Doug O'Laughlin

Doug O’Laughlin is the president of SemiAnalysis, where his responsibilities include semiconductor strategy, market intelligence, competitive analysis and shaping the firm’s research direction. His AI work includes coauthoring “Claude Code is the Inflection Point” and using Claude Code for SemiAnalysis research. He argues that AI can amplify expert analysts while remaining prone to mistakes.[2][4][1]

On February 5, 2026, O’Laughlin published “Claude Code is the Inflection Point” with Jeremie Eliahou Ontiveros, Jordan Nanos, Dylan Patel and Daniel Nishball. They argued that AI’s value would increasingly come from orchestrating tools, memory, sub-agents and verification loops to complete work. Their emphasis was on outcomes: a model’s responses become useful through a system that plans, acts and checks its results.[4]

The authors described information work as a read, think, write, verify cycle: ingest unstructured information, apply domain knowledge, produce structured output and check it against standards. They reported that SemiAnalysis industry-model analysts used Claude Code to generate diagrams and analyses from large datasets. This placed coding agents within a broader research workflow, including processing documents and communicating findings.[4]

The article estimated that Claude Code authored 4% of GitHub public commits at publication and forecast that, at the then-current trajectory, it would exceed 20% of daily commits by the end of 2026. O’Laughlin and his coauthors also argued that agents threatened software businesses built around human-operated interfaces. They framed Microsoft’s strategy as a tension between growing Azure by supplying AI companies and investing in Copilot to protect Office 365.[4]

In his February 24, 2026 Latent Space interview, O’Laughlin offered a more qualified account of AI’s capabilities. He compared it to a junior analyst who gathers information for an expert to make a decision, recalling that he had once done that junior work himself. He doubted that AI yet possessed the “meta level learning” through which repeated analysis develops expertise. Nevertheless, he said that within his firm it “massively amplifies everyone who is an expert.”[1]

Before his 2026 commentary on AI-assisted research, O’Laughlin’s work centered on investing and explaining semiconductors. His earlier career and publication provide context for the domain expertise he brought to those discussions.[6][3][1]

In his first-person career account, O’Laughlin says he joined the Texas investment firm Bowie Capital out of college and focused on compounders and quality-focused investing. He credits his mentor and former boss, Cory Whitaker. He subsequently became interested in the end of Moore’s law and its implications for semiconductor stocks, and says he started Fabricated Knowledge to pursue that interest.[6]

O’Laughlin publicly introduced Fabricated Knowledge on November 23, 2019, identifying semiconductors as its subject. He describes the publication as plain-English explanations of semiconductor companies from an investment perspective. His favored primer series covers topics including packaging, metrology, lithography, timing markets and CXL.[3][6]

On January 3, 2022, Dylan Patel announced Transistor Radio as a planned podcast collaboration with O’Laughlin of Fabricated Knowledge. Their proposed opening conversation concerned the 2022 semiconductor market. Patel described the project as an outlet for broader market projections, semiconductor news and informal opinions that did not fit SemiAnalysis’s full written articles.[5]