AI for Investing
What Actually Works — and What Will Lose You Money
Quant funds are printing money. Robo-advisors manage trillions. ChatGPT reads 10-Ks in seconds. But most of what you've heard is hype. Here's the real story — from someone who ran a hedge fund and uses Claude daily for research.
$400B+
Robo-Advisor AUM
~35%
US Trades by Algos
66%/yr
Medallion Fund Returns
10 min
AI 10-K Analysis
TL;DR — Glen's Take on AI Investing
Where AI Actually Helps
- ✓Reading 10-Ks, earnings calls, and SEC filings in minutes
- ✓Comparing financial metrics across dozens of companies instantly
- ✓Automated portfolio rebalancing and tax-loss harvesting
- ✓Understanding industries and business models you're new to
Where AI Will Lose You Money
- ×Asking “what stock should I buy?” and trusting the answer
- ×Backtested trading strategies that only work on past data
- ×Replacing your judgment with AI confidence
- ×Trying to out-trade quant funds with a $29/mo subscription
AI is the most powerful research tool investors have ever had. But a research tool is not a decision-maker. The best investors in history used simple frameworks and good judgment. AI accelerates the research. You still provide the judgment.
How AI Is Used in Investing Today
AI isn't coming to finance — it's been here for decades. What's new is that powerful AI tools are now accessible to individual investors.
Algorithmic Trading
AI executes trades in milliseconds based on pattern recognition, arbitrage, and market microstructure signals. HFT firms process millions of data points per second.
Who uses it
Hedge funds, prop firms, market makers
Retail access
Minimal. Requires co-located servers and PhD engineers.
Sentiment Analysis
NLP scans news, earnings calls, social media, and SEC filings to gauge market sentiment before it moves prices.
Who uses it
Quant funds, institutional investors, fintech
Retail access
Moderate. StockTwits, MarketBeat offer basic scores.
Portfolio Optimization
AI analyzes correlations, volatility, and returns across thousands of assets to maximize return per unit of risk.
Who uses it
Robo-advisors, wealth managers, pension funds
Retail access
Good. Betterment, Wealthfront, Vanguard Digital Advisor.
Risk Management
ML models predict tail risks, stress-test portfolios, and find hidden correlations traditional models miss.
Who uses it
Banks, insurance companies, asset managers
Retail access
Limited. True AI risk management is institutional.
Fundamental Research
AI reads and summarizes 10-Ks, earnings transcripts, and industry reports in minutes. It can compare financial metrics across hundreds of companies simultaneously.
Who uses it
Analysts, portfolio managers, individual investors (increasingly)
Retail access
Excellent. Claude, ChatGPT, and specialized tools make this accessible to anyone.
AI Investing Tools for Retail Investors
You don't need a quant team or a Bloomberg terminal. These tools actually deliver value — ranked by impact.
⭐ Robo-Advisors Highest impact
Betterment, Wealthfront, Vanguard Digital Advisor — automated rebalancing and tax-loss harvesting at ~0.25%/yr. For the 80% of investors who would panic-sell in a crash, this is the single most valuable AI tool.
🧠 AI Research Assistants (Claude, ChatGPT)
The game-changer for active investors. Upload a 10-K and get analysis in minutes. Compare competitors, model scenarios, understand industries. This is what I use daily. Always verify numbers — AI hallucates financial data.
🔍 AI Stock Screeners
Natural language queries: “profitable mid-caps with accelerating revenue and insider buying.” Finviz, TradingView, Koyfin.
📈 Sentiment Tools
Scan social media for market mood. Useful as a contrarian indicator. The signal is real but noisy.
📊 Portfolio Analyzers
Morningstar, Empower, Portfolio Visualizer. Risk exposure, concentration, drawdown scenarios.
How Hedge Funds Use AI
The real AI arms race is inside quant hedge funds with billions in capital and thousands of engineers. Understanding what they do helps you understand why retail AI trading is mostly a fantasy. You're not competing against other retail investors — you're competing against these firms.
Renaissance Technologies
Jim Simons (mathematician)
AUM: ~$130B
Pure quantitative. Medallion Fund: ~66%/yr before fees for decades. Hires mathematicians and physicists, not MBAs.
Key insight: They don't understand WHY a pattern exists. They find statistical edges and exploit them before they disappear.
Two Sigma
David Siegel & John Overdeck
AUM: ~$60B
ML + distributed computing + massive alternative data (satellite imagery, credit card transactions, weather).
Key insight: Their edge is data engineering. They process data most investors don't know exists.
D.E. Shaw
David Shaw (computer scientist)
AUM: ~$60B
Systematic + discretionary since 1988. One of the earliest quant firms. Jeff Bezos worked here before Amazon.
Key insight: Using computational finance before 'AI investing' was a phrase anyone used.
Bridgewater Associates
Ray Dalio
AUM: ~$125B
Systematic macro. 'Pure Alpha' uses hundreds of rules-based systems modeling economic relationships globally.
Key insight: Macro cause-and-effect relationships, systematized into repeatable algorithms.
Can AI Beat the Market?
The honest answer is nuanced — and nuance doesn't go viral.
Yes, If You're Renaissance Technologies
- •Medallion Fund: ~66%/yr before fees for 30+ years. Real AI alpha, but closed to outsiders since 1993.
- •Top quant funds employ hundreds of PhDs in math and physics, spending $500M+ annually on data and infrastructure.
No, If You're Using a $29/mo App
- •Any strategy available to everyone has no proprietary edge. Backtests look amazing; live performance doesn't.
- •A simple S&P 500 index fund has beaten ~90% of active funds over 15+ years. AI hasn't changed that math.
Bottom line: AI can beat the market — at a scale and cost you can't replicate. For retail investors, AI's real value is making you a better researcher, not a faster trader.
Using Claude & ChatGPT for Investment Research
This is where AI delivers the most value. Not trading. Not stock picking. Research. I can analyze a 10-K in 10 minutes instead of 3 hours. Here are the exact prompts I use.
Analyze a 10-K Filing
Read this 10-K and summarize: revenue trends (3-year), gross/operating/net margins, debt levels, cash flow generation, key risk factors, and management's forward guidance. Flag anything unusual.
Why it works: A 10-K is 100-300 pages. AI reads it in seconds and surfaces what matters.
Compare Competitors
Compare [Company A] and [Company B]: revenue growth, profitability, valuation multiples (P/E, EV/EBITDA, P/FCF), balance sheet strength, and competitive moat. Which is the better value today?
Why it works: Side-by-side comparison that would take hours manually. AI spots the differences instantly.
Understand an Industry
Explain the economics of the [industry] business. What drives revenue? What are the key cost structures? Who are the major players and what are their competitive advantages? What disruption risks exist?
Why it works: Before investing in a company, understand the industry it operates in.
Earnings Call Analysis
Analyze this earnings call transcript. What did management emphasize? What questions did analysts push back on? Were there any changes in tone or guidance from last quarter? Any red flags?
Why it works: Earnings calls contain subtle signals. AI catches tone shifts humans miss on first read.
Warning: AI hallucates financial data. I've seen Claude confidently state wrong revenue numbers and fabricated quotes. Use AI to find what to look at. Verify in the actual filing.
Get Glen's Musings
Occasional thoughts on AI, Claude, investing, and building things. Free. No spam.
Unsubscribe anytime. I respect your inbox more than Congress respects property rights.
Glen's Experience: From Hedge Fund to AI-Powered Research
I ran Global Speculation LP, a deep-value activist hedge fund focused on mortgage REITs and GSE investments (Fannie Mae, Freddie Mac). I published on Seeking Alpha, filed 13D/13G reports with the SEC, and spent 12 years as a vocal activist investor.
Back then, my “AI” was a spreadsheet and 14 hours of reading SEC filings. Today, I use Claude to analyze 10-Ks in minutes, compare entire sectors, and stress-test theses by asking Claude to argue the bear case.
“AI doesn't replace judgment. It accelerates research.”
My workflow: (1) Screen with value metrics. (2) Feed the 10-K to Claude. (3) Read the filing myself, focusing on flagged sections. (4) Build a valuation model. (5) Decide. Steps 2-3 used to take 80% of my time. Now they take 20%.
The Dangers of AI Investing
These are the traps most “AI stock trading” content glosses over.
Overfitting
An AI model that 'predicts' past stock prices perfectly will fail spectacularly on new data. Backtests always look incredible because the model was trained on that exact data. This is the #1 trap in AI investing.
Backtest Bias
Every AI trading strategy shows amazing backtest results. That's because you only see the ones that worked. The thousands of strategies that failed were discarded. This is survivorship bias dressed up as machine learning.
Herd Behavior
If everyone uses the same AI models trained on the same data, they'll all make the same trades at the same time. This amplifies volatility and creates crowded trades that blow up spectacularly when everyone exits at once.
The 'AI Told Me To' Trap
Outsourcing your judgment to an AI chatbot. 'Claude said this stock is undervalued' is not due diligence. AI is a research accelerator, not an oracle. It hallucinates financial data, gets numbers wrong, and has no skin in the game.
What Happens When AI Runs the Hedge Fund?
I wrote a screenplay about this exact question. “THE ALGORITHM” is a 15-scene screenplay about an AI that gradually takes over a hedge fund's trading decisions — and the humans watching their own obsolescence in real time. Fiction, but the tension is real.
Read “THE ALGORITHM” →Frequently Asked Questions
Can AI beat the stock market?
For retail investors, almost certainly not. Strategies that consistently beat the market (like Renaissance's Medallion Fund) require billions in infrastructure, proprietary data, and PhD-level talent. The AI on the other side of your trade was built by 200 PhDs with a $500M data budget. Use AI to research better, not to trade faster.
Should I use a robo-advisor like Betterment or Wealthfront?
For most people, yes. They offer automated rebalancing and tax-loss harvesting at ~0.25% annually vs ~1.0% for human advisors. They won't beat the market, but they prevent behavioral mistakes like panic selling. If you have under $500K and don't want to actively manage, a robo-advisor is a smart choice.
How do I use ChatGPT or Claude for investment research?
Use AI as a research assistant, never a decision-maker. Summarize 10-K filings, compare competitors, explain industries, analyze earnings calls. Always verify numbers — AI hallucates financial data regularly. Best workflow: AI surfaces what to investigate, you verify in primary sources. 80% time savings, 100% your decisions.
What is algorithmic trading and can retail investors do it?
Algo trading executes trades based on programmatic rules. Platforms like QuantConnect and Alpaca let retail investors build strategies, but retail algos rarely produce consistent alpha. Institutions win on speed (microseconds), data (proprietary), and scale (millions of signals). Great for learning about markets, poor for generating returns.
Is AI sentiment analysis useful for stock picking?
Moderately. Social media sentiment has some predictive power for short-term moves, but the signal is noisy and decays fast. By the time you see it, institutions have already traded on it. More useful as a contrarian indicator (extreme fear = buying opportunity) than for specific stock picks.
Will AI replace human financial advisors?
Partially. AI replaces the computational parts: rebalancing, tax optimization, basic planning. But a good advisor's real value is behavioral coaching — keeping you from panic selling at 3 AM during a 40% drawdown. The future is hybrid: AI handles the math, humans handle the emotions.
How does Glen Bradford use AI for investing?
Glen ran Global Speculation LP, a deep-value activist hedge fund. Today he uses Claude to analyze 10-K filings, compare financials, and stress-test theses. AI accelerates research from hours to minutes, but every investment decision is still driven by human judgment about moats, management, and valuation.
What are the biggest risks of using AI for investing?
The top risks: (1) Overfitting — strategies that memorized noise, not signal. (2) Hallucination — AI states incorrect financial data confidently. (3) Herd behavior — everyone using the same models crowds into the same trades. (4) False confidence — AI outputs feel authoritative, so investors skip due diligence. The meta-risk: AI makes bad decisions faster.
Recommended Resources
Tools & books I actually use and recommend
SeekingAlpha Premium
Quant ratings, earnings transcripts, and the stock analysis community where I published 300+ articles.
Try SeekingAlphaA Random Walk Down Wall Street
Burton Malkiel's classic case for index investing. The book that convinced millions to stop stock-picking.
View on AmazonThe Little Book of Common Sense Investing
John Bogle's manifesto on why low-cost index funds beat everything else. Straight from the founder of Vanguard.
View on AmazonSome links above are affiliate links. I only recommend products I personally use. See my full disclosures.
Keep Exploring
Best AI Tools
Ranked breakdown of the AI tools worth your time in 2026 — and the ones that are just hype.
Read moreCompareClaude vs ChatGPT
Head-to-head comparison of the two leading AI assistants. Which one is better for what?
Read moreTrendingAI Replacing Jobs
Which jobs are safe, which are toast, and how to position yourself for the AI economy.
Read moreMarketsTop 25 Market Crashes
Every major crash ranked and analyzed. What caused them, how bad they got, and what happened next.
Read moreGuideHow to Invest $10K
A step-by-step guide to investing your first $10,000 wisely. No gimmicks, no day trading.
Read moreCompareIndex Funds vs Mutual Funds
The data-driven case for why index funds beat actively managed mutual funds over time.
Read moreDisclaimer: AI is a research tool, not a financial advisor. Nothing on this page is financial advice. AI models hallucinate financial data — always verify against primary sources. Past performance is not indicative of future results. Amazon links are affiliate links (tag: glenbradford-20).