‘Something Big Is Happening’ + A.I. Rocks the Romance Novel Industry + One Good Thing

What It Covers

Hard Fork examines AI's accelerating impact across industries, from software company stock crashes to automated romance novel production. The episode explores why Washington DC is alarmed about AI capabilities, how coding tools like Claude are automating engineering work, and how romance authors now produce 200+ books annually using AI assistance.

Key Questions Answered

• •SaaS Business Model Disruption: Software companies like Salesforce, Workday, and Monday.com experienced significant stock declines as investors recognize AI enables businesses to build custom tools internally rather than purchasing per-seat licenses. Outcome-based pricing models emerge as alternatives, exemplified by Sierra's customer service platform charging per resolved inquiry rather than per user seat, fundamentally threatening traditional enterprise software revenue structures.
• •AI Coding Acceleration Metrics: Claude Code currently authors 4% of all public GitHub commits, with projections reaching 20%+ by 2026. OpenAI's GPT 5.3 Codex represents the first model instrumental in creating itself, enabling recursive self-improvement. Engineers report coding work is approximately 90% automated currently, with full automation predicted within twelve months, driven by AI's relentless 24/7 operation without fatigue or morale decline.
• •Romance Novel Production Pipeline: Author Coral Hart created 21 pen names and published 200+ romance novels in one year using AI, generating six-figure income through volume-based strategy. Effective workflow requires detailed prompting with specific subgenres, blocking overused phrases like "turgid manhood," and instructing models to "slow down" since they rush to scene conclusions. Claude writes elegant prose but fails at banter; ChatGPT frequently refuses requests; Grok accepts any prompt.
• •AI Content Quality Limitations: Romance novels generated by AI lack emotional nuance and character depth despite following genre tropes correctly. Models struggle with pacing—converting enemies-to-lovers arcs into immediate romance within one chapter. Writers must provide extensive guidance including unusual settings (winery fermentation tanks, stalled ski lifts) and detailed inventories of non-standard romantic scenarios to avoid generic bedroom/shower scenes and clichéd language patterns.
• •Political and Labor Market Response: Washington DC policymakers express heightened concern about AI's workforce impact as unemployment remains near record lows. Bernie Sanders proposes legislation for data center moratoriums. The comparison to February 2020 pandemic awareness suggests exponential adoption curves that most people haven't recognized yet. Constituents need to engage lawmakers about job displacement scenarios and demand government responses beyond industry reassurances about job creation.
• •Spotify Prompted Playlists Feature: Premium subscribers in US, Canada, and New Zealand access AI-powered playlist generation using natural language requests and personal listening history. Users can request songs played 20+ times but not recently, opposite-taste recommendations, country-specific hits, or abstract concepts. The system employs "world knowledge" beyond song metadata, enabling complex queries like identifying songs whose titles don't appear in lyrics, with automatic daily updates available.