Dad, husband, dog dad to Loki. I write about AI, product, tech & leadership. Views my own. New Mountain Capital, ex Amazon, GE & Koch Industries, Cornell alum
Relationships still win deals in PE.
That’s not changing.
But the firm that shows up to that relationship with better signal wins more often.
AI is not replacing your network.
It’s making your network more valuable.
The winners in AI aren't the ones with the best model — they're the ones who figure out where the leverage actually sits. If McKinsey is mapping 4 distinct agentic roles, it means the build phase is officially underway. Pick one and go deep. 🚀
Agentic AI is changing how tech services create value.
We’re starting to see four distinct roles take shape, each with a different set of capabilities, bets and trade-offs.
The question isn’t whether to play but where to focus and how to build around it. https://t.co/Kg9LETGDur
In honor of 50 years of Apple, we're sharing - for the first time ever - Don Valentine's original 1977 memo for Sequoia's investment into Apple Computer. #Apple50
@andrewchen running this playbook right now. email automation is live. async AI updates are replacing half my 1:1s. hiring is the hardest one — not because AI can’t do it but because no one wants to own the decision when it goes wrong.
How do you actually apply "don't use AI here"?
Simple test I use: Ask — "If this decision is wrong, what's the downside?"
If it's recoverable, automate it.
If it shapes trust, reputation, or major financial outcomes — keep a human in the loop.
AI on the work. Humans on the judgment that matters.
@SahilBloom Great read, thanks for posting. Truth is always in the middle, the Industrial Revolution destroyed jobs but created millions more. The question isn't 'will jobs disappear' ... it's 'what new jobs will emerge that we can't yet imagine?
Just sayin…. Most of us are still running “AI transformation initiatives”: new platforms, new model, training sessions, change management programs.
The ones winning? They’re quietly eliminating administrative tax while their people work exactly like they always have.
The best AI implementations are the ones people don’t even realize they’re using.
Your people are already generating insights in emails, Slack, meeting notes.
The AI opportunity isn’t making them work differently.
It’s capturing the value they’re already creating but losing to administrative friction.
@karpathy: “I had this wrong for decades. Agency > Intelligence.”
This changes everything.
High agency: “I’ll figure it out” → actually does
Low agency: Waits for luck, permission, perfect conditions.
AI makes intelligence free. But it can’t make you DO anything.
Agency is the new scarce resource.
Are you acting like you have 10X more agency than you think you do?
That’s the question.
@karpathy: “I had this wrong for decades. Agency > Intelligence.”
This changes everything.
High agency: “I’ll figure it out” → actually does
Low agency: Waits for luck, permission, perfect conditions.
AI makes intelligence free. But it can’t make you DO anything.
Agency is the new scarce resource.
Are you acting like you have 10X more agency than you think you do?
That’s the question.
@karpathy: “I had this wrong for decades. Agency > Intelligence.”
This changes everything.
High agency: “I’ll figure it out” → actually does
Low agency: Waits for luck, permission, perfect conditions.
AI makes intelligence free. But it can’t make you DO anything.
Agency is the new scarce resource.
Are you acting like you have 10X more agency than you think you do?
That’s the question.
@balajis Yeah, moving in that direction. The scarcest resource I see isn’t compute, tech or capital, it’s organizational trust, and people.
Until technologists learn to build inside existing systems, capital still decides velocity.