroy the software as a problem-solving industry, and make the software as entertainment industry much more competitive.

This style reminds me of those mobile game ads with helpless characters that need YOUR help not to freeze or something. Probably the exaggerated facial expressions.

Or the small Battle Wagon around the height of a child quintupling in size the next frame. Kiddos having fun is the main mission though :)

This nephew game is indistinguishable from 90s arcade game styling.

I disagree. Fable didn't make any mistakes and created a completed game with zero guidance, so it's a one-shot. If additional requests come up afterwards, it doesn't detract from the initial execution.

One shot should mean a single prompt. Any follow up prompts are no longer one shot. Two shot, three shot, etc

Thanks for pointing this out. Before reading, I had assumed that the author was using "one-shot" in the standard vibecoding way, meaning "one conversation turn". The title is slightly misleading. But he did have an enjoyable, playable game after a single prompt—but I think that prompt may have included the earlier, boring version of the game?

"Great work dude"

you can't be serious, the animations are completely off here.

It's crazy how even the intro text to the Fable 5.2 Extra's game still reeks of AI. Like, wow, the models can make a game that almost makes sense, but still can't write text that sounds remotely like a human was involved!

"find all documents, descriptions, walk through of game XYZ, generate a consistent and declarative document that another agent would read to recreate the game without using existing IP or protected data, do not reuse art but describe general details with the ability to extend and add specifics later. focus on functionality first, then mechanics and then later art and gameplay. generate a 10 page document that will follow MVP to Production quality gameplay"

give it a link if you want

Based on what I've seen, there's a good chance this would work today (though Anthropic recently added some restrictions around this sort of prompt) -- try it!

It's a one-shot. Iterate and you can make magic happen. https://www.youtube.com/watch?v=UiWKfq4niC8

This is a bizarre complaint. There's plenty to complain about with regards to you know, normalizing kids consuming AI content, giving a gift from a slop machine instead of something hand-made, etc. But come on. The game itself is insanely impressive for something generated in one prompt from a bunch of doodles (cool doodles, but still!) If a single human hobbyist dev created this in a few days we wouldn't call it "kind of embarrassing".

This is back with Opus 4.1 when I was still amused with these things. It definitely took better part of two weeks of intense non-stop prompting, but I wanted to see how far I could go with just prompting for everything (now called vibing) when this became viable https://www.susmel.com/stacky/ (press h for help)

Turns out, you could go far but you never lose that AI touch to it which I find absolutely horrible and (s)tacky, albeit technically impressive to be able to do that by not touching any code whatsoever.

I Googled it as I had never heard of that game, and it turned out that I knew it as Bop'n Rumble. Great game.

I think a lot of people know their great grandmother's name.

Yes, these are great questions, and I appreciate your politeness.

Basically, I think it's okay to have a bit of fun playing AI-generated video games at your techie uncle's house, and their parents agree. And for what it's worth, Aiden drew every character and boss himself - the game was built from his drawings.

My nephews are otherwise very active, learning music, martial arts, and other sports. But when they visit, I let them play games in between other activities. My parents refused to let me play any video games at home, and I still think that was really lame. Video games are another art form to be appreciated - they just need to be enjoyed in moderation.

Generally, though, I worry about the future and the potential for cognitive decline that comes from having AI agents do everything. I don't have a good answer here, except that I suspect we'll have to intersperse AI use with periods of deliberate daily cognitive struggle to avoid decline.

I'm not OP, but it seems to me that LLMs offer extraordinary access to new skills and information to the uninitiated. Someone who has a positive experience creating a game like this is only a question away from learning details of how it was implemented, the language used, the architectural choices and rationale, etc. The LLM is a competent tutor in all. What I would have given as a child for such access... I had to make due with the help docs, and an occasional book from the library.

I first learned to program in middle school with Logo, and the extent of that was "telling the commander" (or whatever the term was) what to do in almost plain english.

That's one of the things that got me into learning to write code in the first place.

As long as the child in question gets to explore their creativity like this and it instills that sense of learning and diving deep then I'm not sure there's a problem here.

>The first thing I felt when I saw this post was concern on the impact of this type of thing on the mental development of children.

What impact did being able to take a picture have on the mental development of children?

Recording a moment went from something you needed a court painter and two weeks to do, to something that everyone could do with the click of a button.

A century later kids are still happy to draw things.

Counterpoint, I think it’ll get them more into making things, because they don’t have that big frustrating period where their output is nowhere near as interesting as the giant smorgasbord of interesting stuff being pushed on them constantly. I’d much rather they spend their time making stuff with AI than passively consume the garbage streaming on the Disney Channel and Nickelodeon 24/7, which was the status quo with my generation. Making can now compete on instant gratification with all the other stuff now.

When you buy your kid a cake for his birthday, do you think that it's hurting the development of his baking skill, because now he knows cakes are something that he can just hire someone to make with money?

I mean… it was a birthday present. Everything in moderation, and there are a lot of ways to involve the kids in the process more.

I definitely have some concerns about AI, but I would much rather my kid is actively involved in the game design process than playing the “buy more gems to level up!” casinos in the App Store wastelands.

I don't think this effects the kids more than playing any other video game.

> concern on the impact of this type of thing on the mental development of children.

Do we, or our children, ever learn to farm, mine, chop wood, lay bricks, make operating systems, or program in assembly language?

Does that affect or limit your or your children's ability to eat food, sit on chairs, sleep in beds, or use computers and phones?

the crazier possibility is that little kids' drawing skills - whatever pedagogical value that has, i agree that it has some - may stop improving, but there might be way more videogames authored by them that turn out to be really good.

Yeah I for one make my children design and code every video game they play by hand. And before you ask, yes, in assembly.

2029 has you covered.

No.

I have been slopping out games for years now. The graphics don't matter unless the gameplay does. The gameplay doesn't matter if the game is too hard or too easy.

The games that do well, like literature, survive because they touch the soul.

I worry about them never having the satisfaction of working hard on something and truly creating something that is their own. I definitely sound like a boomer (even though I'm a millennial) when I say this but I totally believe that limitations breed creativity. The AI here is just looking at stuff artists from the 90s laboured t create for Arcades and made its own slop version. But then again, one could argue that Mario Paint let you create slop too back in the 90s. Yet, at least with Mario Paint you had to actually hustle quite a bit arranging pixels or musical notes to have something. There was effort and difficulty with overcoming a learning curve and that effort paid off with the creation of something you could be proud of even if it wasn't professional. But here, on the other hand, even a badly generated prompt still gets you a lot.

There's like, no challenge. Reminds me of that one Twilight Zone episode "A Nice Place To Visit" where a bank robber dies after getting shot by the popo five-o and he is greeted by a man (angel like guy) who takes him to a cool mansion or house where he fulfills everyone one of the guy's wishes: women, money, gambling, etc. etc. And he gambles all day long and never loses but after a while he gets bored and then depressed that he always wins and there's no effort and no risk, and then tells the angel dude that he wonders why he was put in Heaven and that he thinks he doesn't belong there and wants to go to the "other place" only for the angel guy to tell him he is in fact in "the other place".

Stuff that AI generates sorta feels like that. Meaningless and synthetic.

So in a sense, I don't worry about generations not necessarily having hardships like us, cause, look at us, we don't write in assembly code like other folks had to do. But on the other hand I do worry that AI is basically removing all manner of human touch and the satisfaction that comes with suffering to generate that human touch.

I suppose there will be suffering later when you want to polish things, but it will be a different, more frustrating kind of suffering. The kind where you don't know what the hell the LLM did and don't even know where to begin.

They are already growing up on slop. Look at brainrot and Roblox games, and YouTube shorts. This will fit right in with what they're used to. They won't even notice.

Every generation believes the next generation is growing up on slop.

Remember 2025 when we thought future generations growing up on tiktok was going to be bad? Can't wait to see how 2027 will ruin society even more.

You registered an account called “slopinthebag” 9 months ago, but you weren’t worried about this until you read this post?

Most games and shows for kids have been slop since long before AI.

Who says it has to stay slop? Nobody is forced to stop at the single, first iteration of something. They can see the outcome, play the game, and then iterate faster than any professional game dev studio could in the past.

Its funny that the same CS saying "garbage in garbage out" seems even more true in the AI modern day.

I agree with you. It’s cool that this tech exists and that something this complicated can be something that regular people have access to.

On the other hand, I’ve got two issues with it:

1. The result gives off the AI slop vibe. It’s like the 2026 version of bitmoji. You can immediately tell this is AI art, AI music, AI game.

2. Art made with AI always feels gross. It feels like substituting creativity. At least when you use AI to make a CRUD app or automate some tedious process there’s an excuse. The time spent doing this could have been spent actually making art. Maybe it wouldn’t have been anywhere near as good or complicated art, but it would have been authentic.

> On the other hand, it's very much like you photocopied some comic books, cut them out, and did a little home theater thing with them

I mean, to be honest, that seems like a fun craft for a thing to do with kids!

I share your general ambivalence though. The generic openai-looking art is kind of gross. At the same time, iterating with a child so they have to think about feedback, how could this be better, what do you like, what do you not like - that's a meta-engagement that's probably positive at some age brackets.

Reminds me of how I can probably make the Team Rocket playable terror cell game I always imagined without needing to worry about its economic viable or liability

Thanks

Who cares about IP rights for a one off game you made for your nephew? Are you worried your nephew won't be able to publish it on Steam? I could imagine reasons this could make you sad, but disney's IP isn't one of them.

Could you even call it "working" for video games, based on the result, for the same kind of reasons you'd say it doesn't work for those other media?

(To qualify the rest of this, clearly it's job done in terms of the author wanting to make something memorable for their nephew - it's quality uncling, with the best of intents - but in the context of how what's been shared is going to contribute to the discourse around AI, in the shadow of a new model release, during an ongoing pre-IPO hype bubble...)

I'm seeing something which has has the surface-level features of a belt scrolling beat 'em up, but it's missing the result of all the different types of work that would normally happen through "make it actually interesting and enjoyable to play" design and iteration, e.g. you can walk through every stage unchallenged mashing punch, there's no evidence from the 2 stages I played of an interesting mix of enemy types with unique AI which can combine in ways which add challenge and variety to a belt scroller, you can stand and basic punch the first boss to death stepping aside only for its AOE attacks, the second boss infinitely spams a running move when it reaches low health and sometimes bumps into the edge of the screen behind it.

_Something_ was one-shotted, sure, but what we expect of a belt scrolling beat 'em up that someone cared enough to finish and release wasn't, not that people who will point at this in the near future as evidence of "OMG new Fable can one-shot games" will care.