The Great SaaS Gaslight

Just as consumers don't want to pay the "true" price of food or the "true" price of clothing, we don't want to pay the "true" price of software.

You might grumble about $30/mo for something like Postman, and it's true that "back in the day" it might have been a $40 one-off, but that's closer to $90 with inflation now, and there's a good chance that would have only bought you version 4.

Then next year version 5 comes out, and you face a dilemma. Do you pay all over again? Do you end up staying on version 4 until eventually there's a compelling reason to upgrade?

SaaS solves that problem by keeping everyone on the latest version.

Software is economically expensive to produce, we don't do enough to recognise that, in part because of how much free work is contributed by open source contributors and how little we recognise how much work they really do.

I’ll be glad when this brief interlude of “OMG AWS go down! We stupid? Cloud bad?” posts are over and we go back to normalcy.

Yes, we never needed the cloud. We could’ve kept the servers onsite or in data centers like we used to and pay much less to do much more. I worked in a startup that had hundreds of servers, and that was enough, and that served phones from cellular customers and mobile devices from some major cellular companies and media companies. Not many IT staff were needed for that and as devs we never had to think of how it was hosted, even though I did from my former experience.

Now, when I develop, I have to have at least a basic understanding of AWS since our production code is built and hosted there.

We have so many risks now. We can’t even use a package manager without worrying about secrets being stolen. We give away our secrets and code to an LLM that knows our intellectual property before we do, because it’s creating it.

Development has been an increasing dystopia from the beginning.

I don’t know if you’ve seen Silo, but if you have look at the monochrome graphical fileserver-ish UX that they use in the first season. I love that. I wish our shit were like that.

But no. Instead we keep developing features no one asked for requiring new expensive equipment. Our aging parents can’t even operate a TV because instead of an on/off switch with volume and channel buttons, it’s either a complex UI or a cable box/TV combo with input selection- what is HDMI 1? They don’t know.

A good enough solution on par with the competition is exactly the right choice for a business 95+% of the time though. Your company probably shouldn't compete on email/calendar/vpn/filesharing/OS/spreadsheets/CRM/reporting... all of these are great as SaaS.

And Google Workspace is a vastly superior solution to its predecessor offline or semi-online office suites, allows live collaboration and works across devices. Hubspot has a real, usable mobile app. Even Power BI is a step up from whatever pile of excels used to pass as reporting.

Don’t forget the tiers where security is held for ransom!

The elephant in the room: piracy.

I would not say SaaS is a problem in itself(I built and run one myself). Software as a Service simply means that you can run an application on a remote machine, which you do not have to manage yourself and can pay only for what you need. That makes perfect financial sense.

The problem with SaaS today is that the subscription-based model is getting out of hand - the value proposition no longer makes sense, in many cases.

And secondly, vendor lock-in. You cannot get your data out. Or if you can, you will be less likely able to migrate to another provider or local application which means there is no free market where various providers can compete with better features, customer support, availability or price. Therefore, in the end the provider will hold you hostage via your own data. This is, after all, Amazon AWS's famous moat. It is often very expensive and painful to migrate into some in-house solution. And often it is simply not possible at all as it might require such a massive rewrites of your own application(understand as dependency on the provider), that it is not in the realm of possibility.

So with like everything else, you are responsible for your own choices and if you make a bad judgement, like tying yourself to one provider, that is all on you.

People keep yapping about monopolies in tech, yet they vehemently dislike decentralisation or taking care of their own stuff and want to have all the things on one place. But when things do not work out as they wanted, then these monopolies become a problem that is too late to mitigate.

tl;dr do not use SaaS of mission-critical functionality and you'll be fine.

My beef with SaaS is that every client I went to just wanted developers who would glue together these solutions and they would invariably have puzzling overlaps in the products they purchased, particularly around CI/CD pipelines. I would be like "why does this go to this other service then deploy? You know it can just deploy from there, right?" And they would invariably say that "so and so made us buy that other service for their department so we're all standardizing around it, end of story." So you can never consolidate anything and it all just ends up as madness glued together with some Python. And writing that "glue" is shitty work and I can't believe how much they paid for me to do it.

I very much resent the subscription model--it is terrible for the consumer and it's crept into everything and it's just disgusting.

For products like Photoshop and AutoCAD, it would have been great to have the option to use it for a few months and stop. This way, I can subscribe when work is available and I can offset cost by earnings. Of course, I am speaking as a freelancer.

However, what I understand is that this level of on and off of subscription is very difficult.

Too one-sided to be a useful article. Negatives need to be balanced against positives.

B2B SaaS can have some really great advantages for clients. It is best if the market is somewhat competitive (say payroll) to help prevent vendors from upping prices too much.

The vendor wants to keep clients, avoiding churn. So the vendor is incentivised to keep the product current, and often add features for free.

In a non-SaaS model, the product remains as delivered, or requires version upgrades or maintenance contracts (usually crappy).

SaaS can have cheap upfront costs especially if designed for easy onboarding, and often you can quit easily (low sunken costs).

Old-school software delivery is regularly a nightmare, and over budget because the vendor is incentivised to take money from their clients.

Tht buyer needs to choose their compromises.

I worked for a payroll software company with a product that ran on Windows, and they regularly screwed over clients because there was no alignment of incentives.

The fundamental SaaS lock-in comes from bundling two things: 1. A declarative, stable interface 2. An expert support/ops team

I think the path forward is to unbundle them.

We're already solving #1. Nix has the best potential to become that declarative & stable layer, letting us reach the goal of treating cloud providers as the simple commodities they should be (I wrote about this approach here: https://ryanrasti.com/blog/why-nix-will-win/)

The bigger, unsolved question is #2: how to build a viable business model around self-hosted, unbundled support?

That's the critical next step. My hunch is the solution is also technical, but it hasn't been built yet.

Imagine if compilers were only available via the SaaS model. Developers and the tech community in general would never accept this, and compilers were open sourced well before the internet had developed to the point where it would even be possible. Nobody would trust a system that sent their proprietary code off to a data center to be compiled to binaries for their platform on a monthly subscription model - the idea is ludicrous.

Currently the only SaaS product that still makes sense are LLMs, but this is temporary - anyone with sense realizes that the ideal situation is to run an open-source LLM model locally and privately, but this still requires a significant investment in high-end hardware and IT technical people.

That's the basic calculation: is it overall more efficient and less expensive to hire a skilled IT team to manage your in-house solutions, mostly open-source, including security patches, than it is to rely on external providers who charge high monthly fees and use all manner of sneaky tactics to keep you locked into their products? The latter is going to win in the long run.

> Software keeps iterating on solving the same soluble problems because the remaining challenges are really difficult to solve with technology. Communication and coordination are full of nuance and subtleties that defy digitization.

Software will never, ever be more than mechanism.

Peak software is when it fades into the background, and focus stays on the task at hand.

There's only one reason for this: You aren't the customer, so your incentives aren't aligned with the vendor's. Your data is being sold to other companies, so you getting value out of the software is only necessary insofar as it keeps you there. That's not where the vendor makes their money.

Classic enshittification. Unfortunately, I have no idea how to evade it.

This entire blog really resonates. I don't take it all literally, but it is funny and insightful and dark and poetic and accurate in many ways.

More weird stuff like this please.

For smaller projects I actually started to like to vibe code in PHP; API, user login (magic links), webpages etc. Stable, lower maintenance, faster iterations etc. Now its really fun to build.

In my humble opinion, great software, and the ideal to work on is unachievable, because it is great software built in house to solve the house problems, not anybody else's problems, sure you can start with a solid base, but in the end every organisation and every department inside this organisation has their own requirements.

Of course we can't pretend that each business, or government institution is going to be able to have a handful of full stacks around just to keep up with the business requirements, but in reality, if that was the case, thing would operate much more smoothly.

Software for the masses is another category completely separate from anything that is going to help you achieve more at work, social media is an example and I think AI to a large extent is the same, since a generic gpt, one without knowledge of the past and present of your organisation, deeply ingrained in your culture will also fail to provide that particular set of knowledge that you can call invaluable for your use case.