I wonder if we will get to the point where this is all so easy that we'll need something like Flickr to organize all the random things one has vibe-coded.

It's already starting to feel like the days when your uncle would make you sit in a dark room and look at his vacation slide show. Sure, they're fine photos, but we quickly got to the point where a photograph was so easy to create that it needed to be extraordinary not just in execution, but in vision, timeliness, originality, etc. And even then, if you couldn't get it in front of the right editor, it just sat in a pile.

It is amazing, that just a few years ago this would have been a pretty amazing, portfolio-worthy project and today it's "pretty cool", said with a shrug.

It has to do with prompter. If you goal is just the result, make more sense for the AI to do that better. If you structured your prompts, then the resulting code is closer to how senior engineers will write it. Decades ago, probably before you're born, a lot of shareware codes are written way messier than what you see with AI.

I've been seeing lots of demos like this, especially with games. Yes, they are impressive, but I'm surprised nobody is talking about some of the things I've noticed with GPT-6: the code it writes, by default, is surprisingly messy and obviously unmaintainable. It's also very...not human. No human would write code like it does. I get the sense that the optimization of these models on benchmarks is causing their behavior to morph into a very brute-force type of approach. It makes you think: when code is suddenly very cheap, does good style and organization even matter as much? Or were those just important for humans? (I think they still do, for the record).

It matters, for now.

- I prompt better when I can understand the code, when necessary

- I can catch wrong assumptions it made more easily

- Good code helps the AI too, AFAICT

Wonder if someone might use an LLM to specifically vibe code a program that takes LLM code and makes it appear like non-LLM code.

I haven't looked at the code honestly, but it's a interesting observation. My impression of the way OpenAI is going is instead of making a larger model smarter, they're going wide with just extremely high effort and a lot of parallel agents to get towards a goal.

For example, for design and architecture, I prefer fable, but for pull requests I like GPT because it's so pedantic about things.

I hope the future generations of models will stop with the fluff/textual clutter/overexplanations - I don't know the right term. They easily fix those once you tell them, but they still do them by default. What I mean on this page, for example:

> A WORLD, FOUR BILLION YEARS IN THE MAKING

> DEEP TIME · CONTINENTS IN MOTION

> Distance becomes an ocean.

> Continents drift apart. Shallow seas spread across land that will one day be dry.

> EXPLORE 4.54 BILLION YEARS OF EARTH HISTORY

> EARTH THROUGH TIME

> Scroll to travel through time Drag to orbit

> THE PAST IS ANOTHER WORLD. IT IS ALSO OURS.

> ILLUSTRATIVE TRANSITION

> Ga = billion years · Ma = million years / Event spacing is not linear

It gets to the point where if you ask a model to make a tiny toy "OS", it'll write stuff like "REAL MODE. REAL MULTITASKING. DRAG WINDOWS / TAB TO SWITCH" directly in the background of the desktop of said toy OS.

I can’t stand it, one of the most annoying tics to me.

Funnily enough even the image model does it: gpt-image just throws little bits of text all over the picture wherever possible.

It's a landing page trope, and I wonder if it emerges out of overtraining on the front door of web sites.

- poetic invocation of awe

- rhetoric of universal humanity as backdoor to software anthropomorphism

- commercialistic exposition of features (in sentence fragments) and foolproof instructions for use

"Put 'em together and wadda ya got?" TURN OFF YOUR BRAIN

A good 6th prompt would have been "remove the title/subtitle idiom wherever it appears in the application"

The term that Anthropic is now using is "mannered prose". If the creator of this example simply prompted "Remove all mannered prose" then the entire experiment would suddenly become normal sounding. In the fable 5.1 prompting guide, they have a longer prompt for removing mannered prose too if needed. I find it works on Astra as well.

https://platform.claude.com/docs/en/build-with-claude/prompt...

it's flimflam. at least that's what i call it, and banning flimflam in the strongest possible verbiage seems to help

I’ve only heard the term flimflam as a name for a scam that involves tricking a cashier into giving you a bunch of money with confusing overlapping talk.

It looks cool but I just can’t be that bothered to engage with something that contains so little of a human’s input.

If one wants human attention, one must put in human effort

Head, hand and heart are all required to make something worthwhile.

When I see things like this, what I'm most interested in is seeing the full conversation/prompts and their associated changes/diffs.

I'm more interested in seeing how others are prompting and working with LLMs than the result itself.

Care to share?

Not the OP project, but I made my entire 731 prompt Claude Opus 5.0 transcript searchable and organized at https://vertumnus.fyi/about/transcript. The chapter section has all of the prompts chronologically. The entire text is about ~2 MB and can be downloaded at https://github.com/perspectivezoom/vertumnus/blob/main/data/....

This is objectively wrong. Over the north pole, Greenland looks as large as North America does when it is the center of the viewport. This is some strange mercator-esque projection applied to a globe.

That's the ice cap.

You're right. My mistake.

Don’t worry it’ll be gone pretty soon.

At some point I wonder where and when AI will just be taken as a means of exfiltration of knowledge from the traditional cathedrals/bazaars/silos - I mean, obviously, the more that knowledge is shared the more valuable it becomes .. but couldn’t this also mean that the training data included apps like this?

A giant for-profit corporation based in Silicon Valley, funded by titans of finance, and in bed with the government is about as "traditional" as it gets.

Garbage. There is one good joke there though: "created by Ethan".

The 'Civilization' view took non-Eurocentrism to a new level.

Very cool. I built something similar to simulate climate change... my first project built exclusively via prompting.

https://terrashift.io/ (it doesn't work on mobile right now, lame I know)

What a fun time to build!

This does not simulate climate change and you should not describe it as such

How would you better describe it? It's admittedly vibe-coded and not my area of expertise but I explicitly prompted Claude to build a (toy) climate change simulation. As an avid skier I was specifically curious how temperature change might affect localized snowfall.

I think I would describe it as a visualization of how sea level, snow cover and vegetation will change given a change in surface temperature. The trouble is it is not an accurate visualization of climate change . I am on mobile now, but if you would like I could offer more detailed scientific critique later

I hope I have not come across as too snarky. I am a climate scientist and modelling and visualization are a passion of mine so I am excited you are interested in it too. I just am a stickler and think for climate change in particular it is essential we communicate the best possible information we have to the public. I worry that projects like this can detract from more accurate scientifically grounded projections which climate scientists have spent decades on.

Have a good rest of your day.

Here is an online alternative to visualizing climate scenarios: https://mscm.dkrz.de/overview_i18n.html?locale=EN

Here is a sea level rise viewer for the USA: https://coast.noaa.gov/slr/

Appreciate your feedback, fair points and I'll check out those links.

For anyone wondering, it's 2.4K lines of code. Very small, but yes AI is very impressive.

Literally can’t see an earth on safari iOS

Edit> on second refresh it worked

Works fine here.

I’ve wanted something like this for a long time. However, I wanted it with extensive fact-checked information and AI slop is the opposite of that so I feel this is kinda pointless.

Yes, but ... did it ever occur to you that similar sites in the pre-LLM days made as passion projects by one person likely contained a similar number of errors?

they did not, in fact

Hard disagree. It was common for people to point out their flaws here on HN.

I thought this was super cool…

And then I clicked the “Explore” button on the top right and there are even more things that you can see! Pretty neat.

There is an old site that offers the same functionality: https://dinosaurpictures.org/ancient-earth

(Use its dropdown.)

Was it before or after GPT-6 got nerfed?

Publish the prompts or re-run them today to see if you get the same quality result.

before (rip)

is this for real? 5 prompts? this is pretty cool.

Do you have any basis for this collection of claims? I've never heard of this intentional obfuscation.

yeah, but is any of this true?

Did any of the prompts involve the user trying it and observing that the scrubber barely works and that the topography texturing (the displacement map or normal map or whatever technique GPT-6 picked) is outrageously wrong? I'm no expert in tectonic movements, but I guarantee that continental plates don't slide around without the mountains on them moving along with them.

It is shiny, though.

(I have personally once vibecoded an interactive 3D viewer. It, too, was a piece of crap. However, I did it with the full understanding that I was making it to solve one particular problem, once, for myself, and that I would just throw it away afterwards after the problem was solved. I used GPT-5.5, IIRC, which was capable of doing the math I needed correctly, but only after a lot of tries and my prompting it with how to do the math. I would like to imagine that 5.6 or 6 would have struggled less with the math.)