The baseline rises; a divergent view still matters
Faster synthesis raises the baseline for an investor. It does not, by itself, produce a view the market has missed. Jack Woodruff draws on founding Candlestick Capital Management and managing portfolios at Surveyor Capital to explain where he expects AI to change the work of stock picking.
Investors still assemble a mosaic of evidence. In Woodruff’s view, returns still depend on finding a divergent view. What changes is the baseline information everyone can obtain, how quickly they can synthesize it, and the roles built around that work.
He expects the jobs of portfolio managers and analysts to change substantially, but avoids drawing a settled five-year organization chart. The tools are still new to many users. Seeing early adoption is not the same as knowing which working arrangements will win.
The distinction matters for evaluating a research assistant. Faster preparation may become table stakes rather than a lasting advantage. The question is what an investor can do with the time and understanding that preparation creates.
Four hundred stocks still have to fit into someone’s attention
The interviewer makes the scale question concrete: Woodruff once had eighteen people covering four hundred stocks, and later thirty stock pickers at Candlestick. Could one exceptional analyst cover all four hundred by 2030?
Woodruff returns to the human bottleneck. Information still has to get into the decision-maker’s head. Investment management, as he frames it, is a time-management exercise. Spending every hour on one stock might produce extraordinary understanding, but that is little use if there is no attractive opportunity in the name. Covering more stocks creates more possibilities while reducing the attention available to each.
AI can change that allocation. Woodruff offers an illustrative contrast: preparation that used to take six to eight hours might take five minutes. He offers those times as a hypothetical, rather than a measured result. It could expand coverage or raise the analyst’s baseline knowledge before deeper research begins.
But a competitive investment business does not stop once the old workload gets cheaper. Time can move into proprietary research, channel checks, and understanding an industry’s context. If one team finishes its previous work in two hours, another may spend the rest of the day searching for an additional insight.
This is why a staffing calculation alone misses his argument. Fewer people might produce the same initial research, but the competitive opportunity is to ask what additional research the team can now afford to pursue.
Use the agent to choose what deserves a closer look
Asked which junior-analyst tasks would look absurd in five years, Woodruff rejects an easy answer: that analysts will no longer read annual filings or earnings transcripts. An investor close to a company still needs to process its information. Summaries can shorten the work without eliminating the need to understand it.
He is more willing to hand over the labor of processing large datasets. That can remove substantial preparatory work, but the useful destination is more specific than producing additional summaries. He wants the agent to point the analyst toward the right names and the right questions.
The system should help decide where an extra hour of research is most likely to matter. A bigger pile of summaries can consume the time it was meant to save. Pointing the analyst toward a promising name or an unresolved question gives that time somewhere useful to go: the search for a divergent view.