The Inside Track: People, Ideas, & Stories shaping how we work, live, and build the future.

Good evening - Michael here: Here's the big business story you need to pay attention to today:

The CEO of Shopify — a $147 billion company — has personally written 957 pieces of code in the first 45 days of this year. In all of 2024, he wrote 94. What changed, and why it's sending shockwaves through Wall Street, is below.

Let me start with something that's been going viral and then explain why it matters way beyond tech Twitter.

There's a website called GitHub. Think of it as Google Docs for software engineers — it's where developers store, share, and collaborate on the code that powers apps, websites, and the tools you use every day. When a developer saves a change to a project on GitHub, it's called a "commit." It's basically a digital fingerprint that says: I built something here.

Tobias Lütke — Tobi — is the co-founder and CEO of Shopify, the company that powers millions of online stores worldwide. Shopify is worth $147 billion. Tobi runs it. He's not supposed to be writing code. He's supposed to be in board meetings and on earnings calls.

But someone pulled up his GitHub profile this week and noticed something remarkable.

In 2024, Tobi made 94 commits. Respectable for a CEO — most make zero. In 2025, that number jumped to 833. And in the first 45 days of 2026? He's already at 957.

That's more than 20 per day. He's on pace for nearly 8,000 by year's end — roughly 80 times what he did two years ago.

So what changed? The short answer: AI coding tools. Specifically, tools like Claude Code and Cursor that let you describe what you want in plain English and watch it get built in real time. You're no longer typing line by line. You're directing. You're reviewing. You're supervising an extremely fast, extremely tireless junior developer who never calls in sick.

Tobi isn't the only one. Last Monday, Spotify's co-CEO Gustav Söderström told investors something even more striking: the company's best, most senior engineers haven't written a single line of code since December.

Not one. Instead, they use an internal system called "Honk" — built on Claude Code — that lets them fix bugs, add features, and push updates to the Spotify app entirely from their phones. During their morning commute. Before they even sit down at a desk.

"An engineer at Spotify on his morning commute, from Slack on his cell phone, can tell Claude to fix a bug and add a new feature to the iOS app," Söderström said. "Once Claude finishes that work, the engineer gets a new version of the app pushed to him on Slack to test, so that he can then merge it to production — all before he's even arrived at the office."

Spotify shipped over 50 new features in 2025. It hit 751 million users. Record revenue. Record profit. And its top engineers aren't writing code anymore.

Now, here's where this gets bigger than tech.

Wall Street noticed. And it panicked.

The same week these stories circulated, investors started doing math they probably should have done a year ago: if AI tools can make one person as productive as ten, why are companies paying for ten software licenses?

Most big companies pay for business software — Salesforce for managing customers, Workday for HR and payroll, Adobe for design, DocuSign for contracts — on a per-seat basis. Meaning: per employee, per month. It's been one of the most reliable business models in tech for 20 years. More employees = more revenue.

But if AI means companies need fewer employees to get the same work done — or more work done — then those per-seat software businesses have a serious problem.

What followed was one of the sharpest selloffs in software stocks in years. Over five days, ServiceNow dropped 14%. HubSpot fell 19%. Atlassian lost 20%. Salesforce, Adobe, DocuSign, Workday — all down significantly. By some estimates, more than $300 billion in value evaporated. Wall Street is calling it the "SaaSpocalypse."

Workday, the HR software giant, lost roughly $40 billion in market value. Its co-founder Aneel Bhusri just came back as CEO with a $139 million pay package to try to save it. He called AI "a bigger transformation than SaaS." The company also laid off 375 people the same week.

This isn't just theory. Klarna, the buy-now-pay-later company, already tried this experiment. It dropped both Salesforce and Workday, replaced customer service agents with an AI chatbot that did the work of 700 people, and cut its workforce from 5,000 to 3,500. Its CEO bragged about it on every earnings call.

But here's where the counterpoints matter — and they matter a lot.

Klarna had to reverse course. Customer satisfaction dropped. Service quality got worse. The same CEO who said AI would replace everyone was recently "tremendously embarrassed" by how it played out and started hiring humans back. Their new strategy? A hybrid model where AI handles simple stuff and real people handle everything that requires judgment, empathy, or nuance. Which, it turns out, is a lot.

The skeptics are not wrong on several fronts:

An Anthropic study published January 29 found that developers who relied on AI assistance actually scored 17% lower on coding comprehension tests — they completed tasks faster but understood less of what they built. That's a problem when things break, and things always break.

One of the top comments on a Reddit thread when the Spotify story dropped: "This is an earnings call quote, not a technical report. The real translation is probably: our senior engineers spend most of their time reviewing AI output and fixing edge cases instead of typing code from scratch. Which is not that different from what senior engineers were already doing before AI." That's a fair point.

And Jason Lemkin, the founder of SaaStr — one of the most respected voices in enterprise software — pushed back hard: "Nobody is building a homegrown CRM in Replit to replace their Salesforce instance. Shipping a v1 is maybe 2% of the work. 98% is scaling, maintaining, iterating, security audits, compliance, and integrations with 500 other tools."

Goldman Sachs actually thinks the software market grows to $780 billion by 2030 — with AI agents accounting for over 60% of that. The argument is that AI doesn't kill software companies; it transforms them. The survivors won't charge per seat. They'll charge per outcome. Per result. Per problem solved.

So what does this actually mean for you?

If you're an employer: this is not about replacing people with AI. Klarna proved that doesn't work cleanly. It's about what happens when your best people use AI — they become dramatically more productive. That changes how you think about team size, about what "senior" means, about how you invest in software versus talent. The companies that win are the ones that figure out the ratio — where AI handles the volume and humans handle the judgment.

If you're an employee: the message from Shopify could not be louder. Last April, Tobi sent an internal memo — which conveniently leaked — telling every manager at Shopify that before they could request a new hire, they had to explain why AI couldn't do the job. AI proficiency is now part of performance reviews at Shopify. And Tobi isn't just talking about it — he's living it, making 20 commits a day from a company worth $147 billion. The CEO is using the tools. If you're not, that's a gap that's only going to get harder to close.

The bottom line: AI isn't replacing workers or killing software companies — not yet, and maybe not ever in the way the panic suggests. But it is changing what it means to be productive, what it means to be valuable, and how companies think about both. The gap between people who use these tools and people who don't is growing every week. And the stock market is starting to price it in.

Workday's co-founder is back — and the playbook is familiar. Aneel Bhusri returned as CEO on February 9 after the company's stock dropped roughly 40% in a year. He's calling AI "a bigger transformation than SaaS" and getting a $139 million pay package tied to turning the stock around. Same week, Workday laid off 375 people. The founder-return playbook (think: Steve Jobs at Apple, Howard Schultz at Starbucks, Bob Iger at Disney) only works when the founder actually has a plan for the disruption, not just credibility from the last era. Watch the February 24 earnings call for signals on whether this is transformation or nostalgia.

Spotify's real competitive moat might not be music — it might be taste data. Buried in the same earnings call where Söderström dropped the "no code since December" line was an equally important insight: Spotify is building a dataset that maps natural language to music preferences across hundreds of millions of users. "There is no factual answer to what is workout music," he said. "Taste is not a fact. It is an opinion." That data — what different people in different places mean when they say "chill" or "hype" or "sad" — can't be scraped from the internet. It can't be replicated by a general AI model. It can only be built by having 751 million people interact with your product every day. If you're thinking about competitive advantage in the AI era, this is the template: own the data that can't be commoditized.

The $300 billion question the market is actually asking. CNBC pointed out something that exposes the contradiction at the heart of this panic: the market is simultaneously punishing hyperscaler stocks (Microsoft, Google, Amazon) because AI spending might not generate returns, AND punishing software stocks because AI adoption will be so successful it destroys their businesses. Both can't be true. Either AI works and the infrastructure companies benefit, or it doesn't work and the software companies are fine. The selloff is pricing in the worst case for everyone. That usually means opportunity for someone paying attention.

There's a pattern emerging that I think is worth sitting with.

The CEO of Shopify is back to building. The best engineers at Spotify are directing instead of typing. And Wall Street is panicking about what this means for the companies that sell tools to everyone in between.

But the common thread isn't AI. It's leverage. These tools are giving individuals the leverage to do what used to take teams. That's not new — spreadsheets did it, the internet did it, mobile did it. What's different this time is the speed and the scope.

The question worth asking yourself this week — whether you run a company, manage a team, or just want to stay relevant — isn't "will AI take my job?" It's: "what would I build if the bottleneck wasn't capacity?"

Because for the first time in a while, it might not be.

Feedback, thoughts, suggestions? Hit the reply!

Michael Wildes is the founder and CEO of Drive Phase Holding Company, a permanent-capital firm focused on building category-defining companies across business, media (owner of Massif & Kroo), aviation, and impact. After leaving a career as a professional pilot, he spent a year as Business Editor at FLYING Magazine, writing 330+ articles on aviation's transformation. Now he builds permanent-capital companies focused on long-term trends that compound over decades. Based in Arlington, Virginia.

Connect: mikewildes.com | [email protected]