AI is just a tool like a hammer or sledge hammer. If you use one, who is responsible if you hurt your finger?
I think the issue is that the ones getting hurt aren’t the vibe coders themselves but maintainers, users, community, etc. who are suddenly flooded with low quality apps.
“A knife is also just a tool”
Draw your own conclusions…
Not really. AI is a tool like a hammer with a grabber of your knowlegde. And a grabber for third person data. As also said here.
If you use a good tool you are more experienced as before. If you use a AI tool the AI tool will be more experienced after all. But it’s so easy. Relax!
And if you believed all the AI slop maybe you are more .. (
)
Although I dislike AI per se (for rational reasons) I keep coming back to the human problem when it’s used in coding:
Do they even understand what they’re doing?
Does this create an athmosphere where people expect apps to pop up & mature in a matter of weeks, instead of years?
Do AI devs (for lack of a better word) have the patience to do what is needed to improve their app?
One thing I’m hearing quite often with AI devs - often as a reply to why they use AI assistance in the 1st place - is: “well sorry, but I just don’t have the time”.
Something I never heard from developers before, at least as far as creating the software itself is concerned (I have heard it many times as a reply to people complaining about “bugs” without putting any basic effort first, or demanding features).
And where did the very common statement “it’s ready when it’s ready (don’t push us)” go?
That’s where I see the dangers of AI/vibe coding. The expectations - which have already been too high for unpaid FOSS devs - are getting even wilder.
The flooding of the zone.
A new normal, where anybody who needs a year to bring some software to maturity doesn’t even show up in the landscape anymore.
Testing the Driver app from OpenRepos.
The “Developer” apparently did not even bother testing the app, because how can one miss the placeholder text staying in place in the input fields?
We need an AI filter in OpenRepos and Chum now, please!
This would be totally overshooting. It is not a black and white issue. As an experienced developer AI frees my hands/head to invest more time in more testing, to start with TDD right away, in better structure, decoupling, etc. all things that increase quality.
I do not mind if AI is used to build apps, I trust the developer. I am very grateful to all developers developing apps too. Trust is difficult to gain, easy to lose.
While I think such a filter should default to “ON”, it makes sense the user should be able disable it, of course.
In all fairness, it’s not only AI coded apps in my experience. Updates and releases to OpenRepos tend to go out quite untested regardless if it was coded by human or AI.
That’s a naive crutch. The problem IMO is not AI, it is developer mindset.
All the ethical problems with AI aside, somehow this did not happen that frequently before the AI bubble.
In some cases ai brings more quality into it… It all depends how experienced is the guy who use ai and the user have to decide in which cases its ok for him to use ai coded app..
Would I use a vibe coded password manager from a random guy? Of course not. Would I use a vibe coded app for a simple xyz task that someone pushed to openrepos? Yes I would at least when the sources are available publicly
AI is here to stay and I think best scenario would be good developers which usin ai in an appropriate and responsible way
What problems? Local AI does not take part in wars. As I said the problem is not black and white.
PS: Sharing is caring
You’re technically correct. It’s not all AI apps and not only them.
But before AI we had maybe 1 or 2 bad apps every now and then. Now it’s like the majority of new apps.
Then why not give users the choice and let everyone decide for themselves?
Option to choose the shade of grey?
I’m all for choice. I just recommend against making no AI a default, because users will miss out on a lot of high quality software. AI is not going to go away and with such defaults you only further hinder growth of the platform.
Somehow I see AI like an amplification. If the dev uses it with a clear conscience and knows what happens, the result will be fine. If he has no clue what he does, it will perhaps works. But it’s not viable in the long-term.
Most of the devs now inform if they use AI or not and that’s the best thing we can have.
Taking aside all ethical issues, AI is unfortunately an integral part of our lives, even if the current situations is not viable.
