Odd Lots: How Substack Creators Are Thinking About AI (Podcast)

The article introduces a live panel discussion held in New York on May 28 with Substack writers and a journalist discussing how to cover an unusually pressured market environment. The only substantive market theme mentioned is the pervasive influence of AI across industries and media. No company-specific, macroeconomic, or price-moving information is provided.

Analysis

The important read-through is not "AI is everywhere," but that information intermediaries are being re-priced as distribution and interpretation layers get commoditized. That shifts value away from generic market commentary toward people and firms that can supply proprietary data, workflow integration, or truly differentiated POV; everything else becomes easier to replicate by AI and harder to charge for. In media, that usually means a wider dispersion of outcomes: premium niches with tight audience loyalty can still gain share, while broad general-interest coverage faces margin pressure and audience churn.

Second-order effects likely show up first in labor, not headlines. If AI reduces the cost of producing baseline analysis, the marginal benefit of large editorial staffs and junior research teams falls, which should pressure lower-end content businesses and benefit smaller, high-trust operators with low fixed cost structures. Over 6-18 months, expect more bundling and consolidation around brands that own audience relationships directly; the middle layer of ad-supported content is the most exposed.

The contrarian angle is that the market may be overestimating how quickly AI erodes trust. In periods of elevated uncertainty, people often pay more for human judgment, especially when the cost of being wrong is high, so the best franchises may retain pricing power even as volumes fragment. The real risk is not content death, but a slower, uglier transition where monetization improves for the top decile while the bottom 90% gets de facto zeroed out.

Key Decisions for Investors

• Favor long-quality / short-mediocrity in media: long MSFT or GOOGL as AI infrastructure beneficiaries, short a basket of low-differentiation media/ad-dependent names if weakness creates a rally to fade; time horizon 6-12 months, thesis is margin compression and distribution disintermediation.
• Build a pair trade: long premium, direct-to-consumer research/content platforms with strong subscription economics versus short broadly diversified legacy media exposed to ad cycles; aim for 12-18 months and look for 15-25% relative outperformance as AI lowers switching costs.
• Use options to express the view that AI hype is not instantly monetized in media: buy 6-9 month put spreads on weak content names after any momentum-driven bounce; risk/reward improves if management teams begin guiding to softer renewal or ad trends.
• On the long side, own picks-and-shovels rather than narrative winners: semis/cloud/platforms tied to AI adoption have clearer spend visibility than media companies debating disruption; prefer names with operating leverage and recurring demand over story stocks.
• Monitor for consolidation catalysts: if a major media brand announces AI-assisted cost cuts or asset sales, the trade becomes a relative-value long the acquirer/strategic buyer and short the stand-alone operator; that setup can work over 3-6 months.