The artist formerly known as FAANG is now rebranded at the top of Big Tech, not due to size, but in accordance to their positioning in the AI arms race:
Google
Apple
Meta
Microsoft
Amazon
What do all of these giants have in common? A seat at the table of AI power consolidation of course. Microsoft owns OpenAI, Meta has their own LLM, Apple is apparently developing one as well. Google has its own LLM and also owns DeepMind, but more presciently they are slowly moving into the compute business with Google Cloud, in direct competition with Amazon’s AWS.
As I’ve touched on in the past, all applications must tap into an LLM to harness the power of AI. All LLMs must power their compute via cloud infrastructure. And all of this is downstream of energy consumption, the elephant in the room that never gets discussed. AI becomes a commodity the same way electricity does. Furthermore, like every other commodity business in America, it is not a monopoly, but a sanctioned oligopoly that ammasses power. To no ones surprise, the power of AI is metered by the same folks who bought and sold our data the first time around. In a way, you could say that the value AI brings, and GAMMA will now control, is derived from all of that data in the first place. Comically, in a world where you are either “bundling or unbundling”, AI is nothing more than the biggest bundle of them all.
What will GAMMA do with this power? Well, they never actually lost power in the first place so one could assume they will continue to do what they have been doing all along: hire a bunch of people and make incremental software updates. But that’s not fun, is it? All those cool tools that are rapidly emerging in the AI community might no longer be developed out in the open, but behind closed doors. All of that 10x and 100x developer talk starts to dry up as software engineers are stuffed with so much money they lose their open source ethos. Does this happen without ZIRPs? Probably, as VC money is already pumping into new AI companies, and the quickest way for those startups to exit is to be acquired by GAMMA. All paths lead to Rome.
SaaS is Dead, Long Live SaaS
I talked in previous weeks how Software as a Service could be threatened, as ambitious people could utilize AI tools to build bespoke software for their own companies for fractions of what a SaaS platform would cost. Despite the looming consolidation of AI tooling power in GAMMA, there will still be plenty of open source AI tooling to enable this reality. It’s hard to predict the future of software, especially as the future of business is warping into a new reality with each passing day. It’s possible that SaaS as we know it is completely absorbed into and commodified by GAMMA.
A new interesting possibility emerges, whether traditional SaaS gets commodified or decentralized. Service as a Service, what an obvious traditionalist concept! In true Middleman fashion, each company is going to need an AI representative the same way they need IT today. This person (could be an agency, although we know the problems they espouse), is a 10x employee in the vein of (and in many many future companies, one and the same of) an executive. Executives get 10x the salary of the average employee for the decisions and visions and execution they impart on an organization. These average employees, leveraging the AI, now have the opportunity to be 10x more valuable than they previously were and order ranked along with the C-suite. These employees, mind you, are very special and have not been snatched up by GAMMA (yet), so they will command the appropriate status for being an individual who can singlehandedly alleviate the everpresent reliance upon enterprise software.
Things are morphing rapidly. Who knows how these scenarios will play out, as timing is the biggest factor. As power consolidates, the AI tools for public consumption may slow down. Or, we might have monumental breakthroughs in the next few weeks that make all these projections sound quaint. All we know for sure is that every industry is shifting, and everyone is scrambling to get their chessboard positioned in a way so as to be prepared for any outcome.
Any thoughts on how open sourced models may effect the GAMMA oligopoly of AI?