Matt Perault

1/ Are the right AI rules... zero rules? Not for Little Tech. To ensure AI thrives in the long run—and that startups can compete at the frontier—we need smart regulation. Here’s what we believe at @a16z 🧵

Matt Cronin outlines three priorities the U.S. needs to pursue become the default AI ecosystem: (1) Keep an unassailable lead at the frontier (2) Win on adoption, with open source playing a key role (3) Make sure AI drives broad-based prosperity

Check out the full conversation here: a16zpolicy.substack.com/p/the-real-ai-…

If you want a real explanation of what’s actually driving the AI race with China, Matt Cronin outlines the CCP’s motivations and the different strategies China and the West are taking to become the world’s default AI system. Great conversation with @Jai_Ramaswamy on the a16z AI Policy Brief!

Thank you, @Gregory_C_Allen, for hosting @jai_ramaswamy and me on the @CSIS_AI Policy Podcast. Appreciate the conversation on what an AI regulatory agenda can look like that protects people, while still supporting innovation and competition. An added bonus was getting to see @CSIS's awesome studio setup. Check it out below.

Read it on the a16z AI Policy Brief: a16zpolicy.substack.com/p/preserving-t…

The Web works because it struck an equilibrium between openness for readers and control for publishers. That balance protects the freedom to learn: if information is lawfully accessible, people are free to read it, analyze it, and build on it. AI now puts that equilibrium to the test. Some proposals would fence off publicly available information through contracts, technical barriers, or expansive interpretations of copyright law. In a new piece, we argue that we don’t need to abandon information openness or publisher control to preserve the freedom to learn in the AI age. 🔗 below!

Ben Horowitz on where AI regulation should actually focus: "The model is a model. It's a mathematical model. It predicts things. It's not a sentient being." "If you ban the technology, or infringe people's ability to do mathematics... that has hundred-year implications." "Don't regulate math. Regulate the applications of that math." "If somebody uses AI to break into a bank or make a robot that shoots somebody, that's illegal. But the technology itself shouldn't be illegal." @bhorowitz with @DavidSolomon and @dhaber

Full conversation here: a16zpolicy.substack.com/p/to-regulate-…

Always enlightening to sit down with @Martin_Casado and @Jai_Ramaswamy, both of whom have experienced major platform and policy shifts up close, from encryption and cybersecurity to cloud and now AI. A core takeaway from this discussion: if you want to protect people from harmful uses of AI, the most effective approach is to target those harms directly. AI shouldn’t excuse wrongdoing. We can build on decades of tech policy and existing law to hold bad actors accountable. As Martin explains, any new AI laws should be grounded in evidence of genuine marginal risk to protect consumers and preserve innovation. Listen here. 👇

Building an AI startup is hard. Small teams must compete with well-resourced incumbents, fighting for access to talent, data, and compute. But in California, there’s another barrier that increasingly makes it hard for startups to compete: a dense and fast-growing web of regulations. What does a small team of founders need to do to comply with California’s AI regulations? We mapped the compliance obligations AI startups face in California in the wake of new laws passed in California in the last few years. A startup’s success shouldn’t be determined by its legal budget. If startups need to write big checks to lawyers just to get an AI product into the market, we’ll end up with fewer challengers, worse products, and a less dynamic AI ecosystem. Greetings from the Garage: a16zpolicy.substack.com/p/greetings-fr…

New York’s RAISE Act is now law, and it looks meaningfully different than it did when first introduced. We applaud Governor Hochul for making amendments to the final bill that recognize the challenges startups face in competing against well-resourced incumbents, including ensuring that burdensome disclosure requirements do not fall on companies unless they exceed a revenue threshold. The intended goal of RAISE is important: keep users safe. But it chooses a misguided approach to try to achieve this goal, targeting the development of AI technology rather than punishing its harmful use, and it adds to a growing, state-by-state patchwork of AI laws that is difficult for Little Tech to navigate. The result will be a legal regime that fails to protect people and undermines competition and innovation. States have an important role to play in AI policy. But governing a national AI market that cuts across state lines requires federal leadership. It is time for Congress to act. a16z has shared a roadmap with nine principles for Congress, outlining the elements of a policy agenda to protect people, foster competition, and safeguard innovation: a16z.com/a-roadmap-for-…

The ongoing evolution of privacy law in the U.S. offers a preview of what a fragmented AI policy landscape could look like: a complex and costly system of overlapping and ever-shifting state requirements. AI is far harder and costlier to govern through a patchwork, with much higher stakes for U.S. innovation and competitiveness. A new piece with Kevin McKinley for the a16z AI Policy Brief. Full article below ⬇️

If there is to be federal preemption of state AI laws, it makes sense to begin with a federal AI law. Here is a helpful discussion (by @MattPerault et al) of a distinct U.S. approach that is a good starting point for discussion of what might make for constructive federal law.

If we want to protect consumers and keep AI development in the U.S., Congress must set a national standard for the technology without delay. In an op-ed for the Washington Post, I argue that the most important action policymakers can take is passing a strong, bipartisan federal standard that positions America to lead, empowers startups to compete, and creates a safe, trusted AI market for all Americans. The world is safer and more prosperous when America leads. Our startup ecosystem is a national asset and the engine of U.S. innovation. But startups simply cannot survive a 50-state patchwork of AI laws. For young companies, navigating conflicting rules with two- or three-person teams is a near–death sentence. Without a national standard, AI development will consolidate in a few large companies, leaving the U.S. less competitive, consumers with worse products and fewer options, and opening the door for China’s rapidly advancing AI systems to shape global norms. Congress must seize this moment. Read more: washingtonpost.com/opinions/2025/…

The stakes for AI governance could not be higher. We need clarity, coherence, and a national framework that protects people while enabling American innovation. Congress has a critical opportunity to act. The roadmap we’ve released today lays out practical, durable principles for governing AI in a way that keeps Americans safe and ensures the US leads the world in this technology. I look forward to continued work with lawmakers on a balanced, bipartisan path forward.

Today we released ideas for federal AI legislation. We’re entering a consequential phase for AI policy. The path forward is clear: protect people, support builders, and keep the U.S. competitive. It’s time for Congress to act.

Check out the full piece here: a16zpolicy.substack.com/p/a-roadmap-fo…

There is real momentum for federal AI legislation, with both Republicans and Democrats calling on Congress to act. A federal AI framework can do two things at once: protect people from harms and ensure startups can compete. Our new legislative roadmap translates our AI policy work into a concrete agenda designed to keep Americans safe while keeping the U.S. in the lead. The goal is straightforward: focus governance on harmful conduct, lower barriers for Little Tech, and strengthen U.S. competitiveness. I’m looking forward to continued conversations with policymakers on both sides of the aisle as we work toward smart, durable legislation that works for Little Tech.

@AdamThierer @RSI @CSIS Check out the full conversation with Adam and Blair: a16zpolicy.substack.com/p/beyond-preem…

If you squint at today’s AI policy debates, you may just see the 1996 Telecom Act in the distance. I had a great time sitting down with @AdamThierer @RSI and Blair Levin @CSIS to hear their reflections on what that landmark deal in tech policy teaches us about building a national framework for AI today.