Hey Reader,

In Star Trek, everything is free. If the last few weeks are any indication, it seems like the software industry is (unwillingly) headed in that direction, too. 🖖

For example, I spent the last couple days coding a really cool new app for the pilot students in our AI Executive System program. So far, I have spent 8 hours "guiding" Claude Code on this – I am the architect and AI is the builder. It is now at a level of quality that would have taken me (a professional software developer with 20 years' experience) 400+ hours of hand-coding to accomplish.

400+ hours of hard work condensed into 2 days of fun.

This is happening in every corner of the software industry right now, and it is why the finance media coined the term SaaS-pocalypse. The software-as-a-service (SaaS) business model – "pay $20/mo for access to software we spent $1 million+ building" – no longer makes sense in a world where useful software just comes out of a Star Trek-style replicator for nearly zero cost.

Pay-per-month software will continue to exist for the time being, but you can see investors changing their assumptions about the business model if you look at the price of public companies like Monday (down 57% in 6 months), Atlassian (down 51% in 6 months) and Salesforce (down 17% in 6 months).

The ability for almost anyone to code almost anything means that the multi-million-dollar investments made by tech companies can now be replicated for hundreds of dollars. This eliminates the "moat" that allowed SaaS companies to get super-high valuations from investors. In the past, the SaaS companies could say:

1. "It is super expensive and time consuming for a competitor to build anything better."
2. "It is a huge pain for our clients to switch to a competitor."

AI coding falsifies both of these assumptions. Now, it takes a meaninglessly small amount of time to replicate the features of Atlassian and Monday's software. That allows new competitors to emerge and try to woo customers away from the incumbents. AI also makes data migrations way easier – so the customers are less "locked in" even if they have a ton of data stored in the incumbent systems. This leaves the SaaS incumbents with little choice but to compete on price or lose customers, and in both cases the business becomes less valuable to an investor.

And then once-vaunted SaaS companies lose 50% of their value in a few month. 😭

(There is a way for the new competitors to capitalize on this trend – I will be talking a lot more about that in the coming months. But it's very bad for incumbents who based all their decisions on assumptions that no longer work.)

If you temporarily look past the pain experienced by the companies, their employees and their shareholders, the SaaS-pocalypse is actually great for everyday consumers.

You'll have better options available for pretty much everything you do online, by virtue of AI coding allowing for competitors to emerge more quickly and cheaply. More competition and a lower barrier to entry means eventually you'll get better products at a lower price. It might even inspire more people to build more truly open-source software – which is a clear public benefit.

At the same time, the SaaS-pocalypse is probably a forerunner of similar things happening in other industries. (Law is probably next.) At a certain point, all those changes will put stress on white-collar workers and the economy as a whole. That was the thesis of a viral "worst-case scenario" for the white-collar economy last week that's worth a read – even if it is unlikely to come true in full.

The most striking thing to me:

AI is incredibly powerful, but it's always been about making things cheaper. Making new things cheaper reduces the value of things that already exist.

This is controversial for a number of reasons, but the most acute is that the writing, photos and videos used to train AI are usually copyrighted works. There is a huge unanswered legal question as to whether companies like OpenAI, Anthropic and Google should be required to compensate or get permission from copyright owners. (It seems unlikely to me that they will be forced to do that – the cat is way too far out of the bag already.)

The reality that AI is built upon copyrighted materials also makes the month's other set of headlines feel bizarre to me:

From a national security standpoint, the American government has an interest in keeping China behind-the-curve on AI. But it's also comical to hear OpenAI and Anthropic, whose entire existence is predicated on questionable use of copyrighted material, complain about their own intellectual property being used in inappropriate-but-probably-not-technically-illegal ways.

It really gets to the central paradox of AI:

AI takes things that already exist and allows you to make new versions in a way that is much faster and cheaper. In doing so, it helps create tons of useful new things, but also reduces the value of all of those things, since they're now so easy to create.

In other words, it's the Star Trek replicator. In a world where you can replicate infinite apples for nearly free, individual apples have no value. So, the United Federation of Planets doesn't need money (although they do still manage to conduct a variety of costly border wars). Nobody goes hungry and everyone exists in a happy state of Fully Automated Luxury Communism. All in all, it's a great universe.

On the one hand, you have to love the idea of a world with nearly infinite, inexpensive nice things.

On the other hand, it's a huge shock to a system built on the scarcity of the things that AI has suddenly made nearly free.

Until next time,

– Rob
CEO of Innovating with AI