Introduction: The Calm Before the Storm
Cast your mind back to February 2020.
If you were paying attention then, you might have noticed fringe news about a virus spreading somewhere far away. But most of us didn’t care. Markets were booming, kids were in school, and we were planning vacations and shaking hands warmly. If someone had told you then that they were hoarding “bread,” you would have thought they’d lost their mind to the internet.
Then, in just three weeks, the entire world changed. Our offices closed, children came home, and life adapted in ways you wouldn’t have believed if they’d been described to you just a month prior.
I’m here to tell you one thing: we are now living through the “February 2020” phase for something far bigger than COVID.
A Message from the Near Future: “It Happened to Us First”
I write this as someone who spends most of their days immersed in technology and following the news of modern companies, where the future is being made. The reason tech workers are sounding the alarm now isn’t because we’re predicting the future, but because the future has already run us over.
In past years, AI development was gradual. Then, in 2025, the gates of technological hell opened. Models no longer improved slowly; the leaps became enormous, and the intervals between them very short.
In February 2026 (a date not far off), new models were released (GPT-5.3 and Opus 4.6). At that moment, I realized the water had reached my chest.
I no longer do my technical work. I simply “describe” what I want in English, and the AI does the rest. It writes tens of thousands of lines of code, builds the application, opens it, tests it, clicks buttons, finds errors, and fixes them itself… then delivers the final product to me and says: “Ready.”
What used to require a team and hours now happens with a click of a button while I sip my coffee.
The Illusion of the “Human Touch” and Taste
You might say: “But AI lacks artistic flair or human judgment.”
I used to say that too. But the new models broke this rule. For the first time, we see AI possessing something akin to “taste.” It makes decisions that are not just technically correct, but “contextually” correct. That elusive feeling of knowing the right decision, which we thought was exclusive to humans? The machine has acquired it.
Why Does This Matter to You as a Pathologist?
You might think this is just programmer talk. You’re wrong.
AI companies focused on “programming” first for a strategic reason: if you make AI capable of writing code, it will be able to program a smarter version of itself. And that’s what happened. AI is now contributing to building the next generation of AI.
Now that they’ve finished with “programming,” they’re coming for everything else: law, finance, and the biggest catastrophe… medicine.
The experience programmers lived through last year – where AI transformed from a “helper tool” to “a tool that does my job better than me” – is the experience you will very soon live through in tissue diagnosis and slide reading.
Don’t Judge Based on 2024 Experience
I hear you saying: “I tried ChatGPT, and it hallucinated and gave wrong information.”
You are judging today’s technology by “ancient history” standards. The difference between 2024 models and today’s models is like the difference between a landline phone and an iPhone.
If you’re using the free version, you’re two light-years behind. Top lawyers, CEOs, and informed doctors aren’t using these tools for play; they’re using them because they accomplish in minutes what takes hours.
Terrifying Acceleration: The Numbers Don’t Lie
2022: AI makes calculation errors (7 x 8).
2023: Passes the bar exam.
2024: Writes programs and explains complex sciences.
2025: Performs the work of software engineers.
2026: Models outperform most PhD holders.
There are institutions that measure the “time” it takes for AI to complete a complex human task without assistance. A year ago, it was 10 minutes. Today, it’s 5 hours of continuous, independent work. And the curve is rising vertically.
Do you really think an “office” job (and pathology diagnosis is cognitive office work in front of a screen) will be safe?
The Intelligence Explosion: Intelligence Builds Intelligence
The most dangerous aspect is what was recently stated in OpenAI’s technical documents: “Our new model helped build itself and correct its errors during training.”
Read that sentence again.
We’re not talking about science fiction. We’re talking about a “Feedback Loop.” Each generation builds a smarter generation, and the smarter generation builds the next one faster. Experts call this an “Intelligence Explosion.”
What Does This Mean for Your Job?
I’m going to be frank with you because you deserve honesty more than comfort.
Dario Amodei, CEO of Anthropic (and one of the most cautious), predicts that AI will displace 50% of cognitive jobs within a few years.
AI doesn’t replace a single “skill”; it replaces “cognitive work” as a whole.
Legal work: writes contracts and analyzes cases.
Medical analysis: reads scans, analyzes slides, suggests diagnoses, and reviews medical literature at a level comparable to or exceeding humans.
My golden rule now: if a model shows a “glimpse” of capability today, the next generation will be “proficient” in it.
Will it replace human empathy? Perhaps not. But if the core of your job is sitting in front of a screen, analyzing images, and writing a report… then AI is coming to take a large part of that work.
What Should You Do? (Roadmap for Survivors)
I’m not writing this to scare you, but to give you the only remaining advantage: primacy.
Use paid tools immediately: Don’t judge AI by the “dumb” free versions. Pay the $20. It’s an investment in your professional survival.
Don’t treat it like a search engine: This is the fatal mistake. Don’t ask it quick questions. Ask it to “work.” Give it a complex case report, ask it to analyze variations, ask it to write code to analyze a slide image. Push it to its limits.
Be “the one who knows”: There’s a very narrow window of time now. The person who walks into a department meeting and says, “I used AI to complete this analysis in an hour instead of 3 days” will become the most important person in the room. Be that person before someone else beats you to it.
Flexibility is the only skill: Tools change every month. The skill isn’t in mastering a specific tool, but in the “muscle of adaptation.” Learn how to learn quickly.
The One-Hour Commitment: Dedicate one hour daily to trying something new with AI. Not to read about it, but to use it. If you do this for 6 months, you’ll be ahead of 99% of your colleagues.
Conclusion: The Future is Knocking
We’ve moved past the stage of “pleasant dinner conversations” about the future. The future is already here.
The train has left the station, and the speed is increasing wildly.
The people who will emerge from this phase safely are those who approach it not with fear, but with curiosity and urgency.
Don’t wait until you read the headlines in 6 months when it’s too late. Start today.
