I guess it's a question of trust and maintenance, and bus-factor. Why would I trust a small one-person project (+agents) that might not exist in two months for anything serious?
Well yes, it's certainly a question of cost-value. Many cathedrals offer real value, but the ones that have used dark patterns and treat their users like blood bags can't be razed quickly enough.
I read that as abliterated at first, and realized that I have read too many AI stories lately, and also that abliterating maybe applies to human agency as much as it does to AI refusal...
That's fine, the bar the originals are setting isn't that high. Both Photoshop vibeslop clones I've tried launch and open a PSD instantly, while double clicking Photoshop is a "reach for the coffee" action.
Vibe coded alternatives have the potential to match functionality of their closed source originals, while being free and open source. LLMs don't get tired or bored, unlike human open source maintainers.
PhotoCraft repo is a couple of weeks old and in early alpha, and already seems to be significantly better than any other open source image editor I tried. And it also opens instantly, while both Photoshop and GIMP take 5+ seconds.
Neat! Independent, curated aggregations of open source software > relying on Google AI Mode / <some LLM> to index these projects.
Some feedback:
- The "Explore the growing wave of vibe-coded projects" pitch does not speak to me personally; I'm more interested in utility, but don't care about how the software is made.
- Consider adding an upvoting feature?
- Typing something custom in the "Alternative to" field of the /submit form should not require hitting tab/enter to avoid a mysterious error.
This looks & feels like https://www.opensourcealternative.to/, which (funnily enough) also has a broken /submit form that always returns "Application Error". I did not try working around this by paying for the $29 expedited review option, maybe that would work.
Someone steal this idea for OSS for vibe coders please:
Stop making full apps. Make modular, well optimized pieces of functionality taken from established apps others can pull into their vibe coded software.
I believe this is what Google wants to do eventually once they big data-crunch the most common vibe code cases, and the new OpenAI Intelligent UI seems to be this as well.
Let's end the endless churn of tokens for subpar todo-app back ends etc and make something streamlined for vibe coders that they can build on.
I suspect this site will end up similar to the google graveyard. The reoccurring theme with the vibe coded apps is the author gets a lot of press then it is abandoned once the bad architecture collapses in on itself
I think that what we really need is a global label that makes it easy to see if a project blocks AI contributions and modified software licenses that make it explicit that AI agents cannot use the code of software under that license. I'm up for that. It sucks to try to find a stable, meticulously handcrafted tool and getting spammed with vibecoded stuff.
On one hand installing these locally is a clear a security risk, but on the other hand, I guess we can now decompile and profile binary programs with AI agents them to determine malicious intent?
Even then, it's clearly a waste for each user to run such an analysis, so perhaps major one role of software companies going forward will be as trusted stewards or 3rd party auditors, if not as builders the software themselves.
I haven't noticed any specific recent Adobe stock drop that would be attributable to PhotoCraft (and similar), but the past 5 years have not been kind to the stock.
Maybe I'm old, but this new era of computing feels... nostalgic? Like we're back in shareware times all over again. The marginal cost of software development has fallen to $0 again in my lifetime.
So that means I have to review the whole app myself before I can put this into use within a company? Some of the ideas are good, but I need to find actual pre-vibe coded alternatives to these.
I see no value in specifically looking for vibe-coded software. The ones that are straight copies are going away soon anyway no doubt, once the lawsuits start hitting
Isn't the entire benefit of AI-driven software that I can easily throw together something that fits my exact requirements? Why would I want someone else's requirements?
I don't really see this ending well. I don't like most "hand-coded" alternatives. They usually "tick all the function point boxes," but with terrible usability.
But hey, I'm often wrong. Maybe I'll be pleasantly surprised.
Is this a joke? The only advantage of vibe coded software is that it can be produced quickly and cheaply and doesn't require specialized knowledge. Assuming cost is not an obstacle, you're better off with the commercial software. And if you want open source, just use a classical open source alternative. If it's missing a feature you want, then you can always vibe the feature into existence. At least then you're starting from a solid baseline.
Why would you not vibe code your own? The popular options theoretically have support, a community, financial backing, some non software service component, etc. without those why would you use someone else's personalized slop abandonware instead of diy?
People mix between vibecoding and using LLM in general, the first one you start prompting what you want as an end product and few prompts later you get something working, no track no audits no specs nothing, just code, and updates later will most likely cause problems. Using llm however isn’t necessary that, you can have your very well detailed specs and design architecture, and the LLM just turns those to code, trackable and auditable, plus easy to revise without losing the focus, it’s like normal old way but instead of the agile approach it’s now spec waterfall one.
Does the label "vibe-coded" even make sense if even in major organizations most of the code is written by LLM-based agents? When do you go from saying "prompt engineer" or "vine coder" to just a regular SWE using productivity tools like IDEs of the pre-LLM era.
Am I the only one who reads "vibecoded" as a negative word?
As far as I'm concerned, that's a term for shit software built by people using AI with minimal oversigt and quality control and often no idea what they're doing.
Using AI as a tool, carefully, would then be "agentic engineering".
Honestly, as of late "vibe-coding" has started triggering me very similarly as the "n" word (or the earlier "c" word for ai). It removes agency and skill from the developers who architected these solutions.
The cathedrals are burning. No doubt there will be collateral damage, but I think it’s a cleansing fire.
I built my own Jekyll markdown editor. Supporting anyone else using it would be annoying.
PhotoCraft repo is a couple of weeks old and in early alpha, and already seems to be significantly better than any other open source image editor I tried. And it also opens instantly, while both Photoshop and GIMP take 5+ seconds.
Some feedback:
- The "Explore the growing wave of vibe-coded projects" pitch does not speak to me personally; I'm more interested in utility, but don't care about how the software is made.
- Consider adding an upvoting feature?
- Typing something custom in the "Alternative to" field of the /submit form should not require hitting tab/enter to avoid a mysterious error.
This looks & feels like https://www.opensourcealternative.to/, which (funnily enough) also has a broken /submit form that always returns "Application Error". I did not try working around this by paying for the $29 expedited review option, maybe that would work.
Stop making full apps. Make modular, well optimized pieces of functionality taken from established apps others can pull into their vibe coded software.
I believe this is what Google wants to do eventually once they big data-crunch the most common vibe code cases, and the new OpenAI Intelligent UI seems to be this as well.
Let's end the endless churn of tokens for subpar todo-app back ends etc and make something streamlined for vibe coders that they can build on.
Minus points for Fullscreen not working in Firefox.
Even then, it's clearly a waste for each user to run such an analysis, so perhaps major one role of software companies going forward will be as trusted stewards or 3rd party auditors, if not as builders the software themselves.
But hey, I'm often wrong. Maybe I'll be pleasantly surprised.
As far as I'm concerned, that's a term for shit software built by people using AI with minimal oversigt and quality control and often no idea what they're doing.
Using AI as a tool, carefully, would then be "agentic engineering".
Now that AI has turned the cost of software to zero, this might as well be a malware distribution website.
There isn't a company or individual behind it, unsigned (Good luck on Windows) and obviously no tech support if it all goes wrong.
But I guess this comes with the 'abundance' of software.
Bait so obvious you couldn't fool a mackerel with it.