When I lecture about AI, or work with clients in workshops, there is one concept I keep coming back to. It is what I call the “tabula rasa”. The idea is simple: in many respects, we are starting from scratch again. Long-standing hierarchies of expertise, access, and even prestige have partially collapsed. The entry point has shifted.
I have written about this before, but it keeps becoming more evident. Today, a 20-year-old with a bunch of subscriptions and a decent internet connection can already produce surprisingly competent work. Not brilliant, not deep, but undeniably useful. Things that once required years of training, institutional access, or expensive tooling are now reachable with curiosity, time, and patience. It has never been easier, or cheaper, to become operationally skilled in something.
This is the part of the story that gets repeated the most, and for good reason. AI lowers the floor, democratizes capability, and shifts responsibility back to the individual. Effort, consistency, and the willingness to spend long hours experimenting suddenly matter again. Excuses evaporate quickly. The tools are there. The knowledge is there. The barrier is no longer access, but discipline.
For those like me, who have spent hundreds of thousands of dollars on education, this shift can be irritating. Much of what once required years of formal training can now be approximated by people who simply started earlier, experimented more, and cared less about credentials. And yet, I cannot bring myself to resent this moment. If anything, it triggers a slightly childish enthusiasm. Not because expertise has lost its value, but because it can no longer hide behind scarcity. Knowledge is no longer something you own by having paid for it. It is something you must keep earning, every day, in public, through use. It may be unsettling, but it is also very much alive.
This, however, is only half of the picture.
On the other end of the spectrum, I see the opposite problem emerging. Consulting companies spending tens of thousands of dollars on tokens. Law firms committing to one million dollar per year or more in licenses for a single AI platform (if you know, you know…). Entire departments built around managing prompts, workflows, integrations, and compliance. AI does not simply reduce costs. It creates new ones, often invisible at first.
As soon as you move from experimentation to production, specificity becomes expensive. Fine-tuned systems, secure environments, domain-adapted tools, governance layers, compliance costs, certifications… Each step adds friction and cost. For some players, this is becoming unsustainable, not because they lack talent, but because the price of staying competitive keeps rising.
This is where the paradox becomes clear.
Being successful has never been cheaper and more expensive at the same time. Cheap at the margins, expensive at scale. Accessible for individuals, exclusionary for organizations. Empowering for newcomers, but increasingly heavy for those who need reliability, governance, and scale. AI flattens the entry curve while steepening the long-term one.
I do not have a clean answer to this tension. I am not even sure there is a single answer. What I am sure about is that this paradox will shape the next phase of AI adoption more than raw model improvements. The real divide will not be between those who use AI and those who do not, but between those who can afford to operationalize it sustainably and those who cannot.
And that, to me, is where the most interesting conversations should start.
