one of my favorite rabbit holes lately is understanding company-building over time. often, when a company solves a hard problem, whether it be beating competition, entering a new market, or figuring out the right product, the method they use becomes a reusable strategy. that ‘sauce’ gets adopted by peers, sometimes crosses industry lines, and becomes table stakes for a certain practice or process.

although it is a silo sometimes, the same can apply to crypto. there are things that we’ve built that create efficiencies - ones that could be used elsewhere. finding a method that changes things usually comes from taking a big risk and defying norms. so when we look at crypto through that lens, the examples worth paying attention to are the ones that did something net new, broke convention, and quietly became a better way to do things.

some points in history that paint the picture:

• singer’s territorial “agency” model for sewing machines (1850s) → one of the earliest examples of scaling a product business by licensing a full operating system: exclusive territories, standardized training, payment plans, and local service. this model was later adopted and perfected by mcdonalds & kfc as the format for franchising in the 1950s → codified operations manuals, uniform branding, and centralized supply - all of which are now table stakes for food, retail, fitness, and service chains worldwide.

• ford’s moving assembly line (1910s) → cut assembly time and cost down dramatically. this was copied and extended by GM in the 1920s with financing + model differentiation → mass production became table stakes not just in cars, but also in healthcare (production-line surgery teams), aerospace (the production of b-24 bombers), and later, consumer electronics.

• sea-land freight containerization (1960s) → spearheaded standardized box sizes, which meant shippers pack once at the factory and then move goods across trucks, trains, and ships without unpacking. this inspired an analogous concept in computing, first seen in mainframe and unix environments in the 1970s, later evolving into the “pack once, run anywhere” model that powers docker/kubernetes, and is now table stakes for cloud deployment.

now shifting focus back to crypto, what are/will we export to other industries (with efficiency at the heart):

• [user behavior] always-on markets → crypto’s 24/7 nature created a new class of traders who are always on, globally distributed, and crave access. they are turning to traditional markets, pushing exchanges to experiment with at least 24/5 equities trading (like forex). though there will be some growing pains around lower liquidity on off-hours and systems that might not be ready to internalize the constant demand, we are seeing uniform push towards always-on markets. the point here is that crypto defined the new age trader, which has pressured traditional markets to keep up.

• [product strategy] composability at the front → modular production and composability have been around forever. modern day api products like stripe and shopify are embedded across industries, but in crypto, composability lies the foundation for the main product, not just as an integration. defi protocols are built on public primitives (amms, oracles, vaults, governance modules) that are meant to be forked and they’re doing it live, while handling hundreds of millions in value, in open markets, without central control. how defi primitives have been used gives us a glimpse of what future software development looks like. you see some of this within ai agent platforms. tools like langchain let you chain together small, reusable components (like calling a specific llm) into workflows that other people can fork and extend.

• [infrastructure] distributed computing → traditional distributed compute relied on trusted parties and closed networks. with the introduction of bitcoin and since, crypto brought in open participation, no central fallback, and real monetary value on distributed systems - pushing forward consensus and verifiability in ways traditional infrastructure hasn’t needed to.

  • ai models today are largely closed systems, from the fitted model and its parameters to how inference is served and abstracted from the user. but when models move into sensitive domains like healthcare and credit, transparency should matter. because deepseek’s open source release showed how much demand there is for visibility. decentralized training will be one of the ways that visibility improves; making it work relies on the same crypto-native rails: incentive alignment, distributed governance, and coordination across untrusted parties. model development in the future will borrow from crypto the systems that were built to operate in adversarial, open, high-value environments.

compared to the traditional examples, these focus on a facet of crypto vs. zooming into single companies. we haven’t seen juggernauts emerge from crypto in the same way we did with companies like apple or google or mcdonalds or ford - but that might be the wrong benchmark. crypto runs lean by design: smaller teams, more composability, less overhead. it also tends to serve narrow but high-value markets.

progress over the past ~16 years has produced a set of primitives that define how crypto works and they’re already starting to leak out. crypto was introduced to enable p2p payments, but it’s now informing how software is built, how markets behave, and how distributed systems hold up under pressure.

one of my favorite rabbit holes lately is understanding company-building over time. often, when a company solves a hard problem, whether it be beating competition, entering a new market, or figuring out the right product, the method they use becomes a reusable strategy. that ‘sauce’ gets adopted by peers, sometimes crosses industry lines, and becomes table stakes for a certain practice or process.

although it is a silo sometimes, the same can apply to crypto. there are things that we’ve built that create efficiencies - ones that could be used elsewhere. finding a method that changes things usually comes from taking a big risk and defying norms. so when we look at crypto through that lens, the examples worth paying attention to are the ones that did something net new, broke convention, and quietly became a better way to do things.

some points in history that paint the picture:

• singer’s territorial “agency” model for sewing machines (1850s) → one of the earliest examples of scaling a product business by licensing a full operating system: exclusive territories, standardized training, payment plans, and local service. this model was later adopted and perfected by mcdonalds & kfc as the format for franchising in the 1950s → codified operations manuals, uniform branding, and centralized supply - all of which are now table stakes for food, retail, fitness, and service chains worldwide.

• ford’s moving assembly line (1910s) → cut assembly time and cost down dramatically. this was copied and extended by GM in the 1920s with financing + model differentiation → mass production became table stakes not just in cars, but also in healthcare (production-line surgery teams), aerospace (the production of b-24 bombers), and later, consumer electronics.

• sea-land freight containerization (1960s) → spearheaded standardized box sizes, which meant shippers pack once at the factory and then move goods across trucks, trains, and ships without unpacking. this inspired an analogous concept in computing, first seen in mainframe and unix environments in the 1970s, later evolving into the “pack once, run anywhere” model that powers docker/kubernetes, and is now table stakes for cloud deployment.

now shifting focus back to crypto, what are/will we export to other industries (with efficiency at the heart):

• [user behavior] always-on markets → crypto’s 24/7 nature created a new class of traders who are always on, globally distributed, and crave access. they are turning to traditional markets, pushing exchanges to experiment with at least 24/5 equities trading (like forex). though there will be some growing pains around lower liquidity on off-hours and systems that might not be ready to internalize the constant demand, we are seeing uniform push towards always-on markets. the point here is that crypto defined the new age trader, which has pressured traditional markets to keep up.

• [product strategy] composability at the front → modular production and composability have been around forever. modern day api products like stripe and shopify are embedded across industries, but in crypto, composability lies the foundation for the main product, not just as an integration. defi protocols are built on public primitives (amms, oracles, vaults, governance modules) that are meant to be forked and they’re doing it live, while handling hundreds of millions in value, in open markets, without central control. how defi primitives have been used gives us a glimpse of what future software development looks like. you see some of this within ai agent platforms. tools like langchain let you chain together small, reusable components (like calling a specific llm) into workflows that other people can fork and extend.

• [infrastructure] distributed computing → traditional distributed compute relied on trusted parties and closed networks. with the introduction of bitcoin and since, crypto brought in open participation, no central fallback, and real monetary value on distributed systems - pushing forward consensus and verifiability in ways traditional infrastructure hasn’t needed to.

  • ai models today are largely closed systems, from the fitted model and its parameters to how inference is served and abstracted from the user. but when models move into sensitive domains like healthcare and credit, transparency should matter. because deepseek’s open source release showed how much demand there is for visibility. decentralized training will be one of the ways that visibility improves; making it work relies on the same crypto-native rails: incentive alignment, distributed governance, and coordination across untrusted parties. model development in the future will borrow from crypto the systems that were built to operate in adversarial, open, high-value environments.

compared to the traditional examples, these focus on a facet of crypto vs. zooming into single companies. we haven’t seen juggernauts emerge from crypto in the same way we did with companies like apple or google or mcdonalds or ford - but that might be the wrong benchmark. crypto runs lean by design: smaller teams, more composability, less overhead. it also tends to serve narrow but high-value markets.

progress over the past ~16 years has produced a set of primitives that define how crypto works and they’re already starting to leak out. crypto was introduced to enable p2p payments, but it’s now informing how software is built, how markets behave, and how distributed systems hold up under pressure.