Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
Ask HN: What is the future of SaaS when things are this easy to build?
10 points by fbrncci 3 days ago | hide | past | favorite | 9 comments
In the past two 3-4 months, I noticed a pattern with myself, where I see an interesting SaaS being launched here on YC (or other places online), I try it out and I like it. Often these are more cutting edge AI tools; which still are really just LLM wrappers with agentic capabilities. I don't want to downplay them, as they are often products that make my professional life smoother and easier.

But the longer I use them, the more issues I notice with them through becoming a power-user and start to understand exactly how they work. Then usually before my first months subscription runs out; if I find them useful, I do not renew the subscription, but I spend a weekend with the latest SOTA LLM in Cursor or VCcode to build out the core capabilities for myself, and then never go back to the service. Often, even as a power-user, if some SaaS has 10-20 features, I really only need 5 of them. And then I can add 2-3 more that they wouldn't ever build. The best part is that I do not need to be "production grade", because I am the only user. I don't even need cloud services, except third party APIs, because I just spin up the repo on my localhost, and launch the apps capabilities when I need them. If there is a bug, I fix it right then and there. Security? Who cares. They'd have to access my computer first.

So quite naturally I am wondering, how many other people are doing this, and how this reflects on the whole SaaS landscape. And at the same time, morals and ethics, because I am basically out here stealing ideas from people who build products, and turning them into private apps for myself with no goal of ever monetizing them. Often I am just going back and forth between those products, and copying their features into my own app to avoid needing to pay for them. And it feels like its becoming easier and easier to do this.





> But the longer I use them, the more issues I notice with them

And that is why AI-generated SaaS apps are hype. Because a reliable SaaS is not easier to build. It is easier to launch a prototype. But building a robust, reliable app that solves a real problem in a way that is worth paying for is a completely different level of effort.

reply


Yes absolutely. A lot of these apps are surface level great, but the you dig deeper and its really just as easy to build the same functionality yourself. Keeping in mind that these are all decently well funded projects.

Often the "issues" aren't even bugs, its more the realization that I want some sort of functionality that they do not have; and immediately realize that if I spend a weekend on my own "base system" for that use case of SaaS; I can just attach anything to it, rather than waiting for them to release something new in 1-2 months.

reply


And yet, there are so many people who have never written a line of code in their life building iOS apps, web apps etc with AI tools and making more money than a super smart software engineer.

Sure these apps could have serious security issues, might have inefficient implementation etc, but in the end they are still able to provide for their families. Not saying this is good or bad, just pointing out that there are at least a few people making a living with these tools

reply


The company I work for has paid subscriptions to Lattice, Bonusly, SalesForce, Gong, Zoom, Jira, Slack, GSuite, Rippling (a YC company), OneRange and Notion.

No we aren’t going to sign a 1000 seat license to pay for a dumb AI, VC bait three person company we have never heard of.

On the other hand, even if we could replace a subscription by a vibe coded internal app, if it isn’t part of our core competitive advantage - ie “it won’t make the beer taste better” - we aren’t going to spend the effort to write it or more importantly maintain it.

While each of us have a yearly “tools and training allowance” that we can use for almost anything job related, we can’t use any random SaaS that could touch proprietary data. If it was offline sure.

reply


When it’s not about the content returned from an online service but instead the actual automation performed by the application, doing something instead of saying something, it becomes exponentially more challenging to replicate via LLM over a weekend. That’s the distinction I think you are looking for.

For example a self hosted application that provides real time service monitoring with performance metrics and can proxy that information between other data systems versus some full stack app that provides service information from the cloud.

On the other side of the coin I do write my own apps to avoid reliance on media streaming subscriptions.

reply


The money in SaaS has never been folks willing/able to build it themselves

risk-tolerant early adopters care purely about your features, and are willing to figure out the rest

the bigger orgs (folks that refuse to move off legacy software) won't move until you can answer questionnaires about industry-specific integrations, security/compliance, enterprise-level support/SLAs, training/onboarding, change management, scalability/performance, data governance and management tools

reply


> If there is a bug, I fix it right then and there. Security? Who cares.

> Security? Who cares.

And there it is.

Would you use a vibe coded accountancy software that stores your credit card, business data, identity information in plain text on a insecure server somewhere?

I know I wouldn't at all even if I made it, there are companies that go through audits and have certifications which cannot be vibe coded.

reply


Of course, I guess then this post is more about the category of apps where it doesn't matter that much. But there are still a ton of apps where all they are doing is bringing together a bunch of API keys and profit the difference.

reply


It’s like the decentralized meme.

Everyone thinks they want their own shit.

Until it becomes awful because it actually kind of sucks.

They centralize to a few solutions.

These solutions suck. We want decentralized.

Repeat.

reply




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: