> Here’s how you know you’re in denial: You use terms like “AI washing”, you think ChatGPT is just a toy and everything built on top of it is “just a wrapper”.
The author provides lots of words explaining how everyone has the wrong answers but doesn't do much to provide any right answers.
> You can’t bolt AI onto SaaS. You have to rebuild around outcomes.
To borrow a quote from Blazing Saddles:
Not only was it authentic <consulting> gibberish, it expressed a courage little seen in this day and age.
arnon · 1d ago
I can't tell everyone how to build their business - but I do actually detail that you need to walk away from the solution you built and revisit the problem you want to solve in an "AI Native" way.
That means think of the future without the usual windows-icons / SaaS-seat solution but rather how you built this with agents, a chat interface, etc.
ogou · 1d ago
This about pitching, not about building. What you can put in a vibe coded demo and a slide deck versus what functions in production a year later are very different. "Three Stanford kids in a trenchcoat" will always be able to get some kind of funding. That's less about technology and business principles and more about the legacy of P. T. Barnum.
jt2190 · 1d ago
Forget about AI and vibe coding and consider that those three Stanford grads might genuinely have found a way to deliver a significantly better product that will put your product out of business. The challenge then is this: How much of your yearly spend should you dedicate to researching these potential improvements yourself? Should you try to be an “early adopter” or a “late mover”?
paulryanrogers · 1d ago
> AI and agentic AI doesn’t improve SaaS. It replaces the need for it. Not the functionality—the entire concept of software as an intermediary between intent and outcome.
This paints with a pretty broad brush. Are agentic AI solutions precise yet? Are they secure? How costly are the mistakes they intrinsically make?
Is questioning the premise of the article just going to be met with dismissive bucketing into recycled stages of grief?
ahartmetz · 1d ago
What the fuck is that?
Summary: Don't sell software to help humans do X, sell software to do X directly for much more money... because that is currently known to work really well, right?!
arnon · 1d ago
Yes, that's the direction we're going in globally. You can keep ignoring it but it's not going away.
ahartmetz · 1d ago
Probably eventually in some way, but for sure not now.
If AGI comes around, why would customers go to an AGI "reseller" instead of paying for the AGI directly?
glutamate · 1d ago
So if I sell project management software for the construction industry, I should pivot to... selling houses?
arnon · 1d ago
No, you should pivot to solve the project management issues without thinking about it in terms of a SaaS CRM that someone uses manually
The author provides lots of words explaining how everyone has the wrong answers but doesn't do much to provide any right answers.
> You can’t bolt AI onto SaaS. You have to rebuild around outcomes.
To borrow a quote from Blazing Saddles:
Not only was it authentic <consulting> gibberish, it expressed a courage little seen in this day and age.
That means think of the future without the usual windows-icons / SaaS-seat solution but rather how you built this with agents, a chat interface, etc.
This paints with a pretty broad brush. Are agentic AI solutions precise yet? Are they secure? How costly are the mistakes they intrinsically make?
Is questioning the premise of the article just going to be met with dismissive bucketing into recycled stages of grief?
Summary: Don't sell software to help humans do X, sell software to do X directly for much more money... because that is currently known to work really well, right?!
If AGI comes around, why would customers go to an AGI "reseller" instead of paying for the AGI directly?