Implementation Is the Task of Our Time (it's the deployment, dummies)

A cyber x AI dinner, anti-anti-woke, lots of links

Dec 12, 2025

I had wrist surgery on Tuesday and can’t type. Since then, I’m really enjoyingwispr for voice to text.

It’s the deployment, dummies

All model development could stop today and we’d still have decades of implementation work to realize the promise of AI. The bottleneck isn’t scaling raw capabilities and the task of our time isn’t developing the machine god. It’s successfully deploying AI throughout the economy to accelerate GDP and raise living standards worldwide.

We’ve done this before, and it took forever

Computerization took roughly two decades. The shift to the cloud is still ongoing after 15-20 years. There’s no reason to think AI will be faster. Most of the economy isn’t using AI at all. A large portion has a top-down mandate to “do AI” but is failing to deploy anything meaningful or quickly. Only a tiny sliver of businesses and developers are credibly AI-native.

Why implementation is failing

Companies that want to move fast can’t. It takes vision and risk. And even when the will exists, organizational inertia, internal politics, patchwork infrastructure, etc. get in the way. It’s not just “change management is hard” (it obviously is).

Most companies don’t have data infrastructure anywhere close to what AI requires. Meanwhile, the complexity of enterprise environments is rising faster than the capabilities and visibility of those tasked with transforming them.

Where I’m looking

Collaboration: Git was built for scarce, human-written code. Now enterprises are generating more AI code and shipping less of it. Whatever replaces Git will be built for that reality.
Services: SIs/consultants are a critical chokepoint ill-suited to the task. There’s a huge opportunity to build new service providers on back of new capabilities and power AI transformation, better, faster, cheaper.
Growth by buyouts: The product company acquires and transforms the operating business directly while owning all the upside.
Apps: When you’re selling work you can fit within existing workflows like a person, just better, faster, cheaper.

Over the next decade there’s a foundation model-sized opportunity in deploying AI into critical systems throughout our economy and around the world. Developing the technology is not enough.

Cyber Dinner

AI is breaking $1.5T of security market cap.

For 30 years, cybersecurity has worked because attacks have had known limitations. Businesses have also had a concept of control: how to secure within a perimeter, how to vet vendors, and how to remediate gaps or accept known risks. The $1.5T security market cap was built on the assumption of visibility and control. That known landscape is gone with AI.

This winter I’m teaming up with ex/ante and Outtake to host a dinner and discussion in NY about the changing threat landscape and how new and existing security companies will meet that.

Apply to join here.

Virtue signaling

Elite consensus is starting to turn on the “anti woke” aesthetic and vice signaling. Sports betting is the canary in the coal mine for the great re-wokening that is coming. Get ready to see CEOs libbing out again within the decade.

New sincerity comes next. Everyone will talk about kindness, craft, moral virtue as the ways to redeem your soul and the character of the nation. Decency will go from being not just not cringe to actively in vogue. Eventually that will metastasize into facsimile and performance and we’ll reach the 2020-style ethics again.

Gen Z pining for peak millennial aesthetics and culture is more evidence of this larger trend. Prepare for libbed out “awesome sauce” imminently.

I Read

The peasant logic of MAGA politics. On an abundance vs. scarcity mindset and the politics/ sociology of zero-sum thinking
Tech bros head to etiquette camp as Silicon Valley levels up its style. More fun coverage of our etiquette school this time in the Washington Post. This is definitely a thing! I expect we’ll be taking etiquette school on the road next year. Hitting more cities, more founders, more venues.
Shock jockey capitalism. Good long peace from Will Manidis on the increasing tendency to compete over “who can say the biggest number” in venture/startups.
Syria after Assad: A year in numbers. Things in Syria are still really bleak but this is a bit of a white pill with some hope on the horizon starting to come through in the numbers.
Power Law And The Long Tail - AVC. I was reminded of this (old) piece from Fred Wilson over the last week because of a reference from Euclid Ventures. Even if all of the money gets made on one end of the distribution (the biggest winners), all the time gets put in on the other (with the relative losers). It might seem non-economic to do that but in a business where your reputation and your network are everything, devoting that time does make sense and the emotional investment is so high.