I'm working on vertical Saas for Infrastructure finance. The vision is to execute on the services-as-delivery Saas technical architecture that can work for this space to close and arbitrage the productivity gaps. We're bootstrapped and have already closed a $300k/yr licensing deal, but my biggest challenge is how to scale the engine that owns the complex, gnarly, high-accountability coordination between AI and the real-world gap beyond myself as the glue. We're too bootstrapped to go build that engine outright and generally I'd like to wait to raise till we have a million of revenue. So for now we're building the trust one new engagement at a time with me in the ownership seat. It's slow though as trust requirements are very high, which is part of the opportunity, but also very slow. Given the size of the market & opportunity, and how much I love your articles, I am interested in what you think a funded approach would look like. Would also love to dig in on where our thesis on this aligns or diverges from yours.
Always worth reading because the insights are grounded in first principles "the specific threat is sector agnostic ... (but) not all software is created equal".
I wonder if you'd consider adding a new immutable primitive for software defensibility at scale: not "trust" per se, but the ability to produce evidence that the software behaved as advertised. Valid inputs -> valid workflow -> measurable performance -> validated outputs.
Since I consider AI to be the new Wizard of Oz, I'd like very much to see the man behind the curtain. Without a foundation to think about data-in-data-out, I think even Bustamante's (terrific) three-point framework needs to be qualified to: as long as it can show it's not GIGO.
Thinking that AI will kill Saas is a short bridge that brings nowhere. Of course, AI will force some traditional horizontal SaaS companies to rethink their models and some will fail. AI will also enhance vertical SaaS companies. But no company will entirely replace a legacy software for AI. If you ask AI to make an image with the same prompt, it will make two different ones and I think it resumes everything.
I'm working on vertical Saas for Infrastructure finance. The vision is to execute on the services-as-delivery Saas technical architecture that can work for this space to close and arbitrage the productivity gaps. We're bootstrapped and have already closed a $300k/yr licensing deal, but my biggest challenge is how to scale the engine that owns the complex, gnarly, high-accountability coordination between AI and the real-world gap beyond myself as the glue. We're too bootstrapped to go build that engine outright and generally I'd like to wait to raise till we have a million of revenue. So for now we're building the trust one new engagement at a time with me in the ownership seat. It's slow though as trust requirements are very high, which is part of the opportunity, but also very slow. Given the size of the market & opportunity, and how much I love your articles, I am interested in what you think a funded approach would look like. Would also love to dig in on where our thesis on this aligns or diverges from yours.
https://www.linkedin.com/in/josh-katz-9249b91b/
Great article!
Always worth reading because the insights are grounded in first principles "the specific threat is sector agnostic ... (but) not all software is created equal".
I wonder if you'd consider adding a new immutable primitive for software defensibility at scale: not "trust" per se, but the ability to produce evidence that the software behaved as advertised. Valid inputs -> valid workflow -> measurable performance -> validated outputs.
Since I consider AI to be the new Wizard of Oz, I'd like very much to see the man behind the curtain. Without a foundation to think about data-in-data-out, I think even Bustamante's (terrific) three-point framework needs to be qualified to: as long as it can show it's not GIGO.
Thoughts?
Thinking that AI will kill Saas is a short bridge that brings nowhere. Of course, AI will force some traditional horizontal SaaS companies to rethink their models and some will fail. AI will also enhance vertical SaaS companies. But no company will entirely replace a legacy software for AI. If you ask AI to make an image with the same prompt, it will make two different ones and I think it resumes everything.