Netflix Chief Product and Technology Officer Elizabeth Stone told Lenny Rachitsky on July 19 that systems thinking is the most important skill she now hires for, and that Netflix is buying fewer narrow stack specialists as a result. Five months earlier, Andreessen Horowitz led a $300 million round into a company whose entire product is systems thinking rendered as infrastructure.
That company is Temporal, which closed its Series D in February at a $5 billion valuation, double its October mark. Temporal does not make code arrive faster. It guarantees that long-running workflows resume from the point of failure rather than restarting, which is the wall teams hit once agents run for hours across half a dozen services. Chief executive Samar Abbas argued in the announcement that agentic AI stalls because the surrounding systems cannot handle real-world execution, not because the models are weak. Investors paid $300 million for the scaffolding.
Kestra raised a $25 million Series A in March to standardize orchestration across data, AI and business workflows after executing more than two billion workflows in 2025. Port raised a $35 million Series B for an internal developer portal built on catalogs, scorecards and golden paths, which is systems thinking sold as a subscription. All of this sits inside an extraordinarily concentrated market. Global startup investment reached a record $510 billion in the first half of 2026, with AI companies capturing more than 70% of second-quarter dollars, according to Crunchbase.
The evidence under the hiring shift
Google’s 2025 DORA study, based on responses from nearly 5,000 technology professionals, found that AI operates as an amplifier. It magnifies what an organization already does well and exposes what it does badly. Ninety percent of surveyed organizations had adopted at least one internal platform, and platform quality correlated directly with whether AI investment produced measurable value. DORA’s stated conclusion is that returns come from the underlying system, meaning workflow clarity, data quality and team alignment, rather than from the tools themselves.
Org charts are following; Gartner forecasts that 80% of large software engineering organizations will run dedicated platform teams by the end of 2026, up from 45% in 2022. Hiring data points the same way. Lightcast found that demand for cross-functional collaboration in AI job postings nearly doubled between the second and third quarters of 2025 and more than doubled again into the fourth, with stakeholder management and operational efficiency close behind. In the 2026 Stanford AI Index, Lightcast reported that the fastest long-term growth in AI postings came from deployment skills such as scalability and workflow management. Postings that name AI skills carry a 28% salary premium, worth roughly $18,000 a year.
Where the thesis could break
The skeptical read is that systems thinking is a hiring platitude that survives because nobody can measure it. METR’s randomized controlled trial complicates that. Sixteen experienced open-source developers took 19% longer to finish real tasks when allowed to use early-2025 AI tools, while estimating afterward that they had gone 20% faster. METR published new data in February 2026 on late-2025 tools that pointed toward a speedup, and labels the original result historical. The models changed, but so did the review paths, the context supplied to the tools and the workflows around them. Isolating the model as the cause is the mistake the data keeps punishing.
Stone’s operational definition is cheap to test and hard to fake; jusy zoom out one level from the task in front of you and ask whether it should become a platform capability rather than a local fix. She pairs that with a trend away from stack-layer specialists toward adaptable generalists, and with a single company-wide expectation of AI fluency applied at every level instead of a rewritten career ladder. Alexis Ohanian, Reddit cofounder and founder of Seven Seven Six, made the blunt version of the argument on X, writing that people who can hold the whole system in their head will run circles around everyone else.
For founders, the sellable form of systems thinking is scaffolding another company can buy: orchestration, evaluation harnesses, service catalogs, design systems, guardrails. Temporal’s valuation doubled in four months on that premise, and the category is still early. For investors, the sharpest diligence question in an AI application deal is no longer which model sits underneath. It is whether the buyer runs a platform team, maintains a source-of-truth data layer and has a review path for what ships, because DORA’s evidence says deployments underperform without them. For operators, the arbitrage has a clock on it. Once 80% of large engineering organizations have platform teams, the scarce skill will be judging which abstractions deserve to exist. Netflix is hiring for that problem now. Most of the market is still hiring for the last one.
