Average person also has heart disease, gets rugged with their privacy, gets scammed on phone calls, is stuffed with microplastics
Knowingly making bad products just because the average person can't tell the difference is selfish and a very FIAT way of doing busines.
...I agree, but the large majority of those do well or have done well in the fiat system. The believe they worked hard, made good decisions, spent wisely, saved hard and ultimately have a good life thanks to the government. There is no way those people will look past their bank account balance and high paying jobs to see their journey was a lie.
There is a difference between the person living in delusion and the person selling them the delusion.
I think with Nostr and bitcoin the risk is reduced since we donāt need to worry about auth or database storage or payments.
Don't think so, since your "encrypted communications" can be leaked before theya re even encrypted.
What are you encrypting?
NIP-17 DMs? Whitenoise MLS DMs? Tons of stuff.
Thatās just one aspect of nostr - kind of weird to point out one use case. But even with DMs - the way MKStack handles it is there are specific human vetted instructions for some components that the llm is instructed to implement - itās not yoloed. Not talking specifically about MLS - but other components in general. I suspect MLS will be much harder, but even that will eventually be a component llm can mostly copy paste.
I hope I understood it correctly from nostr:npub1q3sle0kvfsehgsuexttt3ugjd8xdklxfwwkh559wxckmzddywnws6cd26p ās citadel dispatch episode. It sounded like at least for some things Alex is giving the models very specific ways to implement base levels UX like auth
Send access to your encrypted forms, private posts, self notes, tons of stuff, only pointed out one usecase, since only one usecase is really popular and everyone understands it.
Also even with stacks the tools and the context is owned by the stack, the LLM training is not and no one knows what's going on their even if it's open sourced. There's no alternative to vetting the code.
A better argument would be to put in place automated vetting solutions. Though I have no clue what a trustworthy automated vetting solutions looks like.
You do realize literally no one outside of a few nerds vets any app code they use? Why would nostr be held to any different standard? People trust random apps from random developers but we canāt trust Alex to instruct the LLM to ācopy paste this codeā ?
Yes, but atleast till now they could trust the devs developing the code. Now they don't even know if the dev understands what they've built in the first place.
I trust that they vibed it well š¤£
I dunno man, I look at Nostr apps and they are not rocket science. Outside of handling messaging, itās mostly fetching and signing events to and from a front end. You can literally see whatās happening at the signer extension level and in the console. And the rest of the code is cookie cutter tailwind react stuff - not exactly anything to worry about.
I would only be concerned in privacy focused applications- like private coms or integrated payment handling. But zaps also straightforward and easy to understandā¦ even for non full time devs like myself
Probably you know something I don't, I find it very easy for someone, even an AI to add a trackertl to applications, logging things you don't want to log is surprisingly easy, just a few months ago a dev had logged nsecs, once knowingly, once unknowingly, and this is when it was not even intended, a carefully trained AI (not even prompted) could do it very easily.
Need a bit of code review for sure. I recommend reading your code - also not hard even if you canāt code from scratch. Lots of tools to assist with understanding code too now.
That's all I ask, vet your code, my fear is as "one-shotting" applications make it 70% usable, people just stop doing that, I don't think they do it even now. Those apps should be considered low quality, selling them as not low quality, because the average user doesn't care, is what is disingenuous and selfish.
Iāve vibed plenty of apps without review but always disclaim this up front (use at your own risk / new key). I think this is fine if you do this. Itās better to create new ideas even if they are low quality than not participate at all. You can always go back and get more eyes on your code if / when it gains any traction at all.
Yes, I've vibe coded tons of apps too, it's a great PoC generator, all the criterions you added are correct, if they are followed, which is a major caveat.
Also my post is neither against Alex not against putting nostr apps to any higher standards, it's a post against vibe coded apps and their poor quality in general.
It's not selfish, it's otherful. Very fiat indeed.
