The Edge Has Shifted | Matt Reustle on How the Best Investors Use AI

| Stock Investing | February 25, 2026 | 4.02 Thousand views | 1:03:37

TL;DR

Matt Reustle explains how AI evolved from basic search tools to agentic research assistants capable of automating due diligence and monitoring workflows, allowing investors to leverage technology for junior-level tasks while maintaining human decision-making at the core.

🧠 The AI Inflection: From Toy to Tool 3 insights

Deep research models changed the game

The inflection point arrived in late 2024 and early 2025 with deep research models that function more like agents than traditional LLMs, moving beyond social experiments to professional-grade tools.

LLMs are calculators, agents are analysts

Basic LLMs operate like interns performing single Google searches, while agentic workflows use reasoning to deliver multi-step analysis resembling junior analyst output rather than intern-level summaries.

Premium tiers unlock reasoning capabilities

Free-tier models provide surface-level answers, but premium agentic tiers understand context and deliver synthesized research that investors can actually use for professional decisions.

âš¡ Agentic Workflows: Practical Applications 3 insights

Automated cross-sector monitoring

Investors can set up recurring monitors to track specific commentary across unrelated sectors, such as logistics mentions in CPG earnings calls, delivering synthesized alerts without manual transcript review.

Instant due diligence acceleration

Deep research capabilities eliminate the need for outdated sellside primers by generating comprehensive company summaries and thematic research almost instantly.

KPI tracking and position monitoring

Systems can automate quarterly recaps of key data points across entire portfolios, replacing the manual work of earnings-season transcript reviews previously done late into the night.

🚀 Implementation Strategy 3 insights

Journal your process first

Before adopting tools, document your current research workflow to identify specific friction points where AI can provide immediate leverage rather than applying generic solutions.

Progressive adoption path

Start with premium LLM tiers for communication and basic research, then graduate to specialized investment-specific tools for monitoring and thematic screening.

Compliance remains critical

While scraping capabilities exist, investors must navigate legal restrictions and API requirements when setting up automated data collection systems.

Bottom Line

Treat AI as a leverage tool that handles junior-level research and monitoring tasks, freeing you to focus on high-conviction decision-making and relationship-building that algorithms cannot replicate.

More from Excess Returns

View all
The $2 Trillion Question | Tobias Carlisle on SpaceX, the AI Buildout, and the Rotation No One Sees
58:19
Excess Returns Excess Returns

The $2 Trillion Question | Tobias Carlisle on SpaceX, the AI Buildout, and the Rotation No One Sees

Tobias Carlisle warns that the market is at historic valuation extremes comparable to the dot-com bubble, but argues investors should rotate into deeply undervalued small and micro-cap value stocks rather than exit entirely, as early indicators suggest a potential decade-long rotation away from large-cap growth; meanwhile, he cautions that the massive AI infrastructure buildout risks following historical boom-bust patterns where value accrues to consumers, not creators.

4 days ago · 10 points
The Trillion-Dollar Gap | We Asked Aswath Damodaran What SpaceX Is Really Worth
1:08:27
Excess Returns Excess Returns

The Trillion-Dollar Gap | We Asked Aswath Damodaran What SpaceX Is Really Worth

Finance professor Aswath Damodaran analyzes SpaceX's $2.7 trillion valuation, finding that while the space launch and Starlink businesses hold real competitive advantages, the AI division's projected $26 trillion market relies on terrible unit economics and a contradictory strategy of renting data centers to direct competitors.

6 days ago · 9 points