February 2026 Random Ramblings

Episode Summary

Executive Summary: Andrew Walker’s monthly ramble centers on the market’s fear-driven reaction to AI, especially the selloff in SaaS and spillover into offices, trucking, insurance, and brokers. He argues AI’s exponential progress could hollow out software-heavy businesses with little tangible asset backing, but he also notes network effects, human adoption lag, and the need for investors to balance conviction with humility as they update priors in real time.

Main Topics: AI panic spreading across sectors (Priority: 5/5): Walker describes a market-wide fear trade as AI concerns move beyond SaaS into offices, LTL trucking, insurance, and brokers, producing sharp one-day selloffs across unrelated names. Why SaaS is especially exposed (Priority: 5/5): He argues SaaS businesses may be vulnerable because their value is concentrated in software engineers, contracts, and future cash flows rather than tangible assets, which could be displaced quickly by AI. Exponential progress and investor psychology (Priority: 5/5): Walker stresses that humans are bad at understanding exponential improvement, and AI’s rapid advancement makes it difficult to judge how quickly incumbents could be disrupted. Counterarguments: network effects and adoption friction (Priority: 4/5): He notes that building something that looks like Salesforce is not the same as replacing Salesforce, because network effects, customer switching costs, and enterprise complexity still matter. Hard assets as a relative refuge—and their own risks (Priority: 4/5): He discusses renewed interest in hard assets like coal, steel, cement, and power-related assets as possible AI-resistant or AI-benefiting holdings, while warning AI could also disrupt energy economics via batteries and storage. Updating priors and avoiding permanent bearishness (Priority: 5/5): Walker reflects on how investors should avoid becoming the market commentator who stays bearish forever, emphasizing the need to revise views as new information arrives. Arrogance, humility, and the craft of investing (Priority: 5/5): He closes by describing investing as a mental game requiring a knife-edge balance between conviction and humility, aggressive action when an edge exists, and restraint when it does not.

Key Arguments: AI is improving exponentially, and humans systematically underestimate how quickly that compounds into real business disruption. SaaS firms may be uniquely exposed because much of their valuation is tied to intangible assets and future cash flows that can re-rate quickly if AI substitutes core functions. The market’s one-day, sector-wide selloffs suggest a genuine panic/dislocation rather than isolated stock-specific issues. A basic AI product does not yet equal a full enterprise replacement, especially in network-effect businesses like Salesforce or Twitter-like platforms. Enterprise adoption will likely be slower at large companies, so near-term disruption may first appear in smaller firms delaying software purchases or changing workflows. Hard assets may look safer because they are AI-resistant and in some cases AI-beneficial, but they are not immune to second-order effects from technological change. Investors must constantly update their priors; persistent bearishness or bullishness without revision becomes intellectually lazy and potentially costly. Successful investing requires combining confidence to act with humility to change course when evidence shifts.

Data Points: Podcast length: ~30 minutes - Walker repeatedly describes the episode as a monthy ramble lasting about half an hour. Date recorded: February 12, 2026 - He states the episode is being recorded on Thursday, February 12th. AI model progress timeline: ~3 years - He contrasts early AI video output with current outputs, saying the improvement from horror-like clips to Hollywood-quality is visible over about three years. SaaS stock moves: down 10% on some days - He cites sector-wide moves where disparate SaaS names were falling by around 10% in a single session. Office stocks move: down 5% to 10% - He notes office-related names were hit sharply as investors feared empty buildings if AI reduced labor needs. LTL trucking move: down ~20% - He references RXO and peers selling off after an AI-related paper hit the group. Historic panic examples: banks in early 2023; busted biotechs in early 2025 - He compares the current AI/SaaS panic to prior dislocations that he felt offered strong opportunities. Personal language learning example: 10 words of Polish - He jokes that years of Duolingo led to very limited real fluency, using it as a consumer SaaS example. Employee survey idea: 500 employees vs 25 employees - He suggests comparing AI/SaaS adoption patterns in companies at different stages of growth.

Pivotal Quotes: "I view a panic as somebody shouting fire in a crowded theater. And I want to be running in when everyone’s running out." — Andrew Walker: He explains his instinct to buy into market panics rather than avoid them. "AI is improving exponentially, and humans are really, really bad at dealing with exponential improvement." — Andrew Walker: Core thesis for why the market may be underestimating disruption risk. "Being an investor is a weird, weird job because it requires a level of arrogance and humility that you almost need to balance on a knife edge." — Andrew Walker: He closes with his philosophy on conviction, risk-taking, and intellectual flexibility.

Implications: Listeners should expect more cross-sector repricing as AI adoption accelerates. The episode suggests opportunities in panic-driven dislocations, but also warns that software-heavy businesses may face real structural risk and that investors must keep updating views quickly.