What this covers
Justin Singer talks with a host from Founder's Field Guide about how markets form, using cannabis as the main case. He starts with non-cannabis examples of what he calls regulatory unlocking events: the 1994 Dietary Supplement Health and Education Act, Uber's reliance on independent-contractor classification, and the surviving clause of Section 230. Some of these readings are contested, and so is his claim that no market is truly free.
The cannabis detail is operational. Singer describes the US as effectively 50 separate industries and walks through Colorado's three-tier system. He explains why a manufacturing license tied to a wholly owned building, with setback rules and no mortgage, makes compliant production slow. Co-Manufacturing equipment is built for national scale, and Google and Facebook refuse cannabis ads. He also contrasts tested THC with essentially unregulated CBD. The science thread covers bioavailability (CBD tinctures at roughly 6 percent), wide differences in how people absorb cannabinoids, and THC acting differently at different doses. He pushes back on the claim that 113 cannabinoids are all commercially meaningful, and argues for Magnitude Over Direction when discussing risk.
The remaining turns are historical and methodological. Singer ties early marijuana law to racism against Mexican immigrants and says the Shafer Commission's rescheduling advice was shelved under richard nixon. He expects policy to move through Cohort Replacement as older politicians leave. He also turns to investing, recommending the CIA's material on Structured Decision-Making, forecasting ranges rather than single outcomes, and reconstructing the alternatives behind successes. He treats picking founders on intuition as luck.
Singer argues that regulation, not technology, creates markets, and that cannabis's biggest opportunities lie in building quality infrastructure and product science while the federal rules remain unsettled and the big players are locked out.
- Markets are platforms created by law, property rights, and shared rules, so regulatory change is often the true source of new opportunity
- Federal illegality keeps large CPG companies out, letting small operators build the infrastructure and product science that scale later requires
- Current CBD and THC regulation rewards dishonesty, and regulators must act to create the data and standards the industry needs
Cannabis's missing playbook and unfinished regulation create early opportunity for prepared operators.
- Singer became interested in cannabis in late 2013 because Colorado and Washington were about to legalize it, and the absence of an established playbook, structure, or deal terms was what made it attractive.
“There was no vault guide to cannabis. This was all brand new and was changing rapidly. And that was really attractive.”
- When Singer pitched underbanked and cannabis markets to funds around 2013, he was dismissed as a joke, showing that the mainstream investment community did not take the sector seriously before legalization.
“I kept getting laughed out of the room. Nobody was taking me seriously. Like I got a lot of stoner jokes”
- VCs do not need a dedicated analyst for regulatory change because regulatory shifts are hard to predict a priori; instead, investors should pay attention to the whole picture.
“I don't know if you can do it with an analyst like I think it behooves VCs to actually pay attention to the whole picture.”
- Once a market gets a standardized playbook, all of the real opportunity is gone; the early days of venture, when deal terms were unstructured, were the opportunity, and cannabis offered that same condition.
“once there's a vault guide to something, all of the real opportunity is gone. There's been a vault guide for venture for 15 years now.”
- Obtaining a Colorado cannabis manufacturing license requires finding a wholly owned building that meets city setback rules (often up to a thousand feet from schools, halfway houses, or open space) and is in a locality that permits the activity, because the bank mortgage would otherwise veto the lease.
“Co-Man aren't legal allowed to manufacture marijuana in Colorado. You actually have to go and get a license to be a marijuana manufacturer.”
- Almost all of the regulatory unlocking remains ahead for cannabis: the market lacks national-scale supply chains, low-dose products are restricted, and infrastructure for scaled distribution has not been built.
“Almost all of it is left to be unlocked.”
- Operating in a hard cannabis environment builds operational control that will become an advantage as federal legality reduces headwinds, moving the company from advanced difficulty to easy mode.
“if we can learn to function in an environment as fucked up as this, the headwinds are only receding.”
Legal rules, not technological novelty, unlock the largest markets and decide their viability.
- Section 230's liability protection emerged from a lawsuit over Jordan Belfort's penny stock pump-and-dump scheme on Yahoo forums, where Yahoo was sued as a publisher and the ruling said publishers are not strictly liable for platform content.
“people from his firm were pumping up penny stocks on Yahoo forums, doing pump and dump with them. And Yahoo got sued as part of it that they were the publishers”
- Uber's multibillion-dollar business was built on a scale of independent contractor labor made possible by a legal classification question, not by technological innovation in dispatch, and laws like California's AB 5 threaten it because clearly laws matter.
“that whole business when I say that business, multibillion-dollar behemoth of a business that it is, the capital hoard that it is, is built around a use of independent contractors at a scale that has just never been done before”
- The surviving clause of Section 230 of the Communications Decency Act, which shielded platforms from strict publisher liability for user content, unlocked the user-generated content industry including Facebook and Twitter.
“that sort of became the basis for things like UGC, for user-generated content, for Facebook, for Twitter”
- The best venture opportunities occur where massive, known demand overlaps with a regulatory unlocking event, because laws define the platform on which companies compete and create markets.
“the best venture opportunities are where you have the overlap of massive extent demand, known demand and some sort of regulatory unlocking event”
- The passage of the Dietary Supplement Health and Education Act in 1994 caused the supplement industry to explode, which Singer says shows that nearly every large market can be traced to a regulatory unlocking event.
“supplements, which for better or for worse, exploded after the passage of the Dietary Supplement Health Education Act, DSHEA, back in '93”
- The inversion of the capital gains tax rate relative to ordinary income rates unlocked the explosion of the investment funds industry that followed.
“You look at the inversion of the capital gains tax rate versus ordinary income. And you saw the explosion of the funds industry that comes after that.”
- Singer views blockchain as essentially a shitty database, meaning he sees little technological novelty beyond conventional data storage.
“Every time I hear blockchain, I hear shitty database.”
- Singer is bearish on autonomous cars not because the underlying technology is infeasible but because insurance liability and legal questions around them are unanswerable.
“I'm a bear on things like autonomous cars and it's not because I don't believe the technology is possible, it's because I don't think the insurance question is answerable”
- Crypto and blockchain projects are seeking government sanction, such as recognition as a store of value, because they recognize that official approval is required for them to become what they believe they can be.
“They're desperately trying to get the Fed to declare them a store value.”
- Technology by itself never creates a market; a legal enshrinement of how the technology may compete is required to form the platform on which companies can fight and build economic activity.
“Technology alone never creates a market. There has to be some enshrinement of the competition of that technology”
Fragmented state-by-state rules make cannabis operations far slower and harder than normal markets.
- The US cannabis market is effectively 50 separate industries because each state sets its own laws and market structure, and Colorado uses a three-tier system like alcohol with cultivators, processors, and retailers.
“When you look at the THC industry, you're really looking at 50 different industries. Each state has its own set of laws and has its own market structures.”
- Co-manufacturing equipment in the US is built for national scale, so turning on a line for one state's cannabis market would produce about ten times more inventory than could be sold, and small-scale equipment has largely disappeared.
“turning on their lines would produce 10 times more inventory than you could ever sell. So there was this mismatch”
- Cannabis companies cannot use standard co-manufacturers because Colorado requires a dedicated marijuana manufacturing license tied to real estate that meets zoning setbacks and has no bank mortgage, which makes compliant production a long, difficult process.
“Co-Man aren't legal allowed to manufacture marijuana in Colorado. You actually have to go and get a license to be a marijuana manufacturer.”
- Operating in cannabis takes five to ten times longer than in normal markets, often needing three vendors to find one capable one, each taking about two months to prove competence, so the company adopted a rule to distrust what people say.
“everything takes five times longer if not ten times longer. It takes three vendors to find one who can do the job.”
- Understanding a system at one scale does not imply understanding it at another, so cannabis cultivation and operations require new models at each scale, much as friction behaves differently at molecular, inch, and mile levels.
“just because you understand something the way it operates at one scale, doesn't imply you understand how it will operate at another scale.”
- Models built in Colorado may not transfer to other states such as Michigan, so each state may be truly unique and require relearning everything, a prospect that is both scary and exciting.
“Is this as unique of a market as I believe it to be? Or are there lessons in this market that are transferable to other states”
Judgment about founders should rest on structured tests, ranges, and luck rather than gut instinct.
- Early cannabis operators were not more dishonest than today's participants but were risk takers who operated in cash without controls; when markets legalized, investors using leader-based selection wrongly assumed those same people would build responsible businesses.
“The largest players in an illegal market who have been given the color of law are still the same people they were before they were given the color of law.”
- The CIA's internal library on structured decision making for unstructured situations is among the best resources available, and investors should use it to predict ranges of outcomes rather than a single point outcome.
“CIA's internal library on making structured decision making in unstructured situations is as good as it gets.”
- Organizations that treat employees as humans rather than requiring cult-like commitment would not lose innovation with societal benefit; if anything they would see more innovation, though most current innovation serves equity value rather than employee living standards.
“I don't think you would see any reduction in innovation that actually has societal benefits off of that. If anything, you'd see the exact opposite.”
- Most founders are not special visionaries; they are often poor human beings carried by competent people in their organizations, and the idea that founders must almost bullshit themselves to create companies is a form of idolatry.
“Most of the time, they're actually atrocious human beings who are just being carried by competent people within their organization”
- Teams should routinely reconstruct the alternatives considered before a successful decision, since treating an outcome as inevitable because of skill ignores the large role of luck and the paths not taken.
“I also think it's super important to always play the counterfactuals.”
- Roger Ehrenberg paying Singer out of his own pocket to be an intern at IA Ventures, when it was a friends and family firm, was the kindest act Singer has received, as it led to his return to New York and his later family and company.
“Roger Ehrenberg paying me out of his own pockets to be an intern at IA when they were a friends and family company”
- Because the world is complex and luck is massive, the best strategy is to position oneself to have larger and better options over time, so that eventually one can find and lean into a winning option.
“The best thing that you can do is put yourself in a position to have better, larger options than you had before.”
- Intuitive investing in people is unreliable because people are complex open systems; good outcomes from gut judgments are luck, so investors should build decision structures that probe integrity and the gap between describing and doing.
“There's no such thing as intuitive investing when it comes to people. Those are complex open systems. If you tell yourself, I just like the cut of his jib, they turn out to be great that's luck.”
- Choosing founders based on intuition, such as liking their demeanor, is not repeatable skill because human organizations are complex open systems, so good outcomes from such picks are luck.
“There's no such thing as intuitive investing when it comes to people. Those are complex open systems. If you tell yourself, I just like the cut of his jib, they turn out to be great that's luck.”
- Today's cannabis participants are better credentialed but no less full of bluster, and their overconfidence is unconscious ignorance or arrogance rather than deliberate deception.
“Today's participants are better credentialed, but they're also full of shit. It's not even intentional.”
- In cannabis, people's optimism cannot be distinguished from sales pitch or lying, so a blanket rule is to not believe people's claims, since most failed plans never made sense or materialized.
“Don't believe people when they tell you things. It was a blanket rule for me as an investor, but it's especially a blanket rule for me here.”
Laws and rules define the platform on which markets and companies compete.
- There is no such thing as a truly free market, because a free market without rules is simply whoever holds the biggest gun winning, while property rights and rule of law constrain the market but create a shared platform for transacting.
“there's no such thing as a free market. You want a free market, go to Russia. Whoever's the biggest gun wins.”
- Questions about who owned advertising time rights and how spectrum was auctioned determined the quality of content, the axes of competition, and the growth potential of telecom markets.
“These were all big questions and they all determined both the quality of what was going to be able to produce, what the axes of competition were and what the ultimate growth potential of the market was.”
- Markets can be used or misused, and achieving economic activity requires making a conscious decision about how to design and use them, as argued in the book Reinventing the Bazaar.
“They can be used. They can be misused, but you have to make a conscious decision about how you use them if you want to get economic activity as a result.”
- Bankruptcy law provided a platform of certainty about how outcomes would unfold in different scenarios, which allowed the reorganization industry and high-yield debt markets to be built on top of it.
“suddenly you're able to build all sorts of industry, whether it was high yield debt or junk bonds or whatever it might be, but it was built on top of this platform of certainty”
- Organized sport like the Super Bowl required shared rules, regulations, and contracts, because without rules 22 players fighting in a field is not a game; early college football deaths prompted the rule-making that created the modern game.
“early days of football, college football in the late 1800s, people were dying quite constantly. There was no rule book.”
Regulators should act now, because lax rules reward lying and leave consumers exposed.
- The FDA has claimed a three to five year timeline to regulate CBD derivatives of legalized hemp, which Singer calls wild for a public health regulator ignoring a market where 20 to 40 million people daily consume an unregulated product.
“FDA has claimed a three to five-year timeline. I think it's wild for a food public health regulator to just ignore a market where 20 to 40 million people a day are consuming a product totally unregulated”
- There is absolutely no penalty for lying in the current environment, and the market is rewarding bullshitters from top to bottom, so fraud is a rational expectation.
“There is absolutely no penalty for lying or bullshitting at this point. We are rewarding the bullshitters from top to bottom.”
- Regulators should regulate the market where they can and improve rules over time rather than wait for perfect information, because leaving consumers exposed is worse than imperfect regulation.
“You regulate the market where you can, and then you improve regulations over time. You don't look for perfect.”
- In the CBD market the returns on lying are greater than the returns on doing the right thing, which is a regulatory failure that regulators can fix by putting rules in place.
“the returns on lying are far greater right now than the returns on doing the right thing. And that's a regulatory fault.”
- The popular argument that the modern internet would not exist without Section 230 is overstated, but it is still an admission that regulation matters for market formation.
“Well, I don't know that that's quite the argument in favor the people make it out to be, it is an argument, an admission that regulations matter.”
Drug categories should follow safety profiles, and CBD consumers currently lack testing oversight.
- Raw CBD oil tinctures are an intermediate ingredient rather than a consumer product, because no consumer prefers to drop crude oil on the tongue with an eyedropper.
“It's also basically an intermediate good. Nobody runs out and says, you know what my favorite thing in the world is to do, is to drop raw crude oil on my tongue with an eyedropper.”
- The concept of a drug is a mental frame; alcohol, caffeine, prescription opioids and coffee all augment physiology, and the relevant question is safety profile and effect rather than legal category, with THC plausibly safer than caffeine.
“These are all drugs. They all augment physiology. The question is really, what is the safety profile? What is the effect?”
- Claims that there are 113 identified cannabinoids that all have bioactive effects are overbroad; what matters commercially is which compounds can be produced economically, shown safe, and sold for profit, which is currently unknown.
“I think people make overbroad claims when they say it was 113 identified cannabinoids that all have bioactive effect.”
- CBD is essentially unregulated, while THC faces state-level testing requirements such as cannabinoid testing before products reach shelves in states like Colorado, meaning CBD consumers cannot know what they are taking.
“CBD is just unregulated. THC is actually more regulated than CBD right now.”
Cannabis prohibition and federal illegality were unscientific and shaped the industry's structure.
- Federal illegality keeps large CPG brands from entering cannabis because they cannot risk existing business lines, and it also prevents them from using their national-scale infrastructure, which gives small operators room to find the right formulas.
“federal illegality is holding them back. Anybody who has an existing business line to protect, can't afford to take the risk of operating this.”
- Early US marijuana laws such as the Stamp Act and the Marihuana Tax Act were motivated by racism against Mexican immigrants, and cannabis tincture had been one of the most used drugs in the country before prohibition.
“cannabis tincture, which is marijuana tincture, was one of the top three most used drugs in the country.”
- There was no scientific basis for keeping marijuana in Schedule I for so long, since the Shafer Commission recommended rescheduling early on but the Nixon administration shelved the report.
“There's no scientific excuse for marijuana to have been sitting around schedule one for as long as it was, this was known”
Cannabinoid products should be developed as food-science commodities with verified stability and mass appeal.
- The Ripple product emerged when the company recovered the water-soluble powder fines it had been discarding after granulation, since the 50 percent yield loss made a 10-milligram tea unprofitable, turning the waste into the growth engine.
“our yield was turning out to be like 50% of our THC, which is our core input there. ... what was before waste, became our growth engine.”
- Singer expects three to five follow-on cannabinoids to reach market in the next decade, which he says would be a good outcome given the high scientific interest in these bioactive ingredients.
“I think I'd be happy with three to five additional follow-on cannabinoids in the next decade”
- Cannabinoid products should be formatted so they are mass acceptable, do not alter users' identity, and deliver a consistent experience, since the key question for any functional food is whether it does anything, and THC clearly does.
“How can you put it into formats that are mass acceptable, that don't change people's identity and that are consistent enough”
- Singer's grandmother wanted something to feel better, not to get high, and Singer had no suitable product to offer, which revealed a gap between cannabis products and people seeking mild relief.
“she didn't want to get high. She just wanted to feel better, and I couldn't offer her anything that would make her feel better.”
- Qualifying a cannabinoid raw material is essentially qualifying an agricultural commodity for processing, which is why food scientists from large food companies are well suited to the work.
“fundamentally you are qualifying an agricultural commodity trying to convert it into something that is useful for processing.”
- In 2014 the edibles market was in a race to maximize THC per container, and Singer's team chose to take full cannabis risk, building a food company focused on cannabinoids rather than a cannabis company playing with food.
“we saw this race to potency going on with edibles, where everyone was just trying to shove as much THC as possible into the smallest container possible.”
- Anecdotal reports indicated that people got high faster with Ripple's water-soluble format than with other edibles, which the company later tested in studies.
“people are giving us these anecdotal reports, that they were getting high faster with ripple than they would when they were taking other edibles.”
- Cannabinoid shelf stability must be verified, including whether CBD degrades in base environments and needs an acidic environment, and whether it degrades below label claims before scaling production.
“Will a base breakdown CBD? The answer is yes. Do you need a high acid environment for it to actually live in?”
Rules and constraints redirect innovation toward the frontier rather than polishing existing systems.
- Economic innovation is most valuable in creating things from whole cloth, because firms that must spend most effort polishing the last 5 percent of a rule-bound system have little capacity left for new creation.
“You're not just focusing on spending 95% of your effort to get the last 5% of the way there, you're spending all your time getting that first 95% all over again.”
- Periodically changing rules in F1 racing prevents teams that exploit the current rules from pulling away, restores a level playing field, and shifts competition toward innovation in the first 95 percent of the work.
“if you change the rules and vary them up a bit, then suddenly everybody is looking at this new jumble and they're kind of on a level playing field all over again”
Cannabis effects depend on dose and absorption, so dose control is essential.
- Edibles are the first opportunity to control and reduce the dose variable in cannabis consumption, making measured outcomes observable and useful, while inhaled joints remain unpredictable.
“Edibles are actually sort of our first opportunity to control that input and reduce the variable there”
- THC acts like three different drugs depending on dose: under about 5 milligrams it is a mild anxiolytic for most people, 5 to 15 milligrams produces noticeable intoxication, and above about 50 milligrams it becomes nearly psychedelic.
“THC is kind of three different drugs at one, a sub-five-milligram dose for most people, it's like a light anxiolytic”
- It is almost criminal to discuss the direction of a drug's effect without discussing its magnitude, and the magnitude of THC risks has been wildly overblown relative to their actual size.
“I think it's almost criminal to ever talk about direction of effect without talking about magnitude of effect”
- Absorption of THC and CBD varies widely from person to person, and some people cannot get high at all, some do not absorb CBD, and some can only get high through edibles because they absorb through different pathways.
“Person to person variance in terms of who absorbs what and how much, even in water-soluble versus water-insoluble formats, it's crazy. Some people just can't get high.”
- Budtenders drive about 70 percent of new product sales in dispensaries because customers come in seeking an experience rather than a specific product, so dispensaries function more like jewelry stores than pharmacies.
“budtenders drive 70% of the sales of new product sales because people come in and they know they want an experience, they don't know which product to get”
- Singer discards nearly all research done before 2016 and research on inhalable marijuana because joints vary in dose and potency, making inhaled consumption too uncontrolled to yield useful data.
“I basically throw out almost all research done before 2016. I throw out all research done around inhalable marijuana because we just don't know.”
- The best research suggests CBD tinctures, the most popular CBD format, are only about 6 percent bioavailable, meaning about 94 percent is excreted without entering the bloodstream, which explains poor consumer experience.
“the best research we have suggests those things are 6% bioavailable, 94% of what you consume is getting excreted out without entering your bloodstream.”
Small, lean operators win by building consistent products slowly rather than scaling fast.
- Google and Facebook refuse ads from cannabis companies, so brands must buy affiliate placements and sponsored articles, adding transaction costs, yet the company still reaches about 35 percent EBITDA margins with a salesforce of three.
“Google and Facebook won't take money from CBD or THC companies. ... yet still 35% EBITDA.”
- As a products manufacturer rather than a cultivator or retailer, Singer's company earns roughly 30 to 40 percent EBITDA margins, with the company's self-distribution and small salesforce helping keep costs low.
“You're talking 30 to 40% EBITDA. And the key thing there is we don't have a huge salesforce.”
- Constellation delayed a launch over aluminum can liners eating THC, but that was a roughly $2 million problem treated as a $2 billion one, since can liners were already a solved issue for a small flexible line with good food scientists.
“that's because you treated a $2 million problem like a $2 billion problem. That was a solved problem already, the can liners.”
- Large consumer brands are built for massive volumes and would struggle to produce small batches to find the right formula, while small operators can start small and work up, an advantage in cannabis.
“federal illegality is holding them back. Anybody who has an existing business line to protect, can't afford to take the risk of operating this.”
- A brand without consistent product is just a logo, and building consistency such as water-soluble cannabinoids took years of R&D, so going too fast with too much money leads companies in the wrong direction.
“A brand without consistency is just a logo. You actually have to produce a consistent product, and that is actually really hard”
- The cannabis industry is in an infrastructure phase where brands can only win after shared commodity infrastructure is available to everyone, so differentiation today comes from quality and capabilities built in-house.
“We're in an infrastructure phase in this industry, you've got to be in it for that. Brands win when you have commodity infrastructure that everybody has access to”
- The original 2.5 milligram Stillwater tea never took off because dispensary customers were mostly heavy users and dispensaries had little incentive to reach new customers like grandmothers, who were the target market.
“the people walking into dispensaries weren't that interested in the low dose products. ... Everybody loves making money when they don't have to do any work.”
Cannabis reform stays stalled until old political leaders leave and a supportive administration arrives.
- Cannabis policy stays frozen while a generation of political leaders still treats it as a third rail, even though public approval is about 67 percent and 51 percent among Republicans; change arrives only when those holding the old views are gone.
“a whole generation of political leaders who still believe that this is a political third rail, despite the fact that it's got 67% approval among the population at large and even 51% approval among Republicans”
- A Nixon aide reportedly admitted that the administration could not criminalize being black or anti-war, but could criminalize marijuana use as a proxy linking those groups, which Singer uses to argue prohibition was politically targeted.
“there was an interview with one of Nixon's aides where he straight up admitted, we couldn't criminalize being black.”
- Cannabis made significant progress despite the most anti-marijuana attorneys general, Jeff Sessions and Bill Barr, which implies that reform could accelerate once a more supportive leader takes office.
“if THC and CBD have managed to make this much progress throughout the administrations of attorney general sessions and Bill Barr, the two most anti-marijuana people”
- Singer expects meaningful THC reform, such as clearing banking issues and moving from decriminalization toward legalization, within the next year or 24 months, depending on the outcome of the election and a more pro-regulation administration.
“You can get THC, probably starting with decriminalization, but heading towards legalization with the next 24 months.”
- A future president might embrace marijuana reform the way President Obama came to support gay marriage, being circumspect during the campaign and then delivering reform after taking office.
“I kind of hope that the next president, whoever he might be would embrace marijuana in a similar way to how President Obama embraced gay marriage.”