The Great Software Collapse
The Great Software Liberation: Why Every SaaS Will Be Free
There's a shift coming most enterprise software companies wont appreciate. Not a gradual evolution, more a collapse.Like 3,000 video editing suites when iMovie shipped for free, or GPS hardware when Google Maps arrived on your phone.
Hollywood movie title altert: "This time is it's coming for everything, all at once"
Vibe Coding - The New App Store
Cast your weary mind back to 2008 - oh wait you may not have been born yet. The App Store launched and within a few years, entire software categories were wiped out or permanently repriced. A $400 standalone GPS became a free app overnight.
AI-assisted vibe coding is that moment, but for enterprise software. Tools like Lovable, Cursor, and Bolt let a non-engineer describe what they want and have a working application running in minutes on Postgres via Neon or Supabase with instant production access. The barrier between "I have an idea" and "I have a product" is approaching zero.
Monday.com and Zapier are the canary in the coal mine. They made complex workflow logic configurable by ordinary people. AI vibe coding makes it generatable by ordinary people. Why pay $49/seat/month for a workflow tool when you can describe your exact workflow in plain English and have something custom-built before lunch?
The Incumbents Have a Problem
Atlassian charges enterprise rates for project tracking and documentation. Solved problems. The moment AI can generate a bespoke project management tool tailored to your team in an afternoon, Atlassian's pricing power evaporates.
Salesforce and Microsoft Dynamics built vast, expensive, consultant-dependent ecosystems on the quietly brilliant business model of making things just complicated enough that you can never leave. Most companies use 15% of the features. A custom-generated CRM built around your actual sales process, running cheaply in the cloud, is not a fantasy, it's a prompt..
Xero is the a nuanced case. Born in Wellington, built for small business owners who just want their books to work, Xero understood early that software should get out of your way. The compliance moat is real. But the interface layer, the workflows, the reporting are all up for grabs.
Claude, ChatGPT, and the AI platforms themselves face the most fascinating disruption of all, because they are simultaneously the disruptors and the disrupted. AI eats itself?
The Personal AI Endpoint
The logical conclusion here isn't better SaaS - it's the end of SaaS as a category.
Moore's Law never stopped - it just changed address. What requires a data centre today requires a high-end laptop in two years and a mid-range phone in five, and with AI model efficiency roughly doubling every year on top of that, the arrival of a fully capable personal AI agent looks...logical.
At that stage your AI isn't a subscription, it's an agent living with you, knowing your work, your relationships, your money, your communication style. It thinks locally, privately, and entirely in your interest. Every application becomes a thin interface on top of this personal model the app layer becomes almost incidental skin on top of something that already gets you completely.
The Identity Problem Nobody Has Solved
All of this runs headlong into the biggest unsolved problem on the internet: identity.
We have defacto global systems - Google, Apple,Visa, Mastercard, SWIFT etc. Because the financial system had strong institutional incentives to agree on shared rails. The internet never got its equivalent We got passwords, then OAuth, then "Sign in with Google" - which is just outsourcing your identity to an advertising company and hoping for the best.
A personal AI transacting on your behalf, representing you in digital spaces, needs a universal identity layer that doesn't yet exist. Not a government ID stapled to a browser. Something cryptographically sound, globally portable, privacy-preserving, enough storage to store your life and acceptable to both authoritarian and democratic governments simultaneously.
The more likely path is a bottom-up defacto standard - the way TCP/IP won, the way QWERTY won. Not because anyone formally agreed, but because enough people just started using it. Pragmatic, no rizz, does the job mate.
The Security Elephant in the Room - DumbOS
Lovable, Supabase, and Neon have put production database access into the hands of millions of people who have never heard the words "row-level security." There are applications handling real user data
Bbuilt in an afternoon, on a publicly accessible Postgres instances, no authentication. It's a bit like handing out power tools at the school fair and being surprised when someone loses a finger. Ooops.
Scale that by an order of magnitude. Two hundred million people vibe-coding luvvies, strutting their way to production, blissfully unaware of what SQL injection is - meanwhile the script kiddies prompt for holes. Awww thats one big attack surface you've got there bro.
The personal AI future is only a good one if it's also a secure one. And security, historically, is the thing that gets bolted on last, usually after an embarrassment and someone has to front to the media.
The Arc
Software gets cheaper, then free, then personal. The platforms that survive will own infrastructure i.e compute, identity, payments, or network effects that a personal AI can't replicate alone. Everything in between faces an existential repricing, and no amount of enterprise sales motion is going to stop it.
There's something genuinely levelling about all of this. The small business in Taranaki gets access to the same tools as the Fortune 500. The rugby field doesn't tilt, it flattens, very much the Kiwi underdog triumph story we love down here.
Comments
Post a Comment