“What about this internet thing? Do you know anything about that?”
David Letterman asked Bill Gates that question in 1995. Gates explained that the internet was becoming a place where anyone could publish information, access the latest news, connect with people across the world in real time. “It’s the big new thing,” Gates said.
Letterman wasn’t convinced. He pointed out that you could now listen to a baseball game on your computer. “Does radio ring a bell?” he asked, to a laughing studio audience. Gates said the difference was you could listen whenever you wanted, not just when it was broadcasting. Letterman shot back: “Do tape recorders ring a bell?”
The audience roared.
Thirty years later, that clip lives on YouTube. Letterman’s entire show is on YouTube. You’re reading this on a device connected to the internet, almost certainly with fifteen other tabs open.
Most people know they’re not supposed to be Letterman in that story. What’s harder to reckon with is that right now, in 2026, most people are.
AI is not a future conversation. It’s a present one. And the people running the largest companies in the world are not being subtle about it.
Dario Amodei, CEO of Anthropic — the company behind Claude, currently one of the most widely used AI models across professional and enterprise teams — said earlier this year that AI will write 90% of code within three to six months, and all of it within a year. He also said he has “a fair amount of concern” about what that means for employment. That’s not a casual comment from someone in his position.
Mark Zuckerberg told Joe Rogan that Meta is building AI to function as mid-level engineers. Spotify’s co-CEO said in February 2026 that their best developers have not written a single line of code this year. All of it handled by AI.
These are not predictions. These are descriptions of what is already happening inside some of the most closely watched companies on earth.
Then there’s what Jack Dorsey did 24 hours ago.
Block — the company behind Square, Cash App and Afterpay — cut roughly 4,000 employees. About 40% of the company’s total workforce, gone. Not because the business was in trouble. Gross profit was up 24% year on year. Cash App revenues up 33%. The company hit the Rule of 40 for the first time in its history. Block was, by every financial measure, doing well.
Dorsey wrote to shareholders: “Intelligence tools have changed what it means to build and run a company.” He added that in December 2024, the models became an order of magnitude more capable, and that most companies haven’t caught on yet. Block’s stock rose 24% on the announcement.
A profitable company let go of nearly half its people, the market celebrated, and the CEO said this is just the beginning.
Is This the Last Year of Meaningful Work?
It’s a fair question and it deserves a straight answer: probably not. But it may be the last year you get to decide how you respond to this on your own terms.
The entry-level engineering pipeline is already under pressure. The bootcamp-to-junior-dev path, the CS degree as automatic ticket into a stable career, the junior developer role as standard on-ramp — those aren’t disappearing overnight, but they are changing faster than the institutions built around them can respond. Microsoft’s own restructuring has hit software engineering harder than any other department. These are internal signals from a company that builds the tools developers rely on every day.
The pace is what catches people off guard. Mustafa Suleiman, Microsoft’s head of AI, has put the timeline for significant restructuring of white collar work at 18 months. Not eventual disruption. Eighteen months.
That said — and this part matters — the same forces making some roles redundant are handing individual people capabilities that didn’t exist before. Someone who works well with these tools can do what a team used to do. That’s not spin. That’s what the tools actually allow right now.
So what do you actually do with this?
The practical answer is to start using the tools. Not to experiment with them in your free time and feel like you’ve covered your bases, but to genuinely integrate them into how you work.
If you write code, Cursor and GitHub Copilot have moved well past autocomplete. They reason through problems, refactor entire codebases, catch bugs before you do. If you work in research, analysis or content, Claude and Perplexity are not just search replacements — they handle multi-step reasoning tasks that used to take hours. Perplexity Computer, launched this month, runs autonomous multi-agent workflows using Claude Opus 4.6 as its core reasoning layer. One person with that setup can output what small teams were producing a year ago.
The people who will be most affected by AI are not the ones in the most technical roles. They’re the ones who decide to wait and see. The ones who stay in the audience.
The people who will do well are the ones who treat these tools the way early internet adopters treated the web — not as a novelty to observe, but as a medium to work inside. AI won’t make you irrelevant. Being slow to engage with it might.
Dread it. Run from it. It arrives all the same.
The question was never really whether AI would change work. That’s settled. The real question is what you do in the time between now and when it becomes impossible to ignore — and whether you use that time or spend it in the audience.
The good news is that time still exists. Use it.