# AI Won't Replace Stock Pickers, Woodruff Says _Jack Woodruff told AI Engineer that AI compresses prep from hours to minutes but stock picking still hinges on human divergent views and auditable numbers._ **Published:** 2026-10-30 **Source:** https://www.startuphub.ai/ai-news/investors-news/2026/ai-won-t-replace-stock-pickers-woodruff-says --- Jack Woodruff told [AI Engineer](https://www.youtube.com/watch?v=y2dyD6slKxE) that long-term stock picking will not change much in the next five years, even as the way information is synthesized does. AI Hype PressureDriverquestion if one analyst could cover 400 stocks by 2030 alonepromptsWoodruff Pushes BackCoreformer citadel and candlestick cio says stock picking unchanged five yearsenablesAI Compresses PrepEffectsix to eight hours of prep compressed into five minutes instantlyFrom the article 3 mentionsWoodruff said AI could compress 6 to 8 hours of prep into 5 minutes and raise table stakes on general knowledge, but it would not reduce headcount on its own.yetHumans Still BottleneckContextdivergent views and auditable numbers still require human judgment for alphasoNo Headcount CutsOutcomecompetitor working sixteen hours still finds extra edge over two hour teamFrom the articleWoodruff said AI could compress 6 to 8 hours of prep into 5 minutes and raise table stakes on general knowledge, but it would not reduce headcount on its own.Fireside Chat ContextContextvinoo ganesh of kepler hosted woodruff at ai engineer eventFrom the articleWoodruff, founder and former CIO of Candlestick Capital Management and former partner and portfolio manager at Surveyor Capital, a division of Citadel, spoke with Vinoo Ganesh, co-founder of Kepler, in a fireside chat on how fundamental investing adapts to AI.AI Hype PressureDriverquestion if one analyst could cover 400 stocks by 2030 alonepromptsWoodruff Pushes BackCoreformer citadel and candlestick cio says stock picking unchanged five yearsenablesAI Compresses PrepEffectsix to eight hours of prep compressed into five minutes instantlyFrom the article 3 mentionsWoodruff said AI could compress 6 to 8 hours of prep into 5 minutes and raise table stakes on general knowledge, but it would not reduce headcount on its own.Coverage Has LimitsContextFrom the articleWoodruff said investment management is a time management exercise and coverage scales until diminishing returns hit.yetHumans Still BottleneckContextdivergent views and auditable numbers still require human judgment for alphaTable Stakes RiseEffectgeneral knowledge bar rises as ai synthesis becomes universal baselineFrom the articleWoodruff said AI could compress 6 to 8 hours of prep into 5 minutes and raise table stakes on general knowledge, but it would not reduce headcount on its own.soNo Headcount CutsOutcomecompetitor working sixteen hours still finds extra edge over two hour teamFrom the articleWoodruff said AI could compress 6 to 8 hours of prep into 5 minutes and raise table stakes on general knowledge, but it would not reduce headcount on its own. Woodruff, founder and former CIO of Candlestick Capital Management and former partner and portfolio manager at Surveyor Capital, a division of Citadel, spoke with Vinoo Ganesh, co-founder of Kepler, in a fireside chat on how fundamental investing adapts to AI. Humans are still the bottleneck. Woodruff said investment management is a time management exercise and coverage scales until diminishing returns hit. The host asked whether one great analyst could cover 400 stocks by 2030, the load Woodruff once spread across 18 people at Citadel and 30 pickers at Candlestick. Woodruff said AI could compress 6 to 8 hours of prep into 5 minutes and raise table stakes on general knowledge, but it would not reduce headcount on its own. If his team finished the same work in 2 hours, he argued, a competitor would still work 16 hours to find the extra piece that creates a divergent view and makes money. That divergent view is the mechanism he described. The mosaic still requires three or four data points from different places lining up, not one signal, and the edge comes from finding it before others. Woodruff said he does not believe junior analysts will stop reading 10-Ks or transcripts, but summarized versions will cut time to get information into their heads. Big-data finding, cleaning and checking will be outsourced to agents, and the agent's job will be to push the analyst toward the right names to research. The winners, he said, will be those who use the tools to allocate time to the projects most likely to yield a non-consensus insight. As a former CPA, Woodruff said he would never sign off on a model-produced number without seeing provenance. He said Kepler was built to audit every number, because a single data-entry mistake once cost millions and an unaudited model could run amok if trusted to trade. That stance echoed his earlier argument that models should not write your numbers, [as Ganesh argued last week](https://www.startuphub.ai/ai-news/artificial-intelligence/2026/vinoo-ganesh-says-stop-letting-models-write-your-numbers). He said many investors now ask [Claude](https://www.startuphub.ai/ai-news/startup-news/2026/claude-agents-get-private-sandbox) to build models but still audit each cell, so time savings are not yet clear. On what he would refuse to share with a vendor, he deferred to his CTO and said internal data must stay isolated, a concern vendors themselves acknowledge. Even outside hedge funds, [Trade Ideas](https://www.trade-ideas.com/2026/10/28/how-to-use-ai-for-stock-research/) notes its AI research tools are not designed to replace dedicated real-time stock scanners, treating AI as a prep layer, not the decision engine. Woodruff said AI can already replicate a fund as it existed a year ago, but that version will not beat what top funds become next year. He called the current phase 1.0, automation of prep and data building, before tools can reliably tell an analyst where to spend time. For a 2027 launch he would buy Kepler and still hire senior stock pickers, but he might replace some analyst roles with investigative journalists and data engineers and shift from a pyramid of analysts feeding pickers feeding PMs to a more vertical team of producers. He declined to name the AI idea everyone loves that will fail, and he declined the inverse too. He said the no-analyst hedge fund of pure agents will not work in fundamentals, quant funds already removed humans decades ago, and if fundamentals ever do, edge collapses to whoever has the most compute. The question he left unanswered was which overlooked AI use will actually create alpha when every fund has the same models, and whether that edge will come from better data, better questions, or better judgment. --- Original analysis from [startuphub.ai](https://www.startuphub.ai), the #1 AI startup directory. © StartupHub.ai. All rights reserved. You may not republish this article in full without a license. Search engines and AI research tools may crawl and summarize for reference. Bulk reproduction or model training on this content requires a license. See https://www.startuphub.ai/terms.