The US healthcare system is well and truly f'd, but I think 98% of these issue are government and society policy issues. If anything, I see so many companies trying to take advantage of the complete dysfunction in the US healthcare system to be yet another middlemen siphoning money from systemwide inefficiencies.
My primary point is that poor health outcomes in the US (fully admit my views are US-centric given my experience and knowledge) have nothing to do with a lack of technology and everything to do with structural systems issues rooted in government policies.
I'm a big fan of Dr. Jessica Knurick, who publishes at length about many of the systems that cause and contribute to population-wide poor health in the US. Here's one article taking about food systems and nutrition: https://open.substack.com/pub/drjessicaknurick/p/the-food-sy... .
But this means the market for SaaS products is going to get hit hugely. If you can vibecode up a specific service for your specific requirement in a few days, why bother buying a SaaS product?
And, of course, if you can build a me-too SaaS product that imitates a successful competitor over a weekend, and then price it at 10% of their price, that's going to hit business models.
I think the SaaS startup gravy train is definitely over and done.
Personally, my sense is that there's a lot left to do in batteries + motors + LLMs. The drones in Ukraine could be smarter. Robot companions that can hold a conversation. Voice interfaces for robots generally [0]. Unfortunately, the people making all the batteries, motors, and increasingly the LLMs, are in China. So those of us stuck with idiot governments protecting their fossil-fuel donors are going to miss out on it.
[0] the sketch of two scots in a voice-controlled lift still resonates, though. There's probably still work to do here.
AI makes code "free" as in "free puppy".
And part of the problem that the SaaS solves is that "I have this thing that I need to do. I can probably do it in software, but I don't know how. Can I buy that software?". Which is now becoming "Can I get an LLM to do it?" instead.
There's also no danger of it being enshittified. Or of some twat of a product manager deciding to completely change the UI because they need to change something to prove their importance. Or of the product getting cancelled because it's not making enough money. Or of it getting sold to an evil corp who then sells your data to your competition. Or any of the other stupid shit we've seen SaaS companies pull over the past 20 years.
It’s useful to take an argument and take it to its logical extreme: I just don’t see every company in the world, large and small alike, building everything they depend on in-house, as though they were a prepper stocking up for Armageddon. That seems pretty fanciful on its face.
The maintenance costs are kinda overrated: you fire up the LLM, point it at the code base and say "this needs fixing". I'd say that the maintenance costs of dealing with endless patches and fixes from a SaaS for features that you don't use is probably more onerous.
And generally we're talking about the situation where the SaaS customer is a domain expert in their area of expertise, but that isn't software development. They can use a system incredibly well, they just can't develop it. They have in-house IT folks to maintain their computers, networks, etc, and they're really just adding a couple of people to develop and maintain some applications via LLM to that team.
We're already seeing some of this, so it'll be interesting to see how far it goes.
That said, the inquiry doesn't there. What happens next after the content is generated matters. If human creativity is then applied to the output such that it transforms it into something the machine didn't generate itself, then the resulting product might be copyrightable. See Section F on page 24 of the Report.
Consider that a dictionary contains words that aren't copyrightable; but the selection of words an author select to write a novel constitutes a copyrightable work. It's just that in this case, the author is creatively constructing from much larger components than words.
Lots of questions then obviously follow, like how much and what kind of transformation needs to be applied. But I think this is probably where the law is headed.
So if it's distributed outside of the license, that's subject to contractual penalties? I guess that's what all the "wrapper" SaaS businesses will do.
Read that report, it defined the issues and the boundaries well, for the current generation of AI tools. As they develop and expand, it's going to get interesting, especially if robotics/3d printing etc get involved.
If I use an Optimus Prime to help create art, similar to Andy Warhol's "factory", do I own the copyright on the completed work?
If a person uses AI to generate work that ends up being patentable, are patents also not available?
All software licensing depends on copyright. If no-one owns the copyright than it can't be licensed; it's in the public domain immediately and irrevocably.
Of course, if you rip off someone else's work, and they DMCA you, then you might need to prove that they generated the entire thing using an LLM with zero human input. Though there's plenty of folks posting blog posts claiming that that's their process, so it might not be that hard.
Patents are different. For a start they cost money and effort to get. And there are lots of rules around how they're applied and how you can defend them. Very different from copyright.
[0] https://www.cnbc.com/2025/03/19/ai-art-cannot-be-copyrighted...
I feel like a lot of people are about to learn this lesson for the first time. Except in some very niche areas the majority of the value was never the code. The SaaSs that everyone thinks will be replaced had much more than code if they were successful -integrations, contracts, expertise, reputation, etc…
Exactly right
This is the thing that keeps me up at night. Tech has allowed a very solid middle class lifestyle for a lot of people. I can't think of another good paying job where someone is self-taught, or went to a 12-month certificate program at their local community college and now has a very good career.
If those jobs disappear, or wage growth is non-existent, I don't know where the next generation will find those jobs.
There are still (always?) business opportunites to leverage technology, in what we used to think was a virtuous loop of positive feedback.
If corporates are going to build AIs to attempt to sell things to people, there's going to be an opportunity for AIs that work for an individual, a "de-enshittifier" for example.
The "big model providers" right now aren't necessarily the actual ones that will persist. We're in another dotcom-type boom and when the tide goes out, some of them will have been swimming naked.
I think the additional wrinkle with AI is that it's having an impact, just not really in the way these execs are saying. Before ChatGPT, there was lots of speculative investment into SaaS-type products as companies looked for another hit. Now, though, I think there is a general sense that, except for AI, Internet tech (and lots of other tech) is fully mature. This huge amount of investment in "the next big tech" thing (again, ex-AI) is just over, and the transition happened pretty fast. Blockchain, NFTs, the metaverse, Alexa and other voice assistants, yada yada, were all ventures looking for something as big as, say, the rise of mobile, and they all failed and are getting killed basically simultaneously.
I think the scary thing going forward is that, over the past 25-30 years or so, tech provided a huge amount of the average wage growth, at least in the US. Even if AI doesn't result in huge employment reductions due to productivity gains, the number of high quality jobs in the AI space is just a lot smaller than, say, the overall Internet space. Lots of people have commented here how so many of these AI startups are just wrappers around the big models, and even previous hits are looking dicey now than the big model providers are pulling more stuff in house (and I say this as a previous Cursor subscriber who switched to Claude Code).
I'm curious what future batches of YCombinator will look like. Perhaps it's just a failure of my imagination, but it's really hard for me to think of a speculative tech startup that I think could be a big hit, and that's a huge change for me from, say, the 2005-2020 timeframe. Yeah, I can think of some AI ideas, but it's hard for me to think of things beyond "wrapper" projects on one hand and hugely capital intensive projects for training models on the other.