Window into the future 🔮

The productivity paradox and what it means for your team

In the 1890s, factories across New England swapped their steam engines for electric motors and waited for the productivity gains to follow. For thirty years, almost nothing changed. The technology was superior but the organisation was identical. It wasn't until the 1920s when factories tore everything down and rebuilt from scratch, with assembly lines and individual motors in each piece of equipment that electrification finally paid off.

a16z published a piece this week that opens with three sentences I haven't been able to stop thinking about: "AI just made every individual 10x more productive. No company became 10x more valuable as a result. Where did the productivity go?"

Their answer: we swapped the motor. We did not redesign the factory.

This is the question I keep bumping into in conversations with teams. Everyone has AI now. Most people are using it to do the same things they were doing before, just faster. That's real value but it's not transformation and the data is starting to reflect that.

You might have seen this chart from Ramp AI index on your feed that shows that 24.4% of businesses now pay for Anthropic up from 1 in 25 just a year ago. OpenAI's adoption actually fell for the first time ever. Anthropic now wins 70% of matchups against OpenAI among businesses buying AI for the first time.

That's a complete reversal of the trends we saw throughout 2025. Ramp's economist thinks it's partly culture. Anthropic has become, and I'm quoting here, "cool." The DoW moment from two weeks ago also sharpened a public contrast between the two companies that seems to have had real business effects.

Which is probably why Anthropic just announced a $100M Claude Partner Network this week. This network is a formal ecosystem of consultancies and agencies to help enterprises go from "we have Claude licences" to "AI is actually embedded in how we work." They're scaling their partner team fivefold. The message is clear that the bottleneck isn't the model anymore but its implementation.

And then Atlassian happened.

Their CEO, Mike Cannon-Brookes, sent a memo to 1,600 people this week telling them their roles were being cut (10% of their global workforce). The stated reason is that AI is changing the mix of skills we need. The company is "reorganising around its System of Work to move faster." In October, Cannon-Brookes said on a podcast that Atlassian would employ more engineers in five years, not fewer, they'd just be more efficient. Five months later, 1,600 emails went out.

I want to be careful here because I think the easy take "AI is taking jobs" isn't quite right either. Atlassian's share price has dropped more than 50% this year. Investors are spooked about whether AI tools will make traditional SaaS products obsolete with some calling it a "SaaSpocalypse". The layoffs look at least partly like financial restructuring wrapped in an AI narrative. Even Sam Altman called this pattern out a few weeks ago calling it "AI washing."

But here's what I think is actually true, and this is the part worth sitting with before your Monday morning:

The companies that come out of this period well won't be the ones that used AI to cut costs. They'll be the ones that used it to do things they couldn't do before. That's a completely different question and it requires a completely different kind of thinking about how work gets done.

Which brings me back to Chipotle's chatbot. Pepper can write Python. That's not a joke anymore. The question is whether your organisation is set up to take advantage of that or whether you're just running a faster version of the same old factory. (If that's where your team is stuck, this is what we built 100 School for.)

How to AI 🤖 

Every week, this section is your shortcut. Here are a couple of ways you could try AI this week that are worth your time:

• How to build skills for Claude

• How to use AI for content without creating slop

• How to use Claude's new "visualizer" feature

• How to make AI think harder (7 prompts)

• How to run a more organized workday using ChatGPT

• How to create designs with Claude Code & OpenClaw

Before you go ✌️

Quick one: does your team actually feel more productive with AI or just busier? Hit reply. I read them all.

See you next Sunday!