cowgorithms and the claude trap

a collection of notes and links I shared with my companies recently.

Every so often, I compile the most important things I’ve recently shared with my founders and team. This week’s been an interesting one.

This week’s list is about a telling Founders Fund investment, and some reflections on my nightly 11pm-6am Claude Code sprints.

Cowgorithms

A company that makes solar-powered collars for cows just hit a $2 billion valuation.

Halter.

They make AI-powered collars that let farmers herd cattle remotely through an app.

Virtual fencing, health monitoring, “grazing optimization.”

Last year, they raised $100M at a billion-dollar valuation. This month, Founders Fund led the next round, doubling it.

The same fund that backed SpaceX, Palantir, and Anduril is now backing cattle tech.

What does that mean?

For the last fifteen years, the highest-leverage thing you could do was build software. The margins were insane, the distribution could be free, and a small team could reach millions of people from a laptop.

The edge is going away.

When AI can write code and spin up products in hours, software and “distribution” stops being the moat.

Case in point: OpenClaw getting bought by OAI, and replicated by Claude 2 weeks later.

The value goes back to the hard stuff. Atoms. Supply chains. Food systems. Energy. The things that are genuinely difficult to build and difficult to compete with once you figure them out.

And if you’ve been paying any attention to the world right now, US-Israel v. Iran is actively breaking these systems in real time.

Supply chains are fracturing. Energy markets are volatile. Food security is becoming a real concern. I just paid $6.97 for gas in San Francisco.

This is not temporary noise - it will be the new baseline.

The people building infrastructure to make food systems more resilient, energy more distributed, and supply chains more adaptable are no longer solving nice-to-have problems.

They’re solving existential ones.

A cow collar sounds funny until you realize it’s a wedge into the entire livestock industry.

One device on one animal that generates data no one else has. That’s not a gadget. That’s a platform.

The most interesting companies of the next decade are probably going to be built in industries most people in tech wouldn’t touch.

Follow the smart money.

---

The Claude Trap

During Ramadan, I found myself working until 6am every single night.

My schedule was the following:

IRL meetings and phone calls, back-to-back from ~12pm to 7pm.
Break between 7pm and 11pm.
Deep work from 11pm to 6am.

Deep work meant this:

Four to five panes of Claude Code running at the same time.

Agents spinning. It felt like I was operating at 5x capacity.

Looking back at that work, some of it turned out very useful.

But a lot of it had to be redone.

Decisions I made at 3 AM that didn’t hold up in the morning. Things I shipped that missed the point entirely.

I had fallen into the same trap I told all the founders in our portfolio to be cautious of.

I wasn’t being productive.
I was doing things to feel productive.

AI tools significantly heightened this.

You can spin up four agents and watch them all work in parallel and it feels incredible.

They’ll then generate a huge summary of the work that was done - fully shareable in Slack.

But if you’re sleep-deprived and making bad decisions about what to build, you’re just creating more work for yourself faster.

In the earliest years, hours matter. Every rep counts.

But at some point, the game changes.

The quality of your decisions starts to matter more than the quantity of your hours.

A bad hire costs six months.
A wrong company direction costs a quarter.
A poorly structured deal costs millions of dollars.

No amount of late nights will compensate for bad judgment.

And no amount of “feeling productive” will compensate for just doing the work.

---

Interesting Links

White collar goes blue
A working theory on the rights and reshaping of the professional laptop class.

Situational Awareness: The 25 year old taking on Wall St
A deep dive into Leopold Aschenbrenner’s L/S Equity Portfolio.

$10,000 Invested in Sheep: The 5-Year Profit Breakdown
Well.