One of the things about this whole death of SAAS thing is -
People buy SAAS to offload the cognitive load of understanding the problem space, not just because it’s hard to write for loops.
But writing code probably is a cost. I think there will be an impact but I wonder if - these big companies keep selling to big companies because no one wants to hire a whole ass compliance team to ensure the business logic is consistently up to date.
But maybe there’s a whole lot of people caught up in edge cases that can be solved by a smaller team now? This I think is mostly what I hear about.
In summary, I suspect this is an overcorrection but there is some level of core concern here.
I clicked and starting reading this expecting some actual connection between said markdown files and losing money. There was nothing at all like that in this blog post.
Responding to the headline and not the article substance: no, the SaaS crash is a crash because valuations are speculative and have been very high. Security price motion is only very loosely coupled to fundamentals in the short term, and this moment in history is both highly volatile and unanimously held to be overpriced. Ergo, crash.
That doesn't speak to the fundamentals, with which I only sort of agree. There are SaaS products that just grease inter-human interactions that would be hard to manage otherwise, and these are dead in the water. There are others that manage data human beings will always need to be able to understand, even if the AI is doing the work[1], which are much safer.
[1] Like bug trackers. We all love to hate Jira, but even if you have an army of Claudes doing your coding and testing for you someone somewhere needs to know what still needs to be fixed before shipment.
I mean, Markdown is just a way of storing information.
If the title said "WS just lost $285B because of 13 documents", people would have said "Why mention the documents? We get that people who move money around use documents".
How long until we have agents talking to other agents in a web of agents, and together deciding on something catastrophic? Not individually, but somehow the result of a group process.
SaaS is generally defined as cloud based software. What has it do with easier replacement by AI?
In fact it should be opposite. Local utility software has higher chance of being replaced by AI as then you wouldn't worry about server and state and all the complexities of storage.
Also the article mention stock drop of "technology companies", which is basically stock drop of top 5 companies, like Nvidia and Apple, which is not even related to SaaS.
> SaaS has a serious issue with agentic tooling being able to replicate software.
[...]
> For example, if you hold a company's accounting transaction data and related records, and expose it over MCP (or an old school API that can be wrapped into a CLI - which works better in my view), agents can use this with remarkable efficiency.
Sure why not wire all that transaction data directly into openclaw, what could go wrong?
Snark aside, it seems like there are two ways to go with a method like that if you want to keep the transaction data safe. The first is locally host your LLM(s) so the data stays on-prem. The second is to trust a third party to jump through all the legal and technical hoops to properly handle that very important, very private data. And if you're doing that, you're still doing SaaS. It's just a different provider.
I think this is a case where commenters are reading more into the price movement than they should. I don’t think it’s easy in this case to say how much of it is directly the model releases as opposed to software having a great run and this gives investors a reason to lighten their positions.
I also think the SaaS has is doomed narrative does not work. There is a whole host of compliance, edge cases, reliability that I don’t you can simply bring in house because of AI assisted tools.
With that said I do think there are increasingly a whole host of more service based businesses where they are under threat. Thinking consultancy, legal, marketing or other similar roles. If you can use a leading model with these massive context windows to get 80% of the value for practically no time and no real cost compared to using a human where the quality will be higher might it might be a multi day engagement and easily 10-20k, you might start deferring some of the initial work to AI and only in certain cases then send it out to the human.
If you think contract law can be accurately represented in a 170 LOC Markdown file, you are incredibly stupid. If you think the current frontier models do not hallucinate non-existent court cases with case IDs like 1234567, check YouTube for all the idiot lawyers and pro se defendants showing their whole ass. They are using ChatGPT with one of the current models.
The thing that keeps on getting glided past in most discussions of this is the distinction between systems of record and systems of workflow.
If you're SAP, Workday, Procore, maybe even HubSpot, you have a shout. Growth won't be fast, but you're okay for now and might even be able to position yourself as an integrator.
If you've raised large rounds and you're just a system of workflow that won't trigger a years long political fight to get you out - a document review startup, good luck.
It is not my intention to offend anyone in particular, but that sell-off has really colored my perception that investors are such _sheep_ when it comes to AI. Serious FOMOs. Not a shred of independent, nuanced, or critical thinking; just black or white opinions.
Does this article say anything of substance? So you found a repo with 13 markdown files; what about it? I see absolutely nothing about it. Does it have to do with "SaaS legal review"? I see no relation whatsoever.
Yes, we're in for more headless interfaces and there are existing products that will struggle to serve these new interaction models due to organizational constraints. But I don't think it's as simple as asking "are they a system of record" as we think about the companies that will adapt and thrive and the new ones that will come. Enterprises are investing AI spend into improving core processes and responding to competitive pressure, not saving money and introducing risk into areas they have historically delegated to vendors. AI is going to give us more software, and increase spending as firms seek efficiency in new areas, and they're going to continue to knock on doors of vendors to do it as they always have. Not to mention the demand for auditable, repeatable workflows is still there and always going to be there and dedicated systems are needed to solve this in each problem domain.
21 comments
[ 4.3 ms ] story [ 31.4 ms ] threadPeople buy SAAS to offload the cognitive load of understanding the problem space, not just because it’s hard to write for loops.
But writing code probably is a cost. I think there will be an impact but I wonder if - these big companies keep selling to big companies because no one wants to hire a whole ass compliance team to ensure the business logic is consistently up to date.
But maybe there’s a whole lot of people caught up in edge cases that can be solved by a smaller team now? This I think is mostly what I hear about.
In summary, I suspect this is an overcorrection but there is some level of core concern here.
But like, how are IFTTT and zapier doing?
It is annoying to see a complex field as law being judge by stock prices of some SaaS products or w/e they are basing this on.
Could have at least gave it some effort and interviewed 2-3 people that use these kind of products but that would be too much work surely.
It is not surprising that people that slop like LLMs are so into LLMs
"Go away or I will replace you with a very small shell script" used to be BOFH lore, now it has become real. This is great.
That doesn't speak to the fundamentals, with which I only sort of agree. There are SaaS products that just grease inter-human interactions that would be hard to manage otherwise, and these are dead in the water. There are others that manage data human beings will always need to be able to understand, even if the AI is doing the work[1], which are much safer.
[1] Like bug trackers. We all love to hate Jira, but even if you have an army of Claudes doing your coding and testing for you someone somewhere needs to know what still needs to be fixed before shipment.
If the title said "WS just lost $285B because of 13 documents", people would have said "Why mention the documents? We get that people who move money around use documents".
In fact it should be opposite. Local utility software has higher chance of being replaced by AI as then you wouldn't worry about server and state and all the complexities of storage.
Also the article mention stock drop of "technology companies", which is basically stock drop of top 5 companies, like Nvidia and Apple, which is not even related to SaaS.
[...]
> For example, if you hold a company's accounting transaction data and related records, and expose it over MCP (or an old school API that can be wrapped into a CLI - which works better in my view), agents can use this with remarkable efficiency.
Sure why not wire all that transaction data directly into openclaw, what could go wrong?
Snark aside, it seems like there are two ways to go with a method like that if you want to keep the transaction data safe. The first is locally host your LLM(s) so the data stays on-prem. The second is to trust a third party to jump through all the legal and technical hoops to properly handle that very important, very private data. And if you're doing that, you're still doing SaaS. It's just a different provider.
I also think the SaaS has is doomed narrative does not work. There is a whole host of compliance, edge cases, reliability that I don’t you can simply bring in house because of AI assisted tools.
With that said I do think there are increasingly a whole host of more service based businesses where they are under threat. Thinking consultancy, legal, marketing or other similar roles. If you can use a leading model with these massive context windows to get 80% of the value for practically no time and no real cost compared to using a human where the quality will be higher might it might be a multi day engagement and easily 10-20k, you might start deferring some of the initial work to AI and only in certain cases then send it out to the human.
microsoft ceo said the same thing - agents will be using windows/office and be the bigger buyers of licenses
replacing SaaS software at this scale is hard, nobody will risk it, at least near term
If you're SAP, Workday, Procore, maybe even HubSpot, you have a shout. Growth won't be fast, but you're okay for now and might even be able to position yourself as an integrator.
If you've raised large rounds and you're just a system of workflow that won't trigger a years long political fight to get you out - a document review startup, good luck.