Matthew Waller

If this continues, the question shifts from: “Is AI interesting?” to “Is AI beginning to shift the economy’s growth frontier?” Still early. But worth watching closely.

Recent data are directionally consistent with that pattern: • Near-5% productivity growth • AI-linked capital formation driving a large share of GDP growth • Inflation holding in the mid-2% range That doesn’t prove causality. But the signals are stacking.

Three recent data releases are worth looking at together. Individually, they’re interesting. Taken together, they may suggest something more structural is happening in the economy. 🧵👇

@CashionEast Great to hear from you, Cash!

@supply_chain This is the kind of analysis we need to really understand the impact of AI integration. Won't be linear. Won't always be obvious. But it's impact will be best understood with data driven thinking.

I’ve been paying close attention to AI, trying to separate signal from noise. There’s no shortage of opinions: “AI isn’t delivering yet” vs. “AI is changing everything.” I don’t see either extreme. But I do see a few economic signals worth paying attention to.

None of this proves AI ROI is universal. But when a technology starts to show up in: • productivity data • unit costs • GDP composition …it’s usually a sign that experimentation is giving way to diffusion. Worth watching closely.

Capital doesn’t chase demos. If AI-related investment is that large a share of GDP growth, it suggests firms are moving beyond pilots and committing real capital to infrastructure, workflows, and capability build-out. That’s a different phase of adoption.

I've been paying close attention to AI and trying to separate signal from noise. I read a lot of opinions. Everything from "AI isn't delivering real benefits yet" to "AI is already changing everything." In my own work, I get a window into how companies across the size spectrum are actually using it. My view is partial, but it's informed by real deployments. And my impression is that over the past year or so, AI is being applied more effectively than the press sometimes suggests. Less toy demos, more workflows that genuinely change throughput and decision speed. Still, one macro signal caught my attention. BLS data shows nonfarm business productivity increased at a 4.9% annualized rate in Q3 2025, while unit labor costs fell 1.9% (the first consecutive quarterly decline since 2019). That combination is unusual. Productivity surging and unit costs falling at the same time is what you'd expect if AI is starting to function as an efficiency shock rather than just a cost center. This doesn't mean every firm has cracked the code. But it might suggest that, in aggregate, AI-enabled tools are now improving output per hour fast enough to show up in official statistics. To me, that points to a real possibility: AI may be crossing the line from experimentation to economic impact. Curious what others are seeing. Where has AI genuinely improved throughput, decision quality, or speed? And where is it still hype?

I just got @supply_chain from the X Handle Marketplace! Get your own at handles.x.com

People ask for one rule to live by: move daily, sleep regularly, build muscle, avoid sugar, seek polyphenols, treat ultra-processed foods like poison, find a supportive partner, call your parents, and have purpose in life I know. That’s the point

In 1776, 3 things happened: The Declaration, Adam Smith’s Wealth of Nations, & James Watt’s commercialization of the steam engine. Each was momentous on its own. But when combined, they transformed the world by forming the basis for a system of democratic capitalism based on individual rights.

The one brain structure that literally grows when you do things you HATE doing — and shrinks the moment you get comfortable. Stanford neuroscientist Andrew Huberman dropped what he calls “one of the most important discoveries in the history of neuroscience” in a conversation with David Goggins. It’s called the anterior mid-cingulate cortex (aMCC). Recent human studies (not mice) show: - It’s smaller in people with obesity → grows when they successfully diet - Super-athletes have an unusually large aMCC - People who live the longest keep this area big their entire life - It enlarges every single time you force yourself to do something you genuinely do NOT want to do - It shrinks almost immediately if you stop or if the same task becomes enjoyable Huberman: “This isn’t the seat of intelligence or memory. This might actually be the seat of the WILL TO LIVE.” The rule is brutal but simple: If you love your ice bath → no growth If you’re terrified of cold water but get in anyway → aMCC gets bigger Skip a day or start liking it → it shrinks again tomorrow Huberman waited years to tell David Goggins about this because Goggins has been unconsciously training his aMCC harder than almost anyone alive. Watch the full clip below — it will permanently change how you think about discomfort, willpower, and longevity. What’s one thing you really don’t want to do today… that you’re going to do anyway? Drop it in the comments. Let’s build that aMCC together.

Being kind boosts mental health more than seeking joy. Evidence: Doing 3 random acts of kindness a week is enough to reduce depression, anxiety & loneliness. It's more beneficial than doing nice things for yourself. Self-care feels good, but generosity builds lasting bonds.