A letter to recent graduates, the kids of Gen X parents

I graduated in the early 90s into what economists now call the jobless recovery. US unemployment peaked at 7.8% and stayed stubbornly high long after the recession was technically over. The recovery was so weak it cost a sitting US president his re-election. Whatever country you graduated into, the door looked closed.

And we got the future wrong. We assumed it would be a smaller, harder version of the past. Fewer jobs, same jobs.

It wasn't. By 1996, four years after I graduated, the ground had shifted under everyone. Windows 95 had arrived. Netscape had IPO'd. Every business on earth was scrambling to get online and would hire literally anyone who understood the internet. It was a people & culture nightmare. Friends who had felt unemployable in 1992 were running digital teams by 1997. People with English degrees became "webmasters." People with no degree at all became network administrators because they had taught themselves on a Pentium in their bedroom. Many jumped ship from large consultancies to start businesses, VC in Boston was a cool place to be (I'm from the East Coast and can only imagine being from CA!). Science start-ups, SaaS (before it meant how you pay), Online...it was all new and cool to be in it.

The roles that absorbed our generation did not exist when we sat our final exams. None of our career advisors mentioned them. None of our parents could help us prepare for them. None of the experts in that Channel 15 news report saw them coming. We made them up as we went.

Sit with that for a second. The thing that saved our generation was a category of work that nobody could have told us to study for.

I am not going to insult you by pretending it is identical. It isn't. In 1996, the new technology needed humans to deploy it. Armies of us. Anyone who could install software, configure a network, build a website, or explain email to a CEO was employable almost overnight. The wave created entry level jobs faster than it removed them.

AI in 2026 is doing the opposite. It is targeting exactly the structured, well scoped, entry level work that used to be the training ground. The bottom rung of the ladder, the one I climbed onto in 1992, is being quietly sawn off in a lot of industries. The average tech internship now attracts 273 applicants. Graduate unemployment in the US is at levels not seen since the financial crisis.

You are not imagining it. The thing you are feeling, that the rules have changed under your feet, is real.

In 1996, two students in a garage could compete with IBM. A cheap PC and a phone line were enough. That is why my generation produced Amazon, Google, eBay and Salesforce, companies started by people who were, frankly, just messing around so to speak.

In 2026, the underlying tools and infrastructure to build the next frontier model cost billions of dollars. The economic moat used to be talent. Now it is compute and energy. Power has moved from people who can write the software to people who can afford to run it. But sitting on top of the expense is that, anyone can build a business if they find a problem to solve. The mountain it sits on is pricey but it's easier to solve problems than ever before.

That sounds like bad news. It mostly isn't. But it changes how you play your cards.

In my era, graduates and early career employees built capacity through grunt work. Yours can't, because the grunt work is being automated. The graduates winning right now are the ones who walked in on day one able to use AI to perform at mid level capacity. They are not faster typists than the seniors. They are better orchestrators. Think of yourself as a conductor, not a player. The most useful skill you can develop in your first two years is to look critically at all information (yes, including the news), don't take it as true. Having an authentic judgement skill is the one thing AI does not have.

In my era, a side hustle was something to hide. Most employment contracts explicitly forbade it. If your boss found out you were freelancing or running something on the weekend, you'd get a warning, or worse. Loyalty was measured in singularity. You worked for one company and you didn't dilute that with your own thing.

That rule is dead, and it died for a reason. In 2026, a side hustle says the exact opposite of what it used to. It says: I can identify a problem, ship a solution, find customers or users, iterate based on feedback, and use AI to do the work of a team I cannot afford to hire. It says I am the kind of person who builds rather than waits to be told. It says I have already developed the orchestration skills employers are now desperate for, and I have done it with my own money and my own time, which is the only proof that matters.

Hiring managers I've spoken to recently now actively scan for it. A candidate with a credible side project jumps to the top of the pile. Not because of the revenue, which is usually small. Because of what it proves about the candidate. It is the single fastest signal in the modern job market that you can solve complex problems end-to-end.

Build the side hustle specifically to figure out AI. Pick a problem you actually have or see, for yourself, for a community you belong to, for a small business owner you know, and try to solve it with AI tools. You will learn more about how these systems actually work in three months of trying to ship something than you will in three years of courses. And by the way, that learning is portable. Whatever AI thing you build for yourself becomes the proof point that gets you the role.

In 1995, the smartest thing my friends did was follow the build. Internet infrastructure was being laid down at scale across North America, Europe and Asia and nobody had enough people to do it. In 2026, the equivalent is the AI build, data centres, energy, integration, evals, governance, and application-layer companies in every vertical that hasn't been touched yet. The capex isn't slowing. It's accelerating. That money will need humans attached to it.

You probably can't go and build the next OpenAI or Anthropic. You can build the company that uses AI to transform an industry that hasn't been transformed yet. That is exactly what the great companies of the 90s did with the internet. Amazon was not a chip company. Google was not a telco. They just took a new platform seriously, earlier and more honestly, than the incumbents did. The same playbook is sitting there for you right now, in dozens of industries, in every country.

After the 2001 dot com crash, computer science majors were told the industry was finished. By 2004, hiring was back at pre crash levels. Markets like this one move faster on the way out than the way in. The graduates I admire most from my own cohort were the ones who kept moving, kept learning, kept making, kept reaching out, even when the door looked shut. The ones who waited for permission lost ground that took years to recover. The way to beat it is to refuse to stand still.

In my era we resented the boomers who seemed to have everything. It cost us. The people who got ahead fastest were the ones who attached themselves to a senior who would teach them, advocate for them, and pass them work. In a market where the bottom rung is being removed, mentorship is the workaround. Find people who have been through a downturn before. Ask them how they did it. Most of us will say yes, because we remember being where you are.

Here is my prediction, and I will stake my reputation on it. As AI absorbs more of the structured, derivative, output-driven work, what is genuinely human will rise. The things you might be tempted to dismiss as soft, or unproductive, or not "career relevant" are about to become the most valuable things you own. Get creative. Find a hobby that has nothing to do with your degree. Learn an instrument. Take a pottery class. Hand write a letter and actually mail it. Write a book without AI, actually write it, with all the false starts and bad chapters and weeks of being stuck, because the muscle you build doing that is the exact muscle you will need to do anything complex in the next twenty years. The graduates who will outpace their peers are not the ones who learn to prompt the best. They are the ones who develop original taste, who can hold ambiguity for long enough to find an unexpected answer, who can sit with a hard problem and not flinch. AI will commoditise the obvious. Your job is to become the kind of person who can produce the non-obvious.

Here is something companies used to do, and need to start doing again. When I graduated, firms like Andersen Consulting, CSC, EDS and Price Waterhouse hired huge cohorts of new graduates and put us through their training centres for four to six intensive weeks. They took French majors, history majors, international business majors like me, kids with no technical background whatsoever, and taught us to code. Cobol, then C, then whatever the next thing was. They invested in us before we had proven anything, because they understood that the next generation of consultants had to be built, not just hired. And while we were learning to code, something else happened that no business school would have predicted: we made friends for life. The people I sat next to in those windowless training rooms in Illinois are still in my network thirty years later. They became co-founders, clients, references, allies.

The training centre was a career accelerator and a relationship machine at the same time.

Companies today are doing the opposite. They are afraid, they are laying off in droves, and they are quietly assuming that someone else, universities, bootcamps, the candidates themselves, will produce the trained graduates they need. That is a strategic mistake, and they will pay for it within five years when the senior talent they failed to develop is simply not there. It is time to take a page out of the old book. Invest in the next generation. Hire the French major. Train them. They will pay you back for the rest of their careers, and they will build the friendships and networks that become the connective tissue of your industry.

I know what it feels like to send out a hundred applications and hear nothing. I know what it feels like to wonder if you have already lost the game before it has properly started. You haven't.

Here is the funny thing about that Channel 15 news report from 1991. Every one of the worried experts in it was wrong, not about the difficulty, which was real, but about the verdict. The cynical kids on camera, the ones the voiceover described as anxious and uncertain, went on to invent the modern internet, build the largest companies in history, and define an entire era of culture. The future was in doubt. It just wasn't the future they thought.

Yours won't be the future you think either. The categories of work that will define your career probably do not exist yet. The companies that will hire you for the most interesting jobs of your life probably haven't been founded yet. The version of you that gets to thirty, looking back at this year, will likely tell a story about it that you can't possibly write today.

Your job, right now, is not to land the perfect role. It is to stay in motion. Keep learning. Keep building. Ship the side project. Find one person a week who is doing something interesting and tell them so. Use the tools that scare your parents. Keep one eye on the ethics and the rules, it's important to figure that out too. The market will turn, it always does, and when it does, the people who kept moving in the quiet years are the ones who are ready for it.

My generation graduated into a closed door and built the internet. Yours is graduating into a closed door and will build whatever the next thing is.

I don't know what it looks like yet. Neither do you. That is the point. That is the gift, even when it doesn't feel like one.

We are watching. We are rooting for you. In 35 years, when some kid uploads a clip of a 2026 report about how your generation has no future, watch it back. Laugh. Then go tell the kids of that era what you wish someone had told you.

With love, and a quiet kind of hope and excitement,

A jobless graduate of the class of '92.