AI Enables Tiny Teams to Match Hedge Fund Alpha
These tools will be interesting to watch With AI, fundamental can get a quant back office and, with some of these tools, quant can get a fundamental front end. If you walk the halls of a large hedge fund today, you will see very large data engineering teams, teams of many data scientists, forensic accounting teams, survey teams, etc. This headcount intensity has been a barrier to alpha. A key question is whether other funds can replicate these 200 bodies with 5 people and a stack of great AI tools, which increases the competitive intensity of alpha. In my view, the distinction between longer duration, fundamental quant strategies (long term is 3m+ in quant world) and high/intermediate velocity fundamental strategies should see convergence - the question is how much. The promise of quantamental, which has so hard to implement due to cultural, talent & "run of cards" LP dynamics (non-core strategies first to go when they go off the rails), now becomes more promising, and the investors of the future in my opinion are bionic and highly "quanty". But what side wins? Fundamental investors becoming more quanty or quant firms encroaching on fundamental alpha? Even arcane concepts like thematic betas that fundamental investors mostly ignore today become highly relevant to the fundamental investor when you can overlay automatic research on thematic cycles. I believe demand for factor models will grow nicely and be a key piece in the New Investor Dashboard. If I were a large quant firm, I would work to build a coalition of other quant firms to systematize corporate access via recorded management calls (with mgmt discretion), conducted by a team of in-house journalists, eventually AI voice agents. The same way some of the large long only complexes have taken corporate access in house. May not happen, but will we see CFO's taking their 2x monthly call with the "quant coalition"? That would not be great for fundamental investors... (not sponsored by Qualitate)
Tech Cuts Tiger-Cub Teams From 100 to 15
In the 1995-2010 era, it took a lot of bodies to run a scaled Tiger-Cub strategy. 100+ headcount was common across research, back office & trading (starting in '08, I was one of a large research team) Largely due to technology,...
Junior Investment Jobs Face Uncertain Future Through 2027
For the first time since the ChatGPT moment in November 2022, I'm starting to get worried about the labor market outlook for junior investment talent into 2027. Maybe the kids won't be alright.
Claude Cowork: Promising Concept, Frustratingly Buggy Experience
Maybe it’s me but I find Claude Cowork to be fascinating in theory (Claude Code abstracted in a user friendly interface, with MCP integrations) but unusably buggy in practice Maybe a skill issue on my part
Agents Will Turn AI Into Intuitive Investment Assistants
Someone used the metaphor to me this week that chatbot LLMs are like a calculator, good at generating answers to specific questions, but you have to remember to use them. The cognitive burden of pulling your prompt off the...
Track Normalized FCF/Share via Long‑Term Cycle Analytics
This is a great question One of the most important output metrics in any model, in my opinion, is normalized FCF/share. There are many assumptions that lead to your FCF/share metric, including revenue & normalized margin assumptions. But quarter to quarter and year...
Deep Cash‑Flow Modeling Beats EPS for True Value
The senior analyst at a Tiger Cub who trained me on financial modeling back in '08 (shoutout Chris Laporte) walked me through the importance of deeply detailed cash flow builds. So much of the world is focused on EPS (certainly...
AI Excel Nails Basics, Stumbles on Complex Updates
Uploaded my Skill files and tried this out 1) Passed Level 1 Test: simple AMZN quarterly update, where I update the architecture, just needs numbers input (17/17) 2) Badly Failed Level 2 Test: real-world situation with UBER model 6 quarters out of...
AI Agent Pipeline Delivers Institutional-Grade Earnings Previews
I could never quite get a chat-bot to give me a good, institutional-grade earnings preview. It required the confluence of too many separate prompts. I was working on this today for $DHR, and I personally was quite impressed with the result...
AI Can't Replace Traditional Analyst Training for Investors
Not the students fault. Asking students to speed-run an investment process with AI then having them pitch to Dan Loeb & David Tepper is a really bad idea. Hopefully an important lesson to young investors that even with AI, an intellectual power...
One Low‑Vol Short Can Power an Entire Year’s Alpha
It is underdiscussed how much great short selling talent sits at multis, and how important alpha shorting really is to success in the multi-manager model. One big short, particularly in a low vol name you can size in dollars, can make...
Foundation Models Win, Yet Token Costs Threaten Adoption
There has been a long standing debate in Finance AI around "who wins": the foundation labs vs. the finance specific AI platforms (pejoratively, "the wrappers") The pendulum in asset management has shifted back towards a strong consensus that foundation labs directly...
AI Still Lacks True Investment Judgment, Yet Perspectives Shift
The common thing for investors to say about AI over the last 18 months is that "AI can help with process, but has no judgment". When I would hear that, I would pretty much agree. LLMs are fundamentally stateless with...
Multi‑model AI Beats Single LLMs for Skill Building
I'm building my AI-native mock portfolio on Perplexity Computer for a few reasons: - It's multi-model (for example, GPT 5.4 is much better at auditing excel than Opus 4.6, and Claude models haven't been as good at images). The orchestra seems...
Orchestrated AI Pipelines Replace Hundreds of Prompts
The most exciting outputs I am getting from my experimentation in building an AI-native workflow are not the result of prompts or single workflow skills. The really exciting outputs are a function of orchestrated pipelines, which are a set of sequentially...
Human Judgment Still Beats Overhyped AI Investing
"Should we outsource investment judgment to a machine?" While this feels like a new question, "quantamental" stock selection processes were the cutting edge of investment process in the 2010s. The argument was incredibly compelling conceptually: we can blend the best...
AI-Driven Healthcare Coverage: Automating Institutional-Grade Research
I'm building out "AI-native" coverage on my old healthcare coverage (153 healthcare stocks ex-therapeutics that I covered institutionally for ~10 years). I am testing how close I can get to institutional-grade coverage while doing as little as possible manually: ramping...
AI as Research Director Optimizes Investment Time Allocation
I'm deep in the process of building an AI-native investment process, and one of my favorite ideas so far is the "AI Research Director", where the AI system works to orchestrate human activity to its best and highest use. A former...
OpenAI's Toolkit Nears Operational AI Investment Platform
Vibes are very pro-Claude right now, but it does seem that OpenAI has all the pieces to build an incredibly good “Cowork”style agentic work platform: > GPT 5.4 Pro (very good at Excel) > The best Deep Research capability > Codex (a...
Right Data & Context Transform Zero‑Shot AI
Very much agree with this Cowork zero shot is meh When connected to the right data and fed with deep, detailed context: it’s remarkable And once set up, the MCP and Skills architecture make it quite usable
Maximize Context, Supercharge AI Outputs
The new prompting is context-maxing I'm getting incredible outputs by simply dumping in as much context as I can For GPT 5.4, Claude and Perplexity Computer Context is the key leverage point
Perplexity Computer Delivers Instant, Deep Guidance Credibility Analysis
I never found a way to adopt the old version of Perplexity in my investment process, but the things you can one-shot in Perplexity Computer are mind bogglingly good, and it has become a daily driver for me (for the...
GPT‑5.4’s Power Needs Cowork Harness to Outpace Claude
Will be very interested to see a Claude CoWork equivalent here One thing I’ve noticed, when carefully prompted, GPT 5.4 is an unbelievable model (higher ceiling than Opus 4.6 for investment research tasks). Some of the 5.4 Pro outputs I’ve...
AI Adoption Needs Firm-Specific Exoskeletons, Not Just Tools
This is the exact light bulb moment I've had over the last two weeks. Helping firms become AI native is going to be much less around the technical complexity of the actual tooling. There's so much capex and engineering ingenuity pointed...
AI Coding Tools Remain Hard to Adopt, Yet Rapidly Improving
Having spent a lot of the last two weeks experimenting with different user interfaces for Claude Code (including via terminal, Cursor and VS Code) one of my primary observations is how frustrating it can be to get up to a...
Systematic Short Books Boost Agentic Edge for Funds
If you are a long-short equity fund thinking about ways to adopt an agentic approach, I encourage you to think about building a systematic short book. 1) Good shorts tend to rhyme more than good longs and and so are fertile...
AI Workflow Replaces WhaleWisdom for Fund Portfolio Tracking
With 13Fs for Q4 out, retweeting this AI workflow (which I quite like). I will be building a stack of a couple dozen funds I follow to get a quick look at their portfolios: new additions, size ups/downs and portfolio...