Same for smartphones. There was a time of unlimited hype and secrery around new iphones. Who remembers the story of an iphone 5 prototype left a bar. Journalists were going crazy, people were signing petitions for Apple to not hunt down but instead forgive the employee that made such a grave mistake. People were offering millions to buy the prototype so they can brag they got the new phone 2 weeks before everyone else did.
Or who remembers the dancing disease of 1518, were people would stop what they are doing and start randomly doing the same dance. The lords? Out of their minds. The priests? Terrified the devil had taken hold of the flock!
I have come to believe that it was probably some tik-tok like hype trend of doing a fortnite dance while waiting in line for bread and communion. And the energy back then, like now, was off the charts.
Hype and memetic trend seeking encoded deep in human psyche.
> Who remembers the story of an iphone 5 prototype left a bar.
That was the iPhone 4 actually which was special for having the first “retina” screen. It was in a case to make it look like a 3GS to be used for field testing. The journalists that got their hands on it and published about it before the announcement could not turn it on past I think the Apple logo and a message saying to return it to Apple (or maybe it was just completely off, my memory is fuzzy), but were able to confirm the pixel density via microscope and I remember it blowing everyone’s minds at the time.
> people were signing petitions for Apple to not hunt down but instead forgive the employee that made such a grave mistake
FWIW I asked about it when I worked at Apple and he was indeed not fired and I think may have even still been working there when I was there around 10 years ago (though don’t quote me on that last part, he may have left already and I’m misremembering).
He was apparently not fired or even really reprimanded since it was a genuine accident and he wasn’t the one that sold it to the press, but that did start a slew of new policies around accounting for work devices.
I had dev fused phones for open carry outside of the office while I worked there but had to register when I got them and when they were returned, which apparently didn’t used to be tracked so tightly until that incident according to my coworkers who had been there longer.
This is part of the reason I wanted to announce it even before I feel ready to ship... The frontier models have such crazy velocity (admirable) and moving so fast that I didn't want them to make an announcement and suddenly trivialize all the work I 've put into this, overnight.
The fun technical challenges (that can also act as any sort of weak moat) are being taken away one by one, on an almost weekly cadence now! :)
So, this is what I have been thinking (and part of the reason why I want to "cook" it more before shipping). Since I have been building it for so long, I am not sure how much "overfit" it is on my data, the way I speak and the phrases I 'm using.
And of course the opposite too, how much the dog trained me to speak to it a certain way to maximize outcome success.
But then I thought, when people play games they are not using highly sophisticated vocabulary and there is probably lots of repetition since they are always under some form of multi-tasking stress (playing and replying/speaking). So maybe... maybe, the system can adjust itself. Use a big LLM offline to say "user said X, we did Y - was that good?" - then retrain itself.
The decomposer is basically a bunch of old-school embeddings/classifiers stitched together, it can train super fast and doesn't need tons of data. Could the thing calibrate itself to the user? Does it even need to? (because as I said I 'm a datapoint of 1 and I am not ready for the potentially huge stream of bug reports when I ship (add some perfectionism to the mix and you get the idea)).
When watching the vids, I remembered talking (mostly text chat) in MMORPG's 20 years ago. Even when we started using Skype, etc., chatting was frequently short, slang, and quick info. Like, "pulling", "OTM", "BRB", "OOM". Or even just "let's go", and so on.
But with some practice sessions, I suspect this system would pretty quickly gravitate to using any players slang speech. Am I wrong?
Yes, the dog has a few levers including a "boredom" counter that will trigger it to do things on its own (like roam around, eat something, chase something (eg a rabbit) if there's one around etc). All connected with its current state (Eg hungry = will probably prioritize eating but not guaranteed). Other things are mapped to game events. Deterministic in nature but its underlying values can evolve over time (both its default "homeostasis value" but also the curve of which it increases). You can think of it as letting an llm design a behavior graph, and then playing that behavior graph until it changes again.
Talking to it alleviates pressure so that mechanism doesn't fire much, but in the void-mode video on the website, at the very end, it chooses to chase down an elk on its own much to my surprise.
After play testing it I had to tone it down quite a bit as it caused continuity issues...
In one scenario the dog was hungry so he went and ate something that was owned (counts as a crime attributed to the player) so we ended up in jail. Being transported to jail fired a new location event and a switch from "exterior->interior" which the dog responded to by having the llm author something along the lines of "It's nice to be finally indoors, it was getting cold outside". Which made for a very funny moment but a frustrating gaming experience.
I will be open sourcing it soon :) my code is a bit dirty (the whole system is 3 pieces. The game adapter / a websocket bridge between game and brain / the brain itself) and it runs on 2 different (local) machines currently.
Ale is what makes this work locally, I felt a little conscious about it as I am not sure if it is a novel approach or somebody comes out and claims I rediscovered BERT or something (though ale runs at 1/10 the cost of BERT).
Hello, author here. I was intentionally a little vague about this because this is kind of the thing that makes the whole approach work.
An LLM predicts the next token. If you're trying to predict the next token in a mathematics competition, or while playing a deep strategy game, being a much larger and more capable model helps enormously. To predict that next token correctly, the model effectively needs to model a bunch of possible future states - even if that is a second order (unintended) effect, it is what is seems to be happening.
This is basically the Ilya (and Dario) argument that prediction, understanding, and compression are the same thing (deep rabbit hole) from a few years ago.
----
In my opinion; this is a beautiful idea, but videogames do not need most of that. Videogames (and games in general) shine when character behavior is predictable, and when NPCs are a little dumb (just a little).
We already have very good small roleplaying models — Qwen 3.5 4B/9B/30B-A3B. Nowhere near frontier models at general reasoning. But they can act and write in a very engaging way. Good at roleplaying but very weak at reasoning. They just need a little nudge at reasoning...
And that's the key. The player has already expressed their intent: attack that guy, go look over there, cover me, find the key that shines and is golden, etc. A constrained world, with a constrained set of actions. Instead of asking the model to reason over an enormous space of possible futures, we're mostly asking it to map: player intent + current world state → a small sequence of plausible actions.
As for the "dump context to an LLM". It's basically. "You are roleplaying as X - you experienced Y - you like/dislike (dispositions) Z, you remember Alpha, your journal says Delta. Player orders you to do Gamma. - "What do you respond and do?"
It kind of works (as you can see in the videos I posted). I am not going against the grain, big models are better, but do we need those models for everything?
What I don't understand is how are you passing the world state to the model?
Say for instance when you ask the dog to do an action when you launch an arrow upwards how is the LLM continuously tracking the state of the game to be able to respond?
At that point it's deterministic and doesn't need the LLM. Once a plan has been decomposed and evaluated, it becomes a sequence of actions. Actions can either wait on a trigger or flow from one to another.
For the arrow example: wait here -> bind: arrow event -> pick_up $target -> goto player -> drop $target.
A plan can bind to multiple steps and actions. eg you can tell the dog, "when I say apple you say banana" this creates a short lived "player has said %apple%" -> say "banana" rule. Hide and seek does "say: count to 10" (this then gets fed to llm that converts it to "one, two, three"). So, the plan decomposition is fuzzy, but it gets decomposed to concrete steps, these steps might invoke the LLM back if needed so, but mostly for color, not logic.
Where the LLM is used afterwards, is for evaluating the result. The dog might ask for a treat if it got right or get frustrated is it misses a beat, etc (depending on emotional state, hunger levels, etc).
Basically trying to use the best of each system (fuzzy/vague/emotional for LLM, "hard" game actions for determinism). You can think of Ale's output as a small flowchart.
I am using a pretty dumb model. It's great for speaking, terrible for thinking, so there's a need for creativity
I think what's still a bit unclear is how the action space gets linked to a plan. Did you create your own action prototypes, or are you using ones exposed via an existing Skyrim world state interface layer?
How often are you running the world JSON through the planner? When do you give up on fuzzy matching if an action doesn't fit?
While I agree with the premise, I think this angle only applies on work one was going to do no matter what. The real power of these tools is that there are so many ideas people would like to try, but never have the time or motivation to pursue.
So the comparison is not only "built with and without LLM" but "would you even build this if you didn't have the LLM?". The gap in productivity in this case is much more wide.
> The real power of these tools is that there are so many ideas people would like to try, but never have the time or motivation to pursue.
This can be a negative multiplier: code I thought I wanted that gets immediately abandoned is a net-negative if no one else wants it (lets face it, this is the safest default posture for software of unknown providence).
In isolation, instant-abandonware takes up hdd space, burns dependabot's CPU-cycles, and wastes human attention when appearing in search results. In aggregate, it floods the zone with a deluge of forks with imperceptible differences between them, based on nit-picks, legitimate stand-out products will have a much harder time going forward.
Not every engineering effort is a product in search of market fit or a community.
Some things are already very useful as just a one shot. I just made a quick app to help me pack for a trip, it updated forecasts every day, let me know when rain entered the forecast at one of my stops and gave me a checklist that helped me quell my travel anxiety.
The greatest thing that LLMs have done is allow many to achieve things that they couldn't have before. I'm incredibly disinterested in "I can do the same thing I was already doing x% faster"
> gave me a checklist that helped me quell my travel anxiety.
it's truly fascinating how many positive descriptions of AI gesture at emotional management. I think that's the killer feature of this technology -- it makes people feel good, capable, reassured -- without the risk and vulnerability of interacting with another human.
I use Google Weather for my forecasts btw, no need to vibecode an app for that
Gathering forecasts for each leg of a trip only on the particular days for the particular location is tedious. Anyone can do it with google weather but why spend the time when you don't have to.
I can keep track of my expenses on a napkin but i'd much rather use a spreadsheet or dedicated app especially when that app is effectively free.
I asked Google Gemini to generate me a basic expense estimator to paste in ... Google Sheets. Every single formula needed editing. At least it generated the basic layout for me.
> Not every engineering effort is a product in search of market fit or a community
I think you misunderstood my point: by "legitimate stand-out products", I meant exactly that, with no connotation of commercialization. Maybe you can agree that having a high signal-to-noise ratio for (open source) projects is a desirable goal?
> Some things are already very useful as just a one shot.
I agree. I too have made or forked about a dozen apps and tools in the past few months. It would be dishonest not to consider the flipside, that this software is overfitted to the needs of a single person. Further, this hyper-bespoke software typically feature-complete within moments of the final prompt, and I have,on occasion, completely forgot about the tool/app I spent a weekend created, it clearly wasn't worth the effort I put in.
> The greatest thing that LLMs have done is allow many to achieve things that they couldn't have before.
I disagree; but see where you're coming from. I'm chuckling at the irony of my word-choice: I initially had used "project" but nixed it because of its frequent association with Open Source. I instead opted for "product" as a broader term. For the sake of clarity, my original comment is referring to commercial and non-commercial software projects/products.
Spending 1 day vibecoding or modding an existing app to fit you better and not using it is much better than spending, 2 weeks and then using it from sunk cost falacy.
Also it teaches you that what you think you need and want is not what you need and want. This is why you are not using it
> having a high signal-to-noise ratio for (open source) projects is a desirable goal?
It’s not obviously true. A higher number of attempts, a larger talent pool, typically doesn’t change the average much (or it might even make the average go down), but tends to produce higher peak outcomes.
We see this everywhere (science, startups, sports, chess, etc).
If you want the best spreadsheet, game, or whatever app you want, you’re only interested in the few highest peaks.
So you do actually get better signal to noise with a larger wasteland of discarded attempts. The higher peaks make it easier to filter out the noise.
The goal you’re intrinsically motivated by seems different than this. That seems to be the whole disagreement.
This is wrong. Someone else wanting it is irrelevant. If it solves a need you have, that is a win. Not every piece of software needs to be a market fit. Not everyone wants to grind out an existence marketing software. And rapidly that type of business is going to go byebye
> This can be a negative multiplier: code I thought I wanted that gets immediately abandoned is a net-negative if no one else wants it (lets face it, this is the safest default posture for software of unknown providence).
obviously not ? negative result is still a result, just like in science. It adds new information ("approach X does not work" / "is useless") which is the only thing that matters
You made me do a double take with "providence". I think the word you wanted is "provenance", which comes from latin "provenire" (pro -> forth, venire -> come). English has too many similar words.
Unless you only run code that's protected by God. That's probably not a bad policy if you can verify it.
Yes, it was meant constructively and for exactly this reason. I interpreted what they wrote as what they meant, then realized it was a mistake that I could have easily made too.
Me too, but we're the minority. Most people are rolling their eyes every time this happens. They don't want constructive criticism any more than they want normal criticism.
I think the ideas worth pursuing still require substantial time and motivation, even with LLMs. Otherwise you just end up with a half-baked sloppy artifact that nobody will ever use or study, for a net negative benefit to humanity.
We have entered the era of small software, and no one knows what it will grow to.
Here's an example: my office has a few cars for employees to use as rentals. The number is small enough that it would never be worth any serious software dev to build a tool to manage, but large enough that its a moderate amount of work for someone to manage the requests/getting supervisor approvals/schedule changes due to breakdowns.
AI one-shot that guy a tool. Its now dead simple, he's got a calendar, automatic emails going to people's supervisors with click-here-to-approve links, rescheduling options, fleet management. It doesn't even look bad.
Who cares if its using some un-backed-up sqlite database in the backend, has some placeholder tab for a feature he changed his mind about, or violates the DRY principles a bunch or uses some inferior authentication mechanism. Its an in house tool, isn't mission critical, and it makes his life significantly easier.
Basically everyone is now a few prompts away from their own bespoke tools, and only they will be able to judge the benefit thereof.
Edit: to tie this more directly to the article, I would argue that this is an example of "infinity-x" coding, because the user was in fact not capable of coding a solution on their own without AI.
In my business, I haven't found much area to use code. It's a pub, and we've long been low-tech. Cash register, no POS. I have a little code surrounding my own processes, but mostly it's manual. Hand-entering numbers in my spreadsheet, etc.
But what's interesting to me is that now I can probably program an esp32, or create a small mobile app for a mounted android tablet. I was a web dev in the past, and programming hardware was outside my skillset without dedicating some serious time to learning. Mobile I just always avoided--mostly the same reason.
Anyway, I've got some CYDs on my desk, we'll see what I can make with 'em. I want a kitchen ticketing system instead of the old hand-written ticket stubs, for starters.
Yes this is true to some extent. I've been using LLMs to run some computer vision tests, and I've certainly noticed myself running into the trap of "just one more AI experiment", or "just one more change" while neglecting to actually properly integrate the learnings into my mental model.
However, without an agent running its own experiments on a cloud GPU, would I realistically have invested my limited work hours and tried evaluating 10 different models, each with 10 different tuned parameters, to solve my specific use case?
Or would I have tried 1-2 models and spent my time trying to optimize those models?
I think there is some merit to the spray and pray approach when one is in the exploration phase of the solution space.
Also, on more than one occasion now, I have had fable halve the inference latency of a model simply because the original implementation from an academic included unnecessary GPU-to-CPU-to-GPU transfers or similarly inefficient operations. Those optimizations came at essentially 0 time cost to me and I can verify that the outputs are byte-identical. Pretty sweet!
People need to abandon the notion of every tool you build needing to fit the market. This is not the direction of software anymore when developers can all build tools that are bespoke to their own workflows and needs.
There were so many things that no one will ever study and won’t give humanity any benefit but I use everyday to make my life better. That’s enough. The value far exceeds $200/mo. I’m getting it for cheap and now that I have my GPUs and my models they can’t even take it from me in the future if they wanted, haha!
LLMs allow for human flourishing on a massive scale. One of the best inventions to occur in my life. Up there with the Internet/Web. Truly a marvelous time.
You know how non-coders basically build crazy business apps in Excel? LLMs, even if they don't get any better, seem at least 10x more useful than Excel. Why do you say it's for a narrow audience?
i meant people who enjoy the idea of software even if they don’t necessarily need or enjoy the resulting software itself. it’s possible that some of the spreadsheet hacking has a similar flavor but lacks the natural language hook.
In this last week, I've performed experiments in an almost fully automated way - parallel to my main work, and easy to orchestrate.
This has led to 3 parallel pieces of adjacent work that each speed up our build by quite a drastic margin. When combined, this is a massive improvement. None of this would have happened in the old days, as the research itself takes a long time to babysit and a lot of options to check.
So I very much agree - the activation energy can be a lot lower on some kinds of tasks, and some of those get big returns for small inputs. It's not all like that, but part of the game is identifying when you can spot those high return efforts.
or the AI could lead you to a completely unusable experiment and lead you to believe it was successful wasting weeks of work. Not saying it will happen to you, but there is a huge "survivorship" bias already in software for what tools are useful or not. Maybe it's the person and not the tools.
Yes, but only in the programming sense of productivity, not in the economic sense. When prices stop being subsidised and the novelty wears off, this will boil down to the question of how much money is that worth? We're already seeing an inflation in such projects, which are becoming hard to even show off. It could be similar to 3D printers. Fun and valuable, with a market size of $25B, but not quite world-changing (I'm not saying AI won't be world-changing, only that this tinkering aspect may not be).
Those sort of ideas are unlikely to yield major benefit, though.
This is basically replicating the plight of the solo open source dev, writ large. Individual programmers have long built the thing they've cared about on their own time (essentially "for free" because, despite kindergarten economics theory, a programmer cannot usually monetize a marginal hour). And it usually goes that the project never gets adopted anywhere. It might acrue more features and total man-hour effort than most of what FAANG does in open source to drown out the solo devs. But the market will decide that "no organizational buy-in" is a signal the project doesn't matter. Other developers will decide, "if he could do it, so could I" and also not adopt.
Same exact thing is happening and will continue to happen with all these generated "but we wouldn't have done it otherwise" projects. It's just very, very unlikely to go anywhere.
That which took very little effort to create will receive very little effort to promote.
I don't think this is the right framing, because I think it confuses the means and the ends.
Its like making a jig in woodworking. The measure of the jig's success is not whether it gets re-used or widespread adoption, its whether it made it easier to achieve some actual objective. Because the jig is a means to some other end.
Lots of these "we wouldn't have done it otherwise" applications are means, not ends.
That is true, but that suggests that the value of software is reverting to the value of a jig. Software that gets deployed to millions of people gets a large valuation because it provides small convenience times a million. In the era of bespoke software, you’re basically 100x-ing the production of a millionth of the value of commercial software. That’s why it isn’t all that impactful in the end and I think it’s worth highlighting.
LLMs open up huge possibilities to do work that you otherwise wouldn't do. I vibe coded a PDF viewer yesterday in a couple of hours using Claude and Grok. I've been itching for a native Mac app that meets some very basic requirements: one window with file tree on the left and tabbed view on the right, with the ability to call out to pdfgrep and jump to results. Like Zotero, but with better multi-file search and a viewer that doesn't choke on big files. Claude built it for me in an hour, and Grok iterated a couple of times to make it better. I didn't look at a single line of code. It's now my default PDF viewer, and I used it today to do a bunch of my real work.
It's a braindead simple program that mostly hooks together pre-existing functionality, it just so happened that none of the widely available apps had the specific mix of features I wanted. I could probably have done it myself in a week if I took time off my non-coding day job to figure out Swift and AppKit. But I wasn't going to do that. I’m psyched. I hate web apps and now I can just write my own for all the little things I use every day.
For me it is exactly the opposite. AI makes it so easy to create test fixtures and run experiments that I now spend much more time validating ideas than I could ever afford to do before. When I write code for production, it's not "this ought to work", it's "here are the figures showing how well this works, on this dataset", where the dataset is also much larger than anything I would have used before, because I used AI to generate the tools which collected and organized it.
> it's not "this ought to work", it's "here are the figures showing how well this works, on this dataset"
The thing with constraints is that you focus of the thing with high value first. So you focus on the most promising ideas first or choose experiments that can get rid of most ideas. Instead of trying to validate each ideas and generate what is most likely noise to the decision process.
Like if I ever hire an assistant, I want like one to three options that are closely aligned to my needs, not a bible size report on 42 choices.
I'm doing analysis on stuff that we previously simply couldn't do in my company, it would take way too much time or effort, and we didn't have the manpower.
Yes, I work in hardware design ans the cost to code up the RTL for an idea and compare it to the existing code in terms of PPA, its very expensive in terms of engineering time. This week I had 5 different architecites coded, pushed through a functional test bench, bit accurate models created and then run through ppa analysis all done via llms. So its a task tbis project could not afford to do without llms, but can save us 10% in ppa
Im not so sure if this is a pro or a con. Take two extremes at product building: brute force vs deep research first. I guess llms allow you to experiment quickly with many approaches and see what sticks. Without llms, you need to spend more time researching and thinking what’s the most economical approach. Im biased towards the non-llm approach
I often use LLMs to build locally running web apps and console apps that I otherwise never would have built, because the value-add wouldn’t be worth, say, 40 hours of my free time.
Or who remembers the dancing disease of 1518, were people would stop what they are doing and start randomly doing the same dance. The lords? Out of their minds. The priests? Terrified the devil had taken hold of the flock! I have come to believe that it was probably some tik-tok like hype trend of doing a fortnite dance while waiting in line for bread and communion. And the energy back then, like now, was off the charts.
Hype and memetic trend seeking encoded deep in human psyche.
reply