@CSUnerd

Forward Deployed Nerd. Weekly extended write-ups on Software, AI & Investing. https://xitter.cf/t.co/wCDQ1Jx4dv

productandinvestingnerd.substack.com/
Joined July 2014
  • Tweets20.7K
  • Following5.3K
  • Followers2.6K
884Media
The old software saying was garbage in, garbage out. AI changed the math on that. Now it is garbage in, a whole junkyard out, probabilistically.
1
4
308
Full piece here: productandinvestingnerd.subs…
92
I once argued Uber's moat is wide but shallow. After listening to Dara Khosrowshahi on Invest Like the Best, I owe the counter-argument its due.
1
3
576
more replies
Here's the moat inversion. In the old model, Uber's moat was the driver network, and AVs threatened it. In the new model, the moat is the demand network, and AVs strengthen it. Every new AV operator needs access to it. Network AVs are 30% busier than independent ones.
1
127
The aggregator wins when the supply side fragments and commoditizes. That's the whole bull case. And the whole risk: if AV supply concentrates instead of fragmenting, the aggregator becomes a middleman nobody needs.
1
77
😬 $META
Alexandr Wang
@alexandr_wangOct 18
just got the weekly status report from the team
1
315
In engineering school in the oughts, I took the road less travelled. I wanted to apply science to how someone uses something. My fellow students brushed it off. Fuzzy math, not real engineering.
1
3
334
Back then, how a gadget worked in the wild had no bearing on how it was engineered. The builders of things didn't design for use in the wild. Not truly. Not as a science.
1
2
121
Fast forward. UX is in everyday conversation now. It's in our vehicles. I had a front-row seat for that transition, from shunned to expected. Which makes me wonder: what are we overlooking in 2026 that will be a given in 2046?
1
83
Palantir and Constellation Software look nothing alike. But they run on the same first principle. Both reject the standard Silicon Valley playbook in favour of two shared truths.
1
3
332
more replies
Truth two: complexity as a moat. Most SaaS avoids messy work. Palantir and CSU run toward it. Solving the gritty problems of the real world is the only way to build an enduring monopoly.
1
1
148
The tech industry is finally catching up to what Mark Leonard and Mark Scianna realized long ago. The best software is not built working from home. It's built on the front lines.
1
92
I ran the $100B test on Uber. If I gave you $100B to displace Uber in a world where AVs are the primary supply, could you? The brand, the routing engine, dozens of AV partners, financing and insurance. That's a harder pitch than I first made.
1
1
370
Overlooked doesn't mean absent. Steve Jobs, Xerox PARC, university labs, they were all working on UX for decades. Overlooked means it's not part of how we do things yet. It's the pink elephant in the room before it turned pink.
2
271
If a Forward Deployed Engineer is an engineer embedded with a customer, then a Constellation Software business unit is a Forward Deployed Company. HQ sets the hurdle rate. Pricing, hiring, roadmap: all left to the manager closest to the action.
1
4
655
Uber isn't building autonomous cars. It's building the marketplace the autonomous cars need to be useful. This is the Levi Strauss play updated for robots. Strauss didn't mine the gold. He sold the pants to the miners.
1
1
401
Largely mostly true. $CSU $CSU.TO
Nikita Bier
@nikitabierOct 12
Technical moats are mostly over at this point. If you’re building something that is meant to be durable, the only moats now are: • Proprietary data • Network effects • A physical component
1
15
2,680
When I was in engineering school in the oughts, applying science to how someone uses something was dismissed as fuzzy math, not real engineering. Now UX is in everyday conversation. So here's my question: what are we overlooking in 2026 that will be a given in 2046?
2
407
Load more