The Best of Times, the Worst of Times: We Are Handing Over Life to the Machine

We are, whether we've admitted it to ourselves yet, systematically handing over life to the AI machine. Not in some distant sci-fi sense. Right now, in 2026, in the granular decisions of hiring, firing, capital allocation, and how a small retailer in Ohio figures out what to do with a warehouse full of dead inventory. The handover is already well underway, and it is going to be, in equal and simultaneous measure, the best of times and the worst of times.

I don't say that as a hedge. I say it because both things are true at once, in the same data, in the same month, sometimes in the same company.

The Best of Times: An Entrepreneurial Explosion

Start with what should be the headline story, because it almost never is: Americans are starting businesses (as I did) at a rate the country has never seen before. Through the first half of 2026, roughly 3.5 million new businesses were formed nationwide — up 14% from the same period a year earlier, and the strongest January-through-June total on record. June 2026 alone was the strongest single month ever recorded, edging out even the post-pandemic surge of 2021.

This isn't an accident of timing. It's the direct downstream effect of AI collapsing the cost of starting something. The solo founder who once needed a developer, a bookkeeper, a marketer, and a lawyer on retainer can now get a rough version of all four from a laptop and a subscription. The barrier to entry that used to protect incumbents is dissolving, and people are pouring through the gap. Confidence is up. Determination is up. Millions of people are looking at economic uncertainty and, instead of flinching, deciding this is the moment to build something of their own.

That is the machine at its best: a leverage tool that turns one person with an idea into something close to a small company.

The Worst of Times: The Jobs Being Quietly Erased

Now the other side of the ledger, and it's not pretty.

Goldman Sachs' current estimate is that AI is displacing roughly 16,000 U.S. jobs a month, against something like 9,000 a month in augmentation gains. A persistent net loss, month after month, that doesn't show up as a single dramatic headline but compounds quietly. Layoff trackers now show over 300 layoff events in 2026 impacting more than 200,000 workers, and more than half of those events explicitly cite AI or automation as a contributing factor.

The individual company numbers are the part that should actually unsettle you. Oracle cut roughly 30,000 jobs. Meta laid off about 8,000 people (roughly 10% of its workforce, though they and others overhired during COVID) while simultaneously shifting thousands of others into AI-focused roles. Block eliminated close to half its total headcount. Salesforce's CEO said outright that AI agents let the company shrink its support staff from 9,000 down to 5,000. Amazon cut over 16,000 corporate roles in a single quarter while AWS grew at its fastest pace in thirteen quarters.

That juxtaposition is the real story, and it's not the one either the boosters or the doomers tell. The capital isn't disappearing. It's being reallocated, away from headcount, toward infrastructure. Google, Amazon, Microsoft, and Meta together are planning something like $725 billion in AI capital expenditure this year, up 77% from last year. The money is very much still flowing. It's just flowing around people instead of through them.

Entry-level and administrative roles are absorbing the worst of it: customer service, data entry, junior developer work, the traditional first rungs of a career ladder. Those rungs are being sawed off in real time, and I don't think we've begun to reckon with what that means for the next generation trying to get a foothold.

The Other Side of That Same Coin

And yet new categories of work are being created at a pace that has no real precedent. AI Engineer is now LinkedIn's #1 fastest-growing job in the U.S. for the second year running, with postings up 143% year over year. The global economy has added roughly 1.3 million new AI-related roles in the last two years alone. The World Economic Forum's most-cited projection has AI creating something on the order of 170 million new roles globally by 2030, against 92 million displaced, a large net positive on paper, but one hiding an ugly detail: the people losing their jobs are, for the most part, not the same people qualified to fill the new ones. That gap is a retraining problem the size of a generation, and right now nobody has a real plan for closing it.

So: record entrepreneurship. Record AI capex. Record new-role creation. And, simultaneously, some of the largest and most AI-attributed layoffs in modern corporate history. Not a contradiction, just what a phase transition actually looks like from inside it.

What I Actually Worry About

Here's where I'll be honest about what keeps me up at night, and it isn't the jobs data, wrenching as it is. Markets absorb shocks. People adapt, retrain, start businesses: the data above proves they're already doing exactly that, in real time, faster than anyone predicted.

What I can't talk myself into feeling okay about is what happens when a government with no interest in individual liberty fully understands what this technology can do. Everything that makes AI a gift to a small retailer trying to move dead inventory: the pattern recognition, the prediction, the ability to model and anticipate human behavior at a scale no bureaucracy could ever staff is the exact same capability that makes it a gift to a regime that wants to watch, predict, and control its own population. Financial surveillance. Movement tracking. Sentiment monitoring at the scale of an entire citizenry, updated in real time, with no fatigue and no conscience. We built the tools to make dead inventory liquid and entrepreneurship frictionless. We also just built the most powerful instrument of social control in human history, and handed the on-switch to whoever gets there first.

Democracies are going to spend the coming years arguing, messily and imperfectly, about guardrails. Authoritarian regimes are not going to hold that debate. They're going to deploy.

That's the sentence that actually keeps me up. Not "will AI take my job" - some jobs, yes, and it's real and it's already happening to real people. The sentence that keeps that vies me a “Maalox Moment” is: what does an authoritarian government do with a tool this good at understanding and predicting human beings, once it fully wraps its head around what it's holding?

We are handing over life to the machine. The optimist in me sees the entrepreneurship numbers and believes we'll come out the other side stronger, freer, richer than we went in. The realist in me knows that whoever controls the machine controls an enormous amount of what happens next, and that fact alone, empire by empire and regime by regime, is the whole story of the next decade.

Best of times. Worst of times. Both, at once, right now.