Enterprise AIStratechery

Mythos, Muse, and the Opportunity Cost of Compute

Read original

Why I picked this

brilliant read

ai-policymarket-consolidationregulatory-impact

The era of Aggregation Theory is behind us, and AI is again making technology expensive. This relation of increased cost from increased consumption is anti-internet era thinking.

Key takeaways

  • Reasoning models (o1) fundamentally break Aggregation Theory by reintroducing marginal costs - compute scales with usage, unlike internet-era products
  • Hyperscalers' business models were built on zero marginal cost assumption; AI inference costs challenge this foundation requiring new economic models
  • The 2010s internet era may be viewed as anomalous 'naive time' - technology returning to capital-intensive, high-marginal-cost paradigm of pre-internet era

Why this matters for operators: Companies building AI products need to rethink unit economics and business models fundamentally different from internet-era playbooks

I cover AI×GTM intelligence like this every Wednesday.

Get STEEPWORKS Weekly

More picks

Human-AI Intersectionr/artificial

The Young Are Being Battered by AI as Hiring Shifts to Older Workers

  • Junior role elimination accelerating (43% of CEOs planning cuts vs 17% last year) as AI automation targets entry-level tasks, creating structural unemployment for early-career workers
  • AI ROI confidence declining sharply—only 27% of CEOs report meeting expectations (down from 38%), yet 74% are still freezing/reducing headcount based on automation assumptions
  • Hiring shift favors mid-level experience (30% vs 10% last year) as companies seek workers who can manage AI tools rather than perform tasks AI might automate—creating experience paradox for new graduates
ai-policymarket-consolidationback-to-basics-gtm
GTM OpsSaaStr — Jason Lemkin

Dropbox Hit $1B Faster Than Any B2B Company Ever. But Now, It’s The End of an Era

  • Dropbox achieved the fastest path to $1B ARR in B2B history with near-zero burn through perfected PLG, but revenue declined -1% in 2025 as file sync commoditized into free features from Google/Microsoft
  • The deceleration pattern is brutal: from 40% growth at $1B (2016) to 8% at $2B (2022-23) to negative growth at $2.5B (2025), showing how even perfect execution can't overcome category commoditization
  • Multiple second-act attempts (HelloSign, DocSend, FormSwift, Dash AI) failed to reignite growth, illustrating the challenge of expanding beyond a wedge product once the core becomes a feature not a product
plg-to-salesmarket-consolidationback-to-basics-gtm

This analysis was produced using the STEEPWORKS system — the same agents, skills, and knowledge architecture available in the GrowthOS package.