Blake Scholl on Supersonic Flight and Fixing Broken Infrastructure
Recorded live at the Roots of Progress Institute’s Progress Conference, Tyler Cowen presses Blake Scholl — founder and CEO of Boom Supersonic — on why so much everyday infrastructure stays needlessly broken, why civilian supersonic flight died after Concorde, and how his own company is trying to revive it against defence-manufacturing habits he considers close to insane.
Key ideas
- Everyday friction is a choice, not a fact of nature. Scholl treats airport terminals, road traffic, security screening, and baggage handling as radically improvable systems that stay broken because of misaligned incentives — airports capped in what they can charge per passenger, roads given away free and then blamed for ‘induced demand’ — rather than because a good fix does not exist.
- Concorde and Apollo were tech demos, not products. Both were government-specified prestige projects built with no serious attention to unit economics; Concorde flew 52 per cent full across 27 years at a $20,000 inflation-adjusted fare, and only 14 aircraft ever entered service. Scholl argues supersonic should have started as a private jet for a small market, the way most successful technology iterates — a path the 1973 US ban on supersonic flight over land also foreclosed.
- Regulation tied to safety is a political one-way door. Scholl’s explanation for why obviously pointless rules persist: the career cost of removing a safety-branded rule and being blamed if something later goes wrong is asymmetric to the invisible cost of leaving it in place — so almost nobody ever removes one. This extends, into public policy, the reversibility logic Jeff Bezos applies to corporate decisions in Two-Way Door Decisions.
- Defence supply chains are optimised for votes, not speed. A jet-engine turbine blade that needs 24 hours of machine time took six months and roughly $1 million to obtain, because each production step sits in a different state — a structure Scholl says exists to place one process step in as many congressional districts as possible. Boom’s answer is full vertical integration: one Denver factory, raw materials in, finished engines out.
- LLMs are collapsing the cost of regulatory change. Aviation certification paperwork that once took an engineer two months to draft — a 100-page test plan citing every relevant federal regulation — can now be drafted in minutes by an LLM working from a RAG system over the regulatory text, then edited by an engineer to fix hallucinations, turning teams away from change-aversion because iteration is no longer expensive.
Content
Fixing everyday infrastructure
Cowen opens by handing Scholl dictatorial power over airports. Scholl’s design: put terminals underground, keep the airside above ground, and route arrivals and departures across two runways with no tugs needed — a model borrowed from a crossbar switch. The obstacle, he argues, is not engineering but a broken business model: airports are limited by regulation to charging a fixed amount per passenger ($5.60), so they are forced to make money by trapping travellers in retail instead. Fixing this requires privatising airport infrastructure and inventing a new revenue model, which he thinks VTOL aircraft could help unlock by making new sites near population centres viable.
The same diagnosis extends to security screening, which Scholl calls ‘a farce’ — illustrated with a personal anecdote of an unnoticed box cutter carried, undetected, through San Francisco, Seattle, and London Heathrow. He credits only two post-9/11 changes with genuinely improving safety: reinforced cockpit doors, and passengers now knowing to fight back rather than comply. Everything else, in his account, survives because safety-tied regulation is a one-way door: reversing a rule exposes whoever does it to blame if anything later goes wrong, while leaving it in place carries no comparable personal risk for the official responsible. International security rules compound the problem by harmonising toward whichever country’s rules are strictest, to preserve connecting passengers without re-screening.
Traffic gets the same treatment. Scholl wants every road tolled and priced like any other scarce good, dismissing the traffic engineers’ concept of ‘induced demand’ — more road capacity just draws more drivers — as an argument for pricing roads, not against building them; the analogy he reaches for is that free, unpriced grocery stores would produce identical shortages. Cowen pushes on enforcement — untraceable or unregistered vehicles, distrust of local tolling authorities — and Scholl treats this as a solvable detail (licence-plate cameras, mailed tickets), not a reason to abandon the idea. Baggage and boarding get a lighter pass: Scholl wants checked baggage fast and reliable enough that nobody carries on, with bags ‘teleporting’ between the trunk of an arriving and departing Uber — the actual constraint, again, is airport and airline process, not technology.
Amazon versus Groupon: long-term thinking as a cultural asset
Scholl worked at both companies and treats the contrast as a case study in decision-making horizon. At Amazon, when record labels tried to dump discounted CD inventory at quarter-end, the company ran discounted-cash-flow analysis, took the resulting short-term hit to one quarter’s numbers, and captured the better long-run outcome — a pattern he says was baked deeply into the culture through a single ‘trump card’ used to settle any internal debate: what is actually best for the customer. At Groupon, running the email business, he describes being told every quarter to ‘send more email’ to fix a soft quarter, which degraded the following quarter by exhausting the same customer list — short-term pressure directly undermining the long-term relationship the business depended on.
Asked whether online reviews can still be salvaged, Scholl proposes replacing the compressed five-star scale (where ‘everything is four and a half stars’) with percentile ranks, but concedes he has no fix for the deeper problem: a strong financial incentive to stuff the review ballot box in the first place.
The rise, fall, and return of supersonic flight
Scholl’s central historical claim is that Concorde and the Apollo programme were both government-specified prestige projects — ‘tech demos, not products’ — built with little attention to whether the underlying economics worked, so both proved a capability existed and then produced no successor for half a century. Concorde, a joint venture between the French and British governments, flew at an inflation-adjusted fare of roughly $20,000, averaged 52 per cent full across 27 years even on its most popular route (Newark–London), and only 14 aircraft ever entered service. Cowen presses back once, noting Apollo’s satellite-infrastructure spillovers as a possible justification for state-led moonshots; Scholl does not fully resolve the counterfactual, but maintains that entrepreneurially-driven aviation innovation — every commercial generation from the Wright Brothers to the 707 faster and safer than the last — worked precisely because it was not run that way.
His preferred counterfactual path for supersonic: start with a private jet for a small, wealthy market, the normal way a new technology iterates into an S-curve. The US ban on supersonic flight over land, in place from 1973, foreclosed that path too, since 85 per cent of business-jet miles are flown over land — ‘effectively, we banned minimum viable product’. Boom instead leapt straight from nothing to an international-route airliner. As of the October 2025 recording, Scholl frames the preceding six to nine months as an inflection point: Boom broke the sound barrier twice (January and February 2025), announced Boomless Cruise — technology that avoids an audible sonic boom at ground level — and the land-over-flight ban was repealed by executive order on 6 June 2025, ending a 52-year prohibition. A minor literary aside: Ayn Rand’s Atlas Shrugged (1957) mentions supersonic transport once, in Galt’s speech, which Scholl reads as arguing that only a profit-motivated capitalist, not a state actor, could actually build one — see Objectivism.
Manufacturing, defence procurement, and the cost of change
Asked about the state of American manufacturing, Scholl offers a single concrete anecdote as a diagnostic case. A 3D-printed jet-engine turbine blade was quoted at roughly $1 million and six months’ lead time; the underlying machine work takes 24 hours, and the $2 million printer itself sits idle for two weeks between jobs. The six months disappears into shipping the part between separate factories in different states, each performing one process step. Scholl’s stated heuristic for moments like this — ‘when something seems stupid, ask yourself what would have to be the case for this action to be smart’ — supplies the answer: this is a congressionally-optimised supply chain, deliberately spread across many states and districts to maximise the number of legislators with a local stake in a defence programme’s continued funding, at direct cost to speed and, he argues, wartime resilience.
Boom’s countermeasure is full vertical integration — a single Denver factory taking in raw materials and shipping out finished jet engines, eliminating inter-factory shipping. The second-order effect, in his account, is on engineering behaviour itself: an engineer who knows a wrong turbine-blade design can be remade in 24 hours, rather than six months, takes more design risk and iterates faster, instead of over-analysing to avoid an expensive mistake.
The same logic extends to regulatory paperwork through LLMs. Aviation certification historically required an engineer to spend roughly two months producing a 100-page test plan citing every relevant federal regulation — after which nobody wants to revisit or change anything, since a change means writing the document again. Scholl says a prompt against a RAG system built on the regulatory text can now produce a draft in minutes, for a creative engineer to edit and correct; this, combined with vertical integration and embedded software-hardware teams, is what he means by reducing ‘the cost of change’ across the company. On learning new domains generally, Scholl — a private pilot but not an aerospace engineer before founding Boom — describes keeping a running ‘confusion list’ during Boom’s first year, aiming to resolve roughly one item a week, and made ‘teach me something’ his standard early hiring question.
See also
- Blake Scholl — speaker
- Tyler Cowen — host
- Two-Way Door Decisions — extended here into regulatory and political reversibility
- Objectivism — the Ayn Rand reference on supersonic transport