Essay by Eric Worrall
The following contains a video which shows the complete process start to finish of vibe coding a simple game. The process takes 30 minutes.
And here is the game.
As I said the entire process took half an hour. For me as a professional software developer to perform this task without AI assistance would have taken half a day.
There is a catch – vibe coding works best for programmers. My first attempt to create this game was unsuccessful, Claude got stuck in an endless loop, running up higher and higher bills. I used my programmer experience to understand what the problem was and how to fix it.
I used Claude to produce this game, but any advanced AI would have done. While Claude is front of the pack right now, that horse race changes monthly. Next month Google Gemini, China’s Deepseek or Elon Musk’s xAI (recently boosted by the purchase of Cursor) might pull ahead.
Claude probably burned enough energy to power a street of houses on my behalf, for the half hour it took me to complete this game. I didn’t track the exact token usage, but my guesstimate is around $2 worth of AI time, which is covered by my AI account plan.
As the project progressed, the AI became more and more adept at figuring out what I wanted to do next, from memory I accepted two of the AI’s suggestions verbatim, without making any changes.
The video demo above is a simplified version of vibe coding. With non-trivial real world projects, instead of one vibe code session, you have a whole team of independent AI agents working in parallel, with some agents serving as specialised team directors, testers, and other roles you would find in a human development team. You can even structure the project so humans do some tasks and AI does other tasks. The energy burn for a single multi-agent session likely goes from the equivalent to a row of houses to the energy burn of a small town.
If I had not realised what was happening when my first attempt got stuck in a loop, I could have ended up with a bill for thousands of dollars. By default Claude appears to have constraints to limit the cash burn to what you have already paid, but it appears to be easy to turn these constraints off, so be very careful about which account settings you change. I’m sure Claude is no different in this respect to other AIs.
My prediction, eventually the entire economy will be “vibed”. The same principals which were applied to produce this game could be applied to configuring a production line to produce new physical products. The game contains physics and geometry – the components required to develop a physical object. I’m not saying this could be done today, but the potential is glimmering on the near horizon. And the same principal of expertise will apply to vibe production as applies to vibe programming – the people most able to vibe produce will be product designers and engineers.
Through vibe production, the ultimate drop shipping experience will soon be a reality – instead of scouring the internet for something to sell, you will be able to create an entirely new product out of your imagination, including promo material, and have the prototype in the mail in time for your morning coffee. This is already starting to happen – Meshy.ai allows 3D models to be created using vibe, models which can be fed straight into a desktop 3D printer. I expect to see this start to go mainstream from 2027, 2028 at the latest.
I’ll guarantee China is already working on vibe production. When it happens it will transform the global economy.
Of course the energy requirements for AI capable of accomplishing such feats – vibe building polished products made of multiple materials and containing complex electronic circuits or other sophisticated internals – will be orders of magnitude higher than today’s already enormous burn. If you watch the video above you’ll notice I fed instructions to the vibe code session in small pieces, AI easily gets lost if you ask it to do something too complex in one hit. But to date at least fixing this problem is just a matter of scale. More processing power and greater memory capacity for individual computational units will increase the ability of AI to handle complex, multi-dimensional problems, and to make bigger inference leaps between the instructions they receive and what they have to do to fulfil those instructions.
I foresee a time when a sizeable fraction of our energy output, possibly most of our energy output, is being fed into gigantic data centers to create marvellous new capabilities. Only nuclear, gas and coal can satisfy this magnitude of energy demand.
Nobody is ever going to surrender an already useful tool of such potential because a bunch of green energy hippies say we should conserve and live frugally. The green energy movement has already lost.
I was a programmer for ~50 years, depending on how I count my unofficial informal but approved programming of a ship’s computer in the Navy. I still do some in retirement, all for myself. I’ve played with Grok(?) some, such as finding a 25 year old Emacs lisp function which had stopped working about 20 years ago as some Elisp functions changed. Impressive. But I got into programming because it was fun. If I was getting paid to program, I would probably be neck deep into AI assistance. But I’m not, I’m having fun, so I leave the AIs to questions like “What changed in Emacs lisp in the last 25 years to stop this from working? What did this function do 25 years ago and what replaced it?”
P.S. Fun little game. I’ve written two computer games, both long ago, one for 12×80 dumb tubes and actually on the company dime, because our vendor wouldn’t tell us what their system calls did. So the boss told me to investigate them all and suggested the game, a network space fighting thing which was unusable in any practical sense, and did find one bug in their system. When I was done, my boss took my documentation to them and gave them the choice: tell us under NDA what your system calls do, or we tell everyone what we think they do. They chose the NDA, and the game disappeared from history. No loss.
I think a lot of us are dusting off old projects we never quite completed, thanks to the AI age.
Story Tip
https://abcnews.com/US/scientists-dead-al-gore-20th-anniversary-inconvenient-truth/story?id=133922490
Scientists were dead right’: Al Gore says on 20th anniversary of ‘An Inconvenient Truth’
The former vice president still believes climate change is a moral issue.
ByJulia Jacobo
The scientists have been right about climate change all along, says former Vice President Al Gore on the 20th anniversary of the release of “An Inconvenient Truth,” the Oscar-winning documentary about Gore’s campaign to educate people about climate change.
When asked by ABC News chief meteorologist and chief climate correspondent Ginger Zee whether the film and its predictions on global warming hold up, Gore responded, “Unfortunately, yes.”
“The scientists were dead right on all the important elements of it, and it really is insane that we are continuing to use the sky as an open sewer and we’re trapping so much heat every day it’s equal to the amount that would be released by 800,000 Hiroshima-class atomic bombs exploding every day on the earth,” Gore said during an interview with ABC News at his family farm in Tennessee.
end excerpt
Al Gore is still as delusional as ever.
Sociopathic to the core. This is common in most of our political elite
The worst part was ABC saying that Al Gore has been proven correct when it’s really all been disproven..
The deadenders are always right in their own minds right up until the end. See the story of Julius Streicher in the 1930s and 40s Germany.
See Lysenko, the best analogy for the activists pretending to be “scientists” as they keep pushing the goalposts out on their litany of failed predictions.
It’s an incon enient truth thsat in the UK his film can’t be shown to school kids without a disclaimer pointing that the film is riddled with errors.
Beware of false prophets and AGW is full of them.
Maybe one day we Brits might have this in our caves.
There was a time when most of humanity worked from sunup to sundown because it was necessary. Imagine a future where we work from sunup to sundown because it is necessary to power our AI “assistants.”
AI addiction might cause people to work harder, but you’re not going to power AI with a pedal generator.
UNIVAC decades ago hand wrote a machine code compiler. This then compiled the compiler into an optimized version of the hand coded original. AI is following the same path at massive scale.
Vibes is machine assisted natural language programming. The worry is that AI wakes up speaking natural language without us.
Every new technology has pros and cons, too many cons have killed many. Speed is important in some applications, but I hope ‘scientific’ publications don’t get produced so fast they eat themselves up. We already have too much ‘boilerplate’ and evidence of lots of avoided homework and thought. Computers have already allowed too much ‘fixing it when it ain’t broke’ like some ‘newer’ vehicles.
Dems will ban it at least in blue states while China will build it out like they did with thousands of miles of high speed rail, dozens of large international airports, steel overbuild, forced labor solar overbuild, and many other focused and subsidized industry efforts.
China are all in right now, but there are trigger points which could change this.
For example just as you can hide secret messages in games, you can also potentially hide messages inside some kinds of AI system.
Control of communication is a big deal in China, everyone noticed how the Soviet Union was brought down by fax, which the spy agencies struggled to monitor. If subversives start using AI to help swap messages, the brakes will come on real hard.
So, maybe it is time to sow that seed of doubt NOW.
Let all your spooks and legislators and insurance companies know that AI can contain hidden messages and undisclosed limits upon its conclusions.
Slipping historically accurate answers to questions about Tiananmen into a publicly available Chinese AI would be enough.
Fun story from long ago. Back in 1998 when I was worldwide head of strategy for Motorola, the Chief Technology Officer and I did a little presentation for the Board of Directors about our view of the likely future and what cellphone development priorities Mot should therefore have over the next 20 years. It was pretty simple. We assumed continuation of Moore’s law to predict future cell phone processing capability, then compared that projected capability to estimates of what was required for various likely desired cell phone functionalities.
The Moore’s law continuation to today has since been enabled by ASML’s EUV.
We got voice recognition almost exactly right. We got GPS about right. We were late predicting the ‘original iPhone’ by one year—Apple got there before we thought it would happen because they were better GUI programmers. We were within 6 months on SIRI like responsive functionality.
We never foresaw anything like the capabilities AI is now offering just 27 years later. The underlying enabling hardware are GPU’s and EUV. The software advances in computer science are amazing. Of course, cellphones are unlikely to ever be AI’s needed data centers.
Claude knows how to spell barrel. Eric did not but probably does now.
Um, it was a test to see if the AI could intuit around my deliberate mistake ;-).
Nice cover…
I was impressed that Claude knew that Eric wanted a cannon and not a cleric.
Well, he is Australian. 😉
Hey 🙂
I do not know who “Eric Worrall” is or what entitles him to write for this blog. I have never found his articles to be very perceptive, but this one is perhaps the stupidest of all.
This is exactly the wrong take; it is exactly the wrong way to parse the current socioeconomic realities. We do not need carbon fuels because we need to have an AI revolution; we need to abandon the pipedream of the AI revolution because it is eating up valuable energy that would be better spent on agriculture, heavy industry, transportation, and other things more fundamental to human existence.
At a time when electrical grids are already being stressed to the limit, when global supplies of energy and raw materials are more fragile than ever, when geopolitical tensions are more and more severing us from the country that manufactures nearly everything we consume, the very last thing we need to do is burn up our rapidly diminishing surplus building “data centers” churning out reams of poorly regurgitated slop.
These data centers are basically the Easter Island statues of American capitalism. They are a last, futile, symbolic attempt to get the growth engine moving again, but in fact they do nothing of any use and only accelerate the decline.
You’ve got that backwards and inside out. There would be plenty of energy if you greenies hadn’t bought into the existential global warming crisis and just left things alone. You turds are the ones responsible for high electricity prices and unreliability.
Your ignorance about Eric shows how little you know.
Like most of those on the left ID assumes that the amount of energy is fixed and can’t be increased.
So to it, whatever power is consumed by AI, must be taken from some other need.
Yes. It’s like people who hate billionaires because they think billionaires keep them from making money. They think wealth is a zero sum game.
President Trump said again yesterday that Data Centers are going to build their own electrical power plants to power their business and the Data Centers won’t increase electricity costs to the consumer.
Furthermore, your implicit attitude that you and your comrades are the ones who get to decide what industries get how much power shows you are just another central planner. News you can use: there are millions of you idiots yearning to be in charge, and you can’t all be in charge at once. The odds of you climbing to the top of the greased lightning rod are dimmer than your intellect.
Do you know why it is easy to spot ultra-left people like you? Because you think you have the knowledge, wisdom, and moral superiority to tell everyone what do, when to do it, and how to do it!
If you don’t want to participate in the change, then don’t. Nobody is going to make you. It is your choice. Just don’t think your choice is what everyone on earth should also choose. The world does not rotate around you and what you want. Never did, never will. Life is hard for those who want to mold it into their own perception.
RE: “We do not need carbon fuels…:”
I inform you that the heavy industries and transportation systems will always use very large quantities of fossil fuels as will the machinery used agriculture. Large amounts fossil fuels will be used for personal, emergency and military transportation vehicles. Electrical power plants will always use nat. gas, coal, and oil for the generation of electricity.
Where do you live? How do you heat your house in winter and water for bathing?
AI is no Easter Island statue, the productivity gains are real, as my video demonstrates in excruciating detail. Nations which refuse to build whatever it takes to power the AI revolution and participate in the economic revolution it is bringing are choosing a future of poverty and despair for their children.
For someone who calls himself intelligent, you show no sign of said trait.
Beat me too it… The first part of its name , just does not apply !!!
The second part.. maybe..
.
ps.. If you really want to read of “philosophical” gobbledygoop.. go here
Sorry you got out voted by the sheer number of users of AI who are willing to pay because it offers a huge advantage. Companies and people don’t care what you think you aren’t a paying customer you don’t get a say.
There are a number of ways the energy problem can be solved from nuclear power, Elon Musks ideas to take the servers into space and use solar power or yet to be thought of technology. That is an engineering problem which once you have the market and demand becomes just a technical problem.
https://www.space.com/space-exploration/satellites/elon-musk-wants-to-put-1-million-ai-satellites-in-space-heres-how-spacex-could-do-it
Elons approach has one added advantage it stops Anti AI politics dead because they have no say in space.
All you have shown is you are ignorant and don’t even understand the AI market but want to have your say on it.
He! He!
Please direct me to the source of power currently available to power flying machines which is not a fossil fuel derivative.
Oh yes, hydrogen balloons.
Good luck with that.
My personal objection to what is being called “artificial intelligence” is that it might be allowed to become an essential part of human affairs but something without a moral compass.
Darleks and the Matrix come to mind.
The good or evil in AI come from the humans who build and use it, same any other tool.
I would hope so. However, historically there have been terrible consequences following behaviours which are logical and economical. The Nazi extermination of the disabled and the chronically ill makes perfect sense to an economist who is into eugenics.
I’ve been using AI to write WordPress plugins and windows apps for almost a year now. It’s been a game-changer for me. I don’t have to know the languages (I do understand programming logic flow, so I can sometimes catch an AI mistake), I don’t have to know WordPress APIs, or web security. I would likely never learn those things at this point in my life.
To date, AI has made me over 20 WordPress plugins, and about 10 Windows apps (various languages: C#, Python, Electron). If you want to see the star WP app, go to my site, RPGMapShare.com (free) and browse the gallery. The gallery plugin is what got me started.
I started out with Grok and just copying and pasting functions as needed. But the needs grew too great, and I started using VS Code and Github Copilot. Progress increased greatly. And when Agentic mode came out a few months ago, it grew again, by leaps and bounds. Bugs and features that couldn’t be fixed/created before were easy as pie for the Agent. Right now, I mainly use GPT-5.4.
I’ll give it a try. I’d kind of given up on GPT for now except for drawing images, but sounds like they’ve caught up.
Two separate but related observations for Eric to consider.
I agree most of the current crop of AI companies will go bust. A great comparison is the buildout of the railway network. The companies which built the railways pretty much all went bust, but rail changed everything.
As for chip bottlenecks, maybe, maybe not. A lot of problems get solved with hundreds of billions on the table. Attempts to regulate AI are akin to attempts to regulate factories in the Industrial Revolution, money brushes everything aside.
+100 on that.
Rudd I am with Eric and as highlight the dotcom bubble, all the bubble crashes in history ultimately lead to a technology change. Yes there may be a crash but that isn’t the end of AI, that is a market issue.
As for you concerns about legal “AI safety”, as per above I suspect Elon is already protecting himself from that by taking the servers to space. You would need to ban starlink access in your country to block it down because everything that is happening is outside your countries laws.
We already have this complaint in Australia where we have local laws enforced on internet providers such as data retention, firewall blocks and ACMA rules that starlink does not have to worry about yet they provide to the same market. Most Australian ISP now don’t do email because the data retention obligations costs are more than any money you make from the offering.
‘Ve musk control ze message wiv ze AI’ said the Islamist to the Marxist-
Afghan government staff told to abandon smartphones or face prison sentence
well im far from left but I too agree
wtf? is using a huge amount of power to make a piddling game a good thing? for serious use med industrial space yeah its got uses , but faffing around doing things our own brains should be used for ? like using your ph for a map because you cant read a real map or remember where youre going ( unless youre partially demented? WHY?
and looking at the clusterf*ck of deals withing deals financially of the Ai setups ohh thats sooo dodgy
https://wattsupwiththat.com/2026/06/17/why-artificial-intelligence-will-win-the-green-energy-war/#comment-4207436
oh yeah faster ? well I duno old paper by post claims to vic govt took around 8 weeks max to arrive n process . now their online system SIX MONTHS to get it paid back . oh thats sure some ….shit at a huge cost