The Everyone Codes Economy
Ben Tossell@bentossell
−2.1%
since Sep 30
The idea
I believe AI is making ordinary people technical enough to build their own software, and that shift plays out over a decade, not a quarter. So I own both sides: the toolmakers turning people into builders, and the layers underneath what they build on, clouds, databases, deployment, plus the things nobody rebuilds regardless of skill, like payments and identity. It's spread across large-cap ballast, mid-sized platform bets, and a compute backbone, with a small cash reserve to add if things get cheaper. I'd rethink it if people stop wanting to build their own tools.
What it owns
- Every new builder needs somewhere to run their software, and cloud platforms rent computing power to them. · AmazonAWS is the default cloud platform where most new builders will run their software., AlphabetIt offers both Gemini for code-writing and Google Cloud for hosting what new builders create., CloudflareIt is the layer a beginner reaches for when they want their app live without learning server infrastructure., DigitalOceanThe platform targets technical developers operating independently rather than through enterprise IT departments, a segment that expands as coding ability spreads across more people.23%
- Payments, identity, security, and storefronts remain outsourced services even as people become more technical, so each new builder is a new customer. · ShopifyEven when people can build their own software, they keep renting the storefront because payments and fraud prevention are too complex to rebuild., VisaPayments is the clearest service nobody rebuilds no matter how technical they get, so more apps means more transactions on its rails., CrowdStrikeA decade of software written by newly technical people means a decade of security holes, and it sells the defense., OktaIdentity and login infrastructure remain too risky and complex for most teams to build in-house, keeping specialized identity platforms essential regardless of developer skill levels., TwilioTexts, calls, and verification codes are plumbing that even a capable amateur rents from it instead of building.22%
- Millions of new builders writing code with AI means sustained demand for the chips and infrastructure that run those tools. · NVIDIAEvery AI coding assistant in this take runs on its chips, so a wave of new builders means sustained demand for its hardware., Taiwan SemiconductorIt manufactures the leading-edge chips that NVIDIA, Broadcom, and other compute providers need, so it captures demand from all sides of the take., BroadcomIt sells the networking and custom chips that stitch AI data centers together, the plumbing underneath the clouds new builders will rent.18%
- The take rests on tools that let non-programmers write working software, turning millions of new people into builders. · MicrosoftIt owns GitHub Copilot and the tools most ordinary people will touch when AI makes them technical enough to build software., FigmaIt is where a non-technical person's idea starts, and its design-to-code tools are the bridge from sketch to shipped software., GitLabIt bets that AI-assisted development becomes the whole platform, not just a plugin, if its AI features keep pulling in new builders.17%
- Amateur-built apps still need databases and monitoring tools, which developers rent rather than build from scratch. · MongoDBTutorials and AI assistants recommend it first, so new builders building their first apps will gravitate to it for data storage., DatadogSoftware breaks in production, and monitoring tools help teams detect and fix failures faster, making observability increasingly valuable as more applications reach production environments., SnowflakeNobody who just learned to build is going to run their own data warehouse, so they rent from it instead.13%
- Cash7%