Most people still think of AI as a chatbot that can answer basic questions, write you a mediocre email and occasionally hallucinates (gives you the wrong answers or makes something up). That was true two years ago, and in some cases it was true just 6 months ago.
The gap between what AI can actually do right now and what the average person believes is enormous and growing larger every day.
I’m not writing this to hype AI up or to scare you about AI taking your job. I’m writing this because AI is inevitable. At this point, AI is here to stay and if you ignore that you will get left behind.
Unprecedented Speed
No technology in history has moved at this pace. Not the internet. Not mobile. Not cloud computing. The difference with AI is that every few months there is an improvement that makes the previous state of the art model look like a toy.
What makes this uniquely disorienting is that most people form their opinion of AI based on a single interaction they had months or even a year ago. They asked ChatGPT something in 2024, got a mediocre answer, and decided AI was overhyped.
That's like using dial-up internet in 1997 and concluding the web would never amount to anything. The thing you tried back then and the thing that exists today are barely the same technology.
The rate of improvement is not slowing down. If anything, it's accelerating. And the people who aren't paying close attention have no idea how far things have come.
Timeline
2022 — ChatGPT launches and goes viral. Millions of users in days and people are interacting with AI for the first time. Seeing the words appear on the screen feels like magic at first, but it still makes obvious mistakes.
2023 — GPT 4 is released. It can pass the bar exam and SAT. Impressive, but the everyday use cases are still limited.
2024 — AI agents emerge. Agents can use the LLMs such as GPT 4 to browse the web, use tools, write and execute code, and chain together multi-step tasks. AI starts writing basic functional code, but can still not be trusted to produce production-quality code.
2025 — Reasoning models and further improvements with agents. AI can now think step-by-step through complex problems and self-correct. We start to see some real autonomy.
Late 2025 — We see our first frontier models. Models that can write better code than humans, can automate complete jobs (not just tasks).
2026 — Where we are now. Models are writing full features, managing codebases, conducting research, and producing work that is genuinely difficult to distinguish from a senior professional's output.
That is four years. Four years from "cool trick" to "this thing might be better than me at parts of my job." No one was ready for that timeline, and most people still haven't caught up to the reality of where things stand today.
Late 2025 it all changed. The release of GPT-5.2 and Opus 4.5 (Claude) were truly the first frontier models. All the other models before them were good, but you couldn't rely on them to complete your tasks without some refactoring from a human. The frontier models is where some of the best coders in the world finally adopted vibe coding and they were blown away.
Opus 4.5 was released on Novermber 24th, 2025 and GPT-5.2 Codex on December 18th. Opus 4.6 and GPT-5.3 were released on February 5th, 2026. Proof that the models are getting better and the release dates are getting closer together.
The models are so capable now, they are building newer and better models. Let that sink in.
Software Engineering is the First Being Impacted
If you want to see where AI is headed for every industry, look at what's happening in software engineering right now. We're the canary in the coal mine.
I use AI every single day in my work. It handles most of the technical work now such as writing code, debugging, and documentation. I can simply write better code than I can at a fraction of the speed.
It is addicting, fun, and scary all at the same time. The possibilities of what to build and what can be built seem endless.
But there's a deeper reason software engineering is being impacted first, and it's not just because the feedback loop is tight. It's strategic. If the AI labs can get AI to write code better than anything else, then AI can build better AI. It's a flywheel. Better code models produce better training infrastructure, better evaluations, better tooling — which produces better models.
This is already happening. Here's OpenAI on the creation of GPT-5.3 Codex:
"GPT‑5.3‑Codex is our first model that was instrumental in creating itself. The Codex team used early versions to debug its own training, manage its own deployment, and diagnose test results and evaluations—our team was blown away by how much Codex was able to accelerate its own development."
Software engineering isn't just the first domino. It's the domino that knocks over every other domino faster. Every knowledge-work industry is next.
All Industries Will be Impacted
While software engineering has rapidly adopted AI, most industries are still barely scratching the surface. At least in the public eye.
A week ago, the public found out that Anthropic (the creators of Claude) have spent the past 6 months working with Goldman Sachs to automate back office work.
The CEO of Anthropic has verbally mentioned serval times that most knowledge work will be automated as early as 12 months from now. Sometimes he is even more aggressive with his timeline.
Again, AI is here and here to stay. Just like software engineering is being impacted now, you will start to see other industries such as legal, marketing, sales, healthcare, etc. have parts (if not all) of their work completely automated by AI.
So What’s Next?
If you’ve made it this far in my article, you understand that AI is a world altering technology. If you don’t adapt in this new world, you will get left behind.
So, how do you not get left in the dust?
It is simple. Learn how to use AI.
I’m not saying learn how to build LLMs or AI agents. I’m saying learn how to harness these powerful technologies to stand out over your colleagues that refuse to use them. Integrate AI agents into your business to automate the “boring” or “tedious” work so you can focus on the money making aspects.
AI is a powerful tool that can give you time back in your life to do what you truly enjoy.
That is handling back office work and paperwork so you can focus on creative tasks for your graphic design business.
It might be ranking and scoring inbound leads for your insurance agency so that you can focus on the quality leads rather than spending hours chasing a lead that will only churn 6 months from now.
It might be creating dozens of pages of marketing copy that saves you days of work.
All of this is possible now with personalized AI agents. You can feed agents a knowledge base of your company, branding, writing style, documentation, past successes and failures. It keeps a memory of your interactions so that it knows what you want and expect. It becomes an “employee” that does not complain, take PTO, or create turnover.
Final Thoughts
I've been in tech long enough to know that not every trend lives up to the hype. I've seen blockchain promises fall flat. I've watched VR headsets collect dust. I know what overblown hype looks like.
This is not that.
AI is the most significant technological shift of our lifetime, and we are still in the early innings. The models are going to keep getting better. The agents are going to keep getting smarter. The people using them are going to keep getting faster. And the gap between adopters and everyone else is going to keep growing until it's no longer a gap.
You don't have to be an expert. You don't have to understand transformers or attention mechanisms. You just have to start using it. Consistently. Intentionally. Starting now.
The future belongs to the ones who adopt. Everyone else is choosing to be left behind.