[jcgl]

Technical people, maybe. Even if we hypothetically grant that vibe-coded backup systems can be trusted, deploying, using, and maintaining are still hurdles. And there’s absolutely no way that’s all tractable for the vast, vast majority of the population.

Mind you, an average person probably doesn’t even really know what a server is. And these average people need backups just as much as (probably more than) people technical enough to vibe-and-deploy.

[mannanj]

I think that product people are very much doing development with these tools; and trusting, deploying and maintaining them. The vast majority won’t even know the trust, deployment and maintenance challenges that technical people know and so they won’t worry about them and probably never will need to. Anything very risky will be eventually protected by exposure in the training data. Not unlike how you see in lovable like building tools having built in scan tools.

The training data for these mechanisms is growing. As a technical person I deployed a clients site to lovable and now he maintains it. Anything risky? I trust lovable will surface it, and he’s comfortable with additional risk, and platforms like lovable may want to offer contracting help from humans In their platform in The future to help.

Why does someone who says “I pay -$x for Google Drive, how much can we make our own clone for?” need to know what a server is? The natural language doesn’t necessitate the technical jargon.

[jcgl]

Even just by narrowing to some definition of "product people," you've cut out the vast majority of the population.

> Why does someone who says “I pay -$x for Google Drive, how much can we make our own clone for?” need to know what a server is? The natural language doesn’t necessitate the technical jargon.

My point wasn't specifically that they need to know what servers are. Rather, the point is that the average person who needs backups doesn't have the faintest inkling of the smallest fraction of what goes into a backup solution. For these people (the vast majority of the population), you'd need a truly new level of vibing than what exists today.

Looking at Lovable's homepage, their tagline is "Create apps and websites by chatting with AI." Your average person probably doesn't even understand that "apps and websites" are primary components of a backup solution.

https://xkcd.com/2501/

[mannanj]

Yes I get that. And yes I did narrow it in by saying Product people. My point is why would anyone bother to know what goes into building x (a backup solution for example) if building it doesn't require that?

Using me as an example, or my non-technical client now doing app development using lovable, we don't care to know how the things being built but just that it works and its reliable and consistent enough.

Less than a year ago I'd never built iOS or Mac apps, and now I'm building those - I come from an HTML and frontend world, not the Mac or iOS screen layout world and I still know nothing about that and don't want to either.

[jcgl]

And I'm glad that works well for you! But this is the context of the thread:

>> most people don't need cloud backups

> What world do you live in?

As put by another commenter:

> Most People™ do not have the technical knowledge to set up a self-owned backup system, and/or will not realize how badly their future selves will wish they had backups if the setup friction for a self-owned solution is too high to conveniently do it right this second.

So, okay, if you're just saying Some Small Percentage of People™ can vibe up a long-term backup solution, then, sure, I grant that.

I just don't think it's a useful or interesting point in the context of this thread, even if it's true.

[mannanj]

No I think we are taking about similar things, but the core of it is this:

- you and other guy believe non-technical people can't build technical things because its still too hard and you need to know the technical things

- I'm saying you don't now, and some the samples I showed are that we are moving in that direction and its possible now, even if most people aren't doing it now. that they could. and its an awareness and comfort problem.

is this right or am I still missing your point?

[jcgl]

Thanks for trying to summarize.

> - you and other guy believe non-technical people can't build technical things because its still too hard and you need to know the technical things

Pretty much. Though maybe I should more softly phrase it as "you still need to have some idea about technical things." Even more importantly, you still need to have some idea about decomposing problems in a technical way.

What you said:

> the samples I showed are that we are moving in that direction

> and its possible now...its an awareness and comfort problem

So which is it? Are we there, or are we only moving in that direction? Core to my position is the idea that there is a long tail of competences that technical and "product" people have or can fill in the gaps with. Other people don't have these things that are needed to smooth out what AI misses.

Said another way, even if AI has made it 90% of the way there, that still leaves the remaining 90%[0]. Long tails are long, and smoothing everything out such that a truly average person would feasibly whip up a bespoke, reliable backup solution really relies conquering so many edges that technical and product people would never even notice.

[0] https://en.wikipedia.org/wiki/Ninety%E2%80%93ninety_rule