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

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

by The New York Times

1h 0m•February 13, 2026
Listen to Episode

Overview of Hard Fork (The New York Times)

This episode of the Hard Fork podcast (hosts Kevin Roose and Casey Newton) examines recent signs of an AI-driven inflection point across tech, business and culture. Segments cover: a viral essay arguing “something big is happening” in AI and its implications for software and labor; how AI is reshaping the romance-novel market (guest Alexandra Alter); and a short “One Good Thing” segment highlighting useful or hopeful tech developments.

Episode structure / main segments

  • Intro + market / policy context: why Washington and investors are suddenly freaked out about AI.
  • Deep dive on the viral essay “Something Big Is Happening” (by Matt Schumer) — implications for software engineering and business models.
  • Interview with Alexandra Alter on AI-written romance novels and the self-publishing ecosystem.
  • One Good Thing: two positive/curious AI-adjacent items (Spotify’s Prompted Playlists; Google’s Perch 2.0 bioacoustics model).
  • Disclosures: hosts’ institutional ties (NYT suing OpenAI/Microsoft/Perplexity; partner at Anthropic).

Key themes and takeaways

  • Momentum & perception shift

    • AI capabilities (especially agentic coding tools) are being perceived as entering an inflection/takeoff phase — not just incremental improvements.
    • This perception is driving political attention, investor reactions, and workplace anxiety.

  • SaaSpocalypse / investor reaction

    • Recent sell-offs hit many enterprise SaaS stocks (Workday, Salesforce, Shopify, Adobe, Microsoft, etc.), driven by fears AI will materially weaken incumbent business models.
    • Two explanations: (1) technology will let customers “vibe-code” their own solutions, or (2) AI enables new business models (outcome-based pricing, fewer per-seat licenses).
    • Practical implication: seat-based pricing faces disruption; expect more outcome- or value-based pricing experiments.

  • Automation of software engineering

    • Engineers report agentic tools automating large portions of coding workflows; some estimate software engineering is now “~90% automated” (still requires human oversight).
    • Evidence cited: CloudCode/GitHub commit share (~4% currently; projection to >20% by end of 2026 if trends continue) — interpreted as an exponential signal.
    • Agents are “relentless” (continuous iteration, no fatigue), which changes productivity dynamics.

  • Security, compliance, and handholding

    • Automation will be buggy/insecure initially but enterprises will pay for compliance-focused, hand‑held deployments (e.g., Anthropic/Goldman collaboration).
    • Sensitive domains will adopt more guarded approaches; startups offering security/compliance become valuable.

  • Political and social implications

    • Rising public anxiety suggests a need for policy discussion (unemployment risk, data and antitrust questions).
    • Expect populist, political pushback (examples: proposed moratoriums, regulatory scrutiny).

Notable quotes & insights

  • Viral essay summary: “Something big is happening” — AI tools are starting to automate the technical parts of skilled jobs, changing career risk calculations.
  • Engineer observation: agents “never get tired” — their persistence is altering work processes and expectations.
  • On business-model disruption: “You’re going to see more companies experiment with outcome-based pricing.”
  • Cautionary note: Labs have incentives to hype models’ self-improvement; treat claims (e.g., models creating themselves) with skepticism.

Romance novels: Alexandra Alter (summary)

  • What’s happening

    • Self-published romance authors are using AI tools (PseudoWrite, NovelAI, general LLMs like Claude/ChatGPT/Gemini) to massively speed output. Example: one author published >200 titles in a year and reportedly earned six figures.
    • Tools are combined and fine-tuned: some models are better at prose, others at erotic content; authors create prompts, blocklists and editorial workflows.

  • Quality, craft & reader experience

    • AI writes competent genre-structured plots and stock beats (enemies-to-lovers, slow burn, forced proximity), but often fails at sustained emotional nuance and believable pacing.
    • Authors must heavily prompt and edit (e.g., instruct the model to “slow down” or avoid certain cliché phrases).
    • Knowing a book is AI-written can change reader engagement; disclosure is rare and contentious.

  • Industry implications

    • Self-publishing is a major feeder for traditional publishing; publishers worry about originality, copyright (AI-generated text’s copyright status), and contract guarantees.
    • Some authors reframe their role as “director/creator” rather than sole author.
    • The tension: lower-cost, high-volume AI content vs. author-brand, audience relationships and literary craft.

  • Memorable anecdote: the recurring AI-generated phrase “like a ragged prayer” appeared across multiple AI-written romance works — an example of AI idiosyncrasy and dataset leakage.

One Good Thing (short picks)

  • Casey’s pick: Spotify Prompted Playlists (Premium feature)

    • Creates custom playlists from natural-language prompts; can use your listening history to craft targeted, refreshing playlists (e.g., songs you’ve played >20 times but not in 2 months).
    • Offers creative uses (opposites of your taste, country-specific hits, thematic prompts), auto-updates, leverages LLM-like world knowledge.

  • Kevin’s pick: Google Perch 2.0 (bioacoustics foundation model)

    • A model trained on birdsong that surprisingly transfers to classifying underwater sounds (whales, dolphins) — demonstrating transfer learning across animal acoustic domains.
    • Practical value: better tools for marine biologists to categorize and detect underwater audio; not yet “translation” but useful classification/monitoring.

Practical recommendations (what listeners should do)

  • For workers

    • Get familiar with modern AI tools relevant to your work (especially if you’re in white‑collar, software, legal, finance, publishing).
    • Financially prepare for disruption (reduce risky debt; consider upskilling/reskilling).
    • Engage politically — pressure lawmakers and employers for plans addressing displacement and worker protections.

  • For businesses

    • Explore outcome-based pricing and agent-enabled internal workflows; evaluate where AI reduces per-seat needs.
    • Invest in security/compliance or partner with firms that can offer auditability and guarantees.
    • Consider the human-in-the-loop model: AI to boost productivity plus human oversight for quality/control.

  • For readers & consumers

    • Be aware AI use isn’t always disclosed (in publishing and other creative industries); your expectations might affect enjoyment.
    • Try new tools (e.g., Spotify Prompted Playlists) and evaluate usefulness first-hand.

People & references mentioned

  • Hosts: Kevin Roose (NYT) and Casey Newton (Platformer).
  • Viral essay: “Something Big is Happening” by Matt Schumer.
  • Guest: Alexandra Alter (New York Times) on romance publishing.
  • Companies/tech referenced: Anthropic (Claude), OpenAI (Codex/GPT), CloudCode, Salesforce, Workday, Goldman Sachs collaboration, PseudoWrite/NovelAI, Spotify, Google (Perch 2.0).
  • Data points: CloudCode ~4% of public GitHub commits now; projection cited toward >20% by 2026 under certain trajectories.

Final note

The episode frames the current moment as a mix of clear technical progress and social/political uncertainty. AI is already reshaping workflows and business models; the immediate tasks are to learn the tools, adapt business strategies, demand policy solutions, and be skeptical about hype while recognizing rapid, real change.