Future-Proof: How Regulators Can (And Must) Avoid Overfitting to Today's New Tech
Audio Brief
Show transcript
This episode explores the delicate balance between technological innovation and government regulation, drawing on insights from Christian Keil, Vice President at Astranis.
There are three key takeaways from this discussion on how regulatory frameworks can either stifle or accelerate progress.
First, regulators often fall into the trap of overfitting rules to current technologies. Much like in data science, where a model might work perfectly on past data but fail on new inputs, laws written specifically for today's tools inadvertently ban future solutions. Keil argues that mandating specific technologies, such as fiber-optic cables for broadband, blocks superior future inventions like advanced satellite internet simply because they haven't been invented yet. This creates a rigid system that assumes today's technology is the final state of innovation.
Second, regulatory strategies generally fall into four distinct categories ranging from restrictive to enabling. The most harmful approaches include the total prohibition of industries or the protection of incumbents through excessive paperwork, which acts as a moat against startups. Keil notes that Americans spent 10.5 billion hours on regulatory compliance in 2023, a figure that exceeds the total hours required for the entire Apollo moon landing program. A more subtle danger is the well-intentioned strategy of enabling the current thing, where governments subsidize today's popular tech at the expense of alternative, potentially better solutions.
Third, the ideal path forward is a tech-neutral, outcome-based approach. Rather than prescribing specific methods, legislation should focus on the desired result, such as connecting people to low-latency internet regardless of the delivery mechanism. This strategy acknowledges the unknowability of the future and prevents laws from becoming obsolete. For entrepreneurs and policymakers, the goal must be to design frameworks that encourage competition and leave room for unimagined innovations to solve societal problems.
In summary, shifting toward outcome-based legislation ensures that legal frameworks remain relevant and supportive of future breakthroughs rather than locking industries into the past.
Episode Overview
- This presentation by Christian Keil, Vice President at Astranis, explores the complex relationship between technological innovation and government regulation, specifically within the context of the space and telecommunications industries.
- Keil categorizes regulatory approaches into four distinct strategies, ranging from complete prohibition to a "tech-neutral" model that encourages future innovation rather than locking in current technologies.
- The talk is relevant for entrepreneurs, policymakers, and technologists interested in how legal frameworks can either stifle or accelerate progress, using real-world examples like rural broadband initiatives and satellite internet deployment to illustrate the consequences of different regulatory mindsets.
Key Concepts
- The "Overfitting" Problem: Just as overfitting in data science leads to models that work perfectly on past data but fail on new data, regulators often "overfit" rules to current technologies. This happens when laws are written so specifically for today's tools (e.g., mandating electric stoves or specifying fiber-optic cables) that they inadvertently ban superior future technologies that haven't been invented yet.
- The Four Regulatory Strategies:
- Allow Nothing: The most restrictive approach, often used intentionally to kill an industry (e.g., nuclear power regulation post-Three Mile Island).
- Protect Incumbents: A strategy that favors established players by creating high barriers to entry through excessive paperwork ("The TSA Method") or granting special legal privileges to current operators.
- Enable the Current Thing: A well-intentioned but flawed approach where regulators try to accelerate a specific, currently popular technology. While this boosts the favored tech, it blocks alternative solutions that might be better (e.g., mandating specific broadband technologies while excluding satellite).
- Tech-Neutral, Future-Friendly: The ideal strategy where regulations focus on outcomes (e.g., connecting people to the internet) rather than prescribing the specific methods, allowing market competition and new inventions to solve problems.
- The Unknowability of the Future: Because we cannot predict future inventions (as predicting an invention is essentially inventing it), regulations must be designed with humility. If laws are built around the assumption that today's technology is the final state, they will inevitably become obsolete and restrictive.
- Incumbency as a Barrier: Large, established companies (incumbents) often benefit from heavy regulation because they have the resources to navigate complex compliance burdens that startups cannot afford. This creates a "moat" that protects them from disruption, even if their service is inferior.
Quotes
- At 1:30 - "The first thing is that regulators are people... People are diverse, people have lots of different viewpoints. There is no such thing as 'a regulator.' There are just regulators that are different people in different contexts... and they all have different tendencies." - Highlighting that regulation is not a monolithic force but the result of individual human decisions and biases.
- At 5:41 - "Americans spent 10.5 billion hours on 'regulatory burden' in 2023... The entire Apollo program took 7 billion hours to get a man to the moon." - Illustrating the sheer scale of compliance costs in the US and how accumulated bureaucracy acts as a massive tax on human effort and innovation.
- At 9:35 - "If you fit regulations specifically to the current thing, it will help now. But it will only help you in the future if the current thing becomes the forever thing... If the future does not look like the present, then you're going to have the exact same situation... just two or three years from now." - Explaining the core danger of well-intentioned regulations that are too specific to today's technology.
Takeaways
- Shift advocacy toward outcome-based legislation rather than technology-specific mandates; ensure laws define what needs to be achieved (e.g., low-latency internet) rather than how it must be achieved (e.g., via fiber optic cable).
- When analyzing a market or industry for a startup, assess the regulatory landscape to identify if it is "encased in amber" (Strategy 2) or hostile to new entrants (Strategy 1), as this significantly impacts the viability of disruption.
- Adopt a "Tech-Neutral" mindset in organizational policy-making by avoiding rules that lock processes into current tools, thereby leaving room for future innovations that are currently unimaginable.
