Claude Code, the Finance Junior Analyst + The Global Memory Shortage: Doug O'Laughlin, SemiAnalysis

| Podcasts | February 24, 2026 | 6.75 Thousand views | 2:07:40

TL;DR

Doug O'Laughlin describes Claude Code as a capable but imperfect 'junior analyst' that amplifies expert productivity rather than replacing human judgment, while recounting his journey from value investing to semiconductor analysis through his early conviction that Moore's Law was ending.

💻 AI Tools as Force Multipliers 4 insights

Claude Code serves as a junior analyst

O'Laughlin views Claude Code as equivalent to a junior analyst that handles painful information gathering tasks, but emphasizes it makes mistakes constantly and still requires expert oversight to synthesize final decisions.

Missing meta-level learning capability

Unlike human analysts who develop intuition and expertise through repetition, current AI lacks the ability to achieve meta-level learning or consistent hit rates that elevate practitioners to expert status.

Real-world hiring case study application

SemiAnalysis has tested Claude Code since March 2024 on financial analyst hiring case studies requiring 24-hour human-equivalent tasks, finding it effective for agentic workflows but obvious when producing 'slop'.

Rapid code generation adoption

Updated data shows AI now generates approximately 5% of code and climbing, which O'Laughlin considers a 'weapons-grade' tool essential for gaining information edges in financial analysis.

Semiconductor Disruption Thesis 4 insights

Moore's Law is dead, creating pricing power

O'Laughlin's foundational 2018 thesis held that Moore's Law's end would terminate free performance gains, forcing the industry to reward companies capable of genuine architectural innovation with significant pricing power.

2020 Nvidia scaling laws prediction

He predicted in 2020 that exploding demand from scaling laws combined with supply constraints would benefit parallel compute leaders, specifically identifying Nvidia as the primary beneficiary months before the market recognized the trend.

Magnitude exceeded bullish expectations

While the directional thesis proved correct, O'Laughlin admits the scale of Nvidia's rise to become the world's most valuable company surpassed even his most aggressive predictions.

ASML as science fiction moat

His investment journey began with ASML in 2018, where the 'science fiction' complexity of EUV lithography revealed how semiconductor manufacturing had become impossibly difficult, creating insurmountable competitive moats.

🎯 Career Strategy & Conviction 3 insights

From Value Mule to semiconductor specialist

O'Laughlin transitioned from anonymous value investor 'Value Mule' to semi-analysis after ASML 'nerd sniped' him into studying chipmaking physics, realizing the industry was misunderstood as 'mature' when it was actually entering a new era.

Collaboration through shared conviction

He partnered with Dylan Patel after discovering he was the only other analyst equally 'semicropilled' regarding the end of Moore's Law, leading to SemiAnalysis's formation around the shared thesis.

All-in trend following philosophy

Describing himself as skilled at trend spotting, O'Laughlin advocates identifying massive waves early—citing his 2019 TikTok obsession as an example—and reorienting one's entire career around high-conviction opportunities rather than diversifying attention.

Bottom Line

Treat AI coding tools as junior analysts that amplify expert productivity but require human oversight for quality control, while the end of Moore's Law has permanently shifted semiconductor value creation toward integrated systems companies capable of architectural innovation rather than process shrinks.

More from Latent Space

View all
Dreamer: the Agent OS for Everyone — David Singleton
1:04:23
Latent Space Latent Space

Dreamer: the Agent OS for Everyone — David Singleton

David Singleton introduces Dreamer as an 'Agent OS' that combines a personal AI Sidekick with a marketplace of tools and agents, enabling both non-technical users and engineers to build, customize, and deploy AI applications through natural language while maintaining privacy through centralized, OS-level architecture.

5 days ago · 9 points