Sep 8, 2025, 9:01 AM ETInterview, Fireside Chat

The Little Tech Agenda for AI

The firm explicitly advocates for "little tech"—startups and entrepreneurs with fewer than 50 employees, garage builders, and smaller model developers—rather than legacy institutional players.
The agenda addresses the disparity in regulatory compliance capacity between small builders and trillion-dollar incumbents like Microsoft, OpenAI, Meta, or Google.
The initiative was formalized as the "Little Tech Agenda" in July 2023 to fill the absence of a specific voice for non-fortune 500 tech entities in Washington D.C. and state capitals.
The firm argues that current advocacy often inadvertently aligns with large incumbents, failing to protect smaller entities that cannot absorb the costs of complex compliance frameworks.
Colin Matt joined the firm in November 2023 after reading the agenda, which highlighted an "empty seat" for small builders in policy discussions.

The firm's primary policy pillar is "regulate use, do not regulate development," which distinguishes between governing harmful application and stifling the creation of technology.
The framework advocates for enforcing existing criminal, civil, and consumer protection laws against AI misuse (e.g., fraud, discrimination, cyberattacks) rather than creating new ex-ante development restrictions.
The firm explicitly rejects "zero regulation" as a goal, stating they have never argued against governance in their portfolio; rather, they oppose misguided regulation.
The philosophy argues that regulating development (e.g., licensing models like nuclear energy) creates unnecessary barriers that consolidate power among existing players and hinder competition.
Policymakers often misinterpret "do not regulate development" as "no regulation," omitting the firm's emphasis on active enforcement of existing laws regarding harmful use.

The AI policy conversation accelerated rapidly following Fall 2023 Senate hearings where CEOs warned of existential risks, spooking Capitol Hill and leading to calls for immediate, restrictive oversight.
The firm attributes the intensity of the "safetyism" movement to a 10-year head start by specific advocacy groups and non-profits in influencing think tanks and legislators, which the firm claims dwarfs industry lobbying spend.
Early policy proposals included a "licensing regime" for AI similar to nuclear energy and bans on open-source models, which the firm argues would have stifled U.S. innovation and lost the race to China.
The Biden administration's Executive Order on AI is criticized for promoting restrictive measures that were not well-calibrated to the startup ecosystem.
The firm notes that while there is bipartisan consensus on the need for safety, the specific policy tools proposed (licensing, audits) often fail to account for the unique constraints of small builders.

The firm advocates for a federal framework to govern AI development and interstate commerce, arguing that a 50-state patchwork is unworkable for software technology.
States are assigned the role of policing harmful conduct, such as enforcing anti-discrimination laws and criminal statutes within their jurisdictions.
The firm released a recent analysis warning state laws against potential "Dormant Commerce Clause" violations where regulations burden out-of-state commerce more than in-state benefits.
The National AI Action Plan is praised for signaling a shift to a "win while keeping people safe" national security and economic stance, moving away from a sole focus on safety over innovation.
Recent legislative efforts like the AI Moratorium failed to pass due to procedural issues in the reconciliation package and a lack of organization among pro-preemption stakeholders.
The firm anticipates a need for federal preemption specifically regarding model regulation to prevent a patchwork of conflicting state requirements.

The firm is prioritizing the creation of a centralized federal resource to lower barriers to entry for startups, including improved access to compute and data.
Upcoming advocacy focuses on federal standards, workforce retraining programs, and AI literacy initiatives outlined in the National AI Action Plan.
The firm recently established a political action vehicle ("Leading the Future PAC") to fund advocacy at federal and state levels to counter the "doomer" narrative.
The firm hired Kevin McKinley to lead state policy efforts, aiming to transition from reactive opposition to proactive agenda-setting in upcoming legislative sessions.
The firm warns that the industry alignment may fracture in the coming months, with potential divergence between large incumbents and smaller builders on specific regulatory approaches.
The firm remains committed to a nonpartisan stance, aligning with any stakeholder (big or small) that supports "little tech" interests and opposing those that do not.

Colorado's passed AI law, which requires risk assessments for "high-risk" uses, is criticized as an amorphous administrative burden that may drive startups away rather than effectively reducing bias.
The firm suggests a simpler alternative to complex impact assessments: explicitly codifying that using AI to violate existing anti-discrimination statutes is illegal and enforceable.
The firm argues that "ex-ante" regulations intended to predict and prevent future harms are ineffective and threaten civil liberties, citing the unworkability of predictive surveillance.
Export control proposals that inadvertently ban the international distribution of open-source U.S. models are opposed as they would cede global market share to Chinese competitors.
The firm contends that the "industry agreement" cited by lawmakers often reflects the views of large tech companies, not the broader ecosystem of small builders.