The AI Revolution is Here: A Warning and a Guide from Matt Shumer

Remember February 2020? A few whispers about a virus overseas. Stockpiling toilet paper seemed…extreme. Then, in three weeks, everything changed. I believe we’re in the “this seems overblown” phase of something far bigger than COVID-19. Having spent six years building an AI startup and investing in the space, I’m writing this for those who aren’t immersed in this world – the honest version, even if it sounds unsettling.

The truth is, even within the AI industry, influence over what’s unfolding is concentrated in a remarkably small group: a few hundred researchers at companies like OpenAI, Anthropic, and Google DeepMind. Most of us are building *on top* of their foundations, watching the ground shake just like you, but perhaps feeling the tremors first.

We’re Not Predicting, We’re Reporting

This isn’t about predictions. It’s about reporting what’s already happening. For years, AI improved steadily. But in 2023-2024, new techniques unlocked a dramatically faster pace of progress. This year, something fundamentally shifted. It wasn’t a light switch, but the realization that the water has been rising around you – and is now at your chest.

I’m no longer needed for the technical work of my job. I describe what I want built, in plain English, and it simply…appears. Not a rough draft, but a finished product, often better than I could have created myself, requiring no corrections. Just months ago, I was guiding the AI, making edits. Now, I describe the desired outcome and step away.

A Real-World Example

Let me illustrate. I tell the AI: “I want to build this app. Here’s what it should do, roughly how it should look. Figure out the user flow, the design, everything.” And it does. It writes tens of thousands of lines of code, then *tests the app itself*. It clicks buttons, uses features, and iterates on its own design, refining until it meets its own standards. Only then does it present it to me for testing – and it’s usually perfect.

That was my Monday this week. I’ve always been an early adopter, but the last few months have been shocking. These aren’t incremental improvements; this is a different order of magnitude. The experience of tech workers watching AI go from “helpful tool” to “does my job better” is about to become universal.

The Impact Across Industries

Law, finance, medicine, accounting, consulting, writing, design, analysis, customer service… not in 10 years, but within one to five years. Some say even sooner. The market reacted this month, wiping out $1 trillion in software value in a single week. And I foresee more disruption, and soon.

If you tried ChatGPT in 2023 or early 2024 and found it unimpressive, you were right. Those versions were limited, prone to “hallucinations.” Today’s models are unrecognizable. The debate about whether AI is “really getting better” is over. Anyone still arguing that point hasn’t used the current models or has a vested interest in downplaying the progress.

The gap between public perception and reality is dangerous because it hinders preparation. Most people are using the free versions of AI tools, which are significantly behind the paid versions. Evaluating AI based on a free-tier ChatGPT is like judging smartphones by a flip phone.

What This Means for You

Dario Amodei, a leading voice in AI safety, predicts AI will eliminate 50% of entry-level white-collar jobs within one to five years. Many in the industry believe he’s being conservative. This isn’t just about automating specific skills; AI is a general substitute for cognitive work. It improves at *everything* simultaneously.

If your job involves reading, writing, analyzing, deciding, or communicating through a keyboard, AI is coming for significant parts of it. And eventually, robots will handle physical work too.

Adapt or Be Left Behind

I’m not writing this to instill fear, but to emphasize the advantage of being early. Start using AI seriously, not just as a search engine. Sign up for the paid version of Claude or ChatGPT ($20/month). Use the most capable model available (currently GPT-5.2 on ChatGPT or Claude Opus 4.6 on Claude). Follow Matt Shumer on X for updates on the best models.

Don’t just ask quick questions. Push AI into your actual work. If you’re a lawyer, feed it a contract. If you’re in finance, give it a spreadsheet. If you’re a manager, paste in your team’s data. The people who are getting ahead aren’t using AI casually; they’re actively automating parts of their jobs.

Prepare for the Future

Get your financial house in order. Build savings, be cautious about debt, and maintain flexibility. Focus on skills that are harder to replace: relationships, physical presence, licensed accountability, and navigating regulatory hurdles.

Rethink what you’re telling your kids. The traditional path – good grades, good college, stable job – points directly at the most vulnerable roles. Teach them to be builders, learners, and to pursue their passions.

This might be the most important year of your career. Learn these tools, get proficient, and demonstrate what’s possible. The window of opportunity won’t stay open forever.

This article is adapted from shumer.dev.