There is probably a huge chunk of people who only need a slither of the features of a SaaS but are paying for everything they don't need.

If your particular use of a SaaS is not susceptible to security issues (for example you can use it on your local laptop) then a slither of features that are insecure is exactly the thing for you.

Not everyone is going to replace SaaS with half-baked personal implementations, especially the big companies that most SaaS are aimed at, but it will gnaw out some of the long tail of SaaS subscription revenue for sure.

> slither

What is a slither in this context? Or should this be "sliver"?

Reminds me of the days I used to spend building my own Linux kernels or configuring X rather than paying the "Apple Tax", thus costing myself hundreds of hours of productive work time.

Well done choosing to invest in your knowledge and neurones rather than in investors wallet :)

Sliver!! Snakes slither, cakes sliver.

“You too can implement a slither of the features you need right now in an insecure form without hosting, support, maintenance or future development.”

Intuit does just that and charges north of $100 per month.

Insecure? What do you think is more secure, trusting a third-party SaaS with your data or processing it on your own server?

Which is more secure, a GCP instance or you managing your own infrastructure?

Remember that TFA is talking about personal tools that run on his own computer, so definitely his own infrastructure is more secure. Right now in Europe, GCP (like any other American cloud provider) is a big no-go as well, so I'm not sure even in professional settings the comparison holds.

This seems like an incredibly defensive take for vibe coding personal apps.

Replacing some subscription app like Any.do, Google Calendar, fitness/diet tracking or basically any other CRUD-centric app, needn't be insecure, and a semi-competent developer can easily host it, continue further development (with or without vibe coding) and secure it. There's huge benefit for software developers that do find themselves using many of these apps with active subscriptions to make their own, tailored for themselves, and cut down their spending.

Yes, when it comes to commercialising such software, more work needs to be done (mostly in support and marketing), but for personal use it's fine. The author explicitly states they don't trust vibe coding enough to turn these into products.

The writing is hardly on the wall for all these companies which make little todo list apps and calendars. The vast majority of people could get a LLM to produce an alternative but the lacking they have in basic software engineering would eventually be a hurdle to further development. Most people will continue spending $1.00/month here, and $2.99/month there. There's no reason why software engineers need to do that anymore, unless paying this gives them access to some sort of content repository (music, books) or actual advanced software.

>There's huge benefit for software developers that do find themselves using many of these apps with active subscriptions to make their own, tailored for themselves, and cut down their spending.

I guess if you're unemployed or in an area with spectacularly low wages, and don't have any ideas of your own that seem monetizable.

>Most people will continue spending $1.00/month here, and $2.99/month there.

If I make 100 dollars an hour as a consultant, and I spend 1 hour to make a local version that never needs any updating on my part to replicate a portion of that functionality I get for $1.00/month, it will take me 101 months to see any profit on my investment of time.

Cut my pay in half and I still need 51 months to see any profit. It would be idiotic for me to waste 1 hour on that.

And let's face it, code when made is a cost center, you will have to keep it up to date (so as to not introduce security hazards etc.) you will never break even much less earn anything for your time.

By that logic, how much did it cost you to write this comment?

Now be the teenager with no college degree that you were before you ever made $100/hr. Most of your options then paid < $10/hr, if that. The calculus looks a lot different then.

The vibecoded threat isn't from Devin, the threat is from 15 year old nerds who no mortgage, partner, or responsibilities. Powered by insecurity, pride, and spite, plus a generous amount of teenage hormones. I would love to see What I could have built if I had Claude back when my reflexes were still powered by youth.

Plus, if you made an app that other people were paying $1/month for. Sell yours to people for $2/month (or $0.50/month, you decide!) and you'd recoup your money much faster than 51 months.

Exactly. Most people are also not in the position to easily make $100/hr and be able to just burn money to "save time" with subscription apps.

There's another competition, your average folks who end up using physical notebooks, stock Notes/Calendar/... apps and Google Sheets when they figure the subscription isn't worth it anymore / they can't keep up with all of them due to rising CoL.

One’s time is only worth money if they would otherwise be making money. If they would otherwise be playing video games or watching Netflix, then the app they spent an hour on cost them nothing in opportunity cost.

I now code for fun with AI because it can handle the boring parts. 80% of anything is mostly drudge. And while we can’t subsist on frosting alone, having assistant that can keep you in the zone is very rewarding.

Three major iterations from now this whole conversation will be quaint and everyone will have always thought this.

Almost nobody can just arbitrarily decide to go make money in a given hour. You work 40 hours a week, you get paid for 40 hours a week, and that’s it.

And if they simply want something tailored for themselves?

Something designed without content suggestions, ads, influence and constant un-necessary redesigns, for privacy and to retain their own data.

Good for you economically. Some people are unemployed and underpaid. In fact, most are. Half of your post just came across as you broadcasting your economic success.

> This reminds me of the people who think they can build docsend in a weekend.

The author is very clear that that's not what they are trying to achieve:

"Now, don’t get me wrong, Jabber is not “production quality.” I would never sell it as a product or even recommend it to other people, but it does what I needed from Wispr Flow, and it does exactly the way I want it to."

And yet for whatever reason, jabber actually works for me. I have tried a few apps, some of them I even paid for and I thought voice recognition is not good for me (I'm a L2 speaker).

Whatever the default model was choose for jabber, it just hits my sweet spot: it recognizes my speech faster than typing (I type relatively quickly and speak relatively slowly, so any lag in the recognition can quickly tip the scale).

I honestly don't know what a real production software of this kind would do better (besides the promise that somebody will actually take care of issuing updates for the next time apple breaks backward compat)

You over-estimate quality of most SaaS products

> This reminds me of the people who think they can build [some commercial app] in a weekend. No, you cannot.

If the article was claiming that you could, you might have a point.

> You can build a wee throwaway app with some of the features of [some commercial app].

Which is precisely and explicitly what the author has done.

But throwaway is not a replacement, it is something that will break after some time unless you maintain it. And now, no, you did not make it in a weekend.

A big pricing problem in software is that people try to sell products as-if you'd use all the features, but people rarely have any interest in most of it.

There are whole products that succeed by simply being less capable, but cheaper, variations of some other product. This is most obvious in the art space, where you can visually see the reduction in tools and menu items.

This people often don't want support, maintenance, or future development. Their needs have been met.

What you say isn't different from what the blog post says: he makes it clear that he wouldn't trust the results to be production-ready and he wouldn't sell nor recommend his vibecoded stuff to others. The only claim is that those are good enough to cover his specific use case, no more and no less.

Missing my point. It isn’t good enough to suit his use case, he just hasn’t realised yet.

Bugs to fix. Dependencies to maintain. Features to add. Data to protect and restore when things go wrong.

You don’t just pay for some code. You’re paying for a service.

Being able to self-host is a plus for me. And being able to support myself avoids being told "no" for a feature I want.

Reminds me of a guy I know who wants to vibe code all his company tools, his latest target was Ashby