Hard Fork

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

February 13, 2026

Key Takeaways Copied to clipboard!

  • The stock market is experiencing a sell-off in software-as-a-service (SaaS) companies due to mounting fears that AI agentic capabilities and the ability to 'vibe code' internal tools will disrupt traditional seat-based business models. 
  • The viral essay 'Something Big Is Happening' by Matt Schumer articulates the growing sense among technical professionals that AI is reaching an inflection point, evidenced by accelerating release cycles and the relentless, 24/7 nature of AI agents in coding tasks. 
  • The romance novel industry is seeing writers use AI to dramatically increase output (one writer published over 200 novels in a year), though the resulting content often requires heavy editing to avoid generic tropes and AI 'isms' like the phrase 'ragged prayer'. 
  • Spotify's new 'prompted playlists' feature utilizes world knowledge, similar to an LLM, allowing users to create highly specific playlists based on complex criteria, such as songs whose titles do not appear in the lyrics. 
  • Google's new bioacoustics foundation model, Perch 2.0, demonstrates surprising success in classifying underwater noises (whales, dolphins, orcas) by being trained primarily on birdsong, illustrating the power of transfer learning in general AI models. 
  • The research on Perch 2.0 suggests a generalizable principle in AI where improving a model for one domain (like bird sounds) enhances its performance in related, seemingly different domains (like underwater acoustics). 

Segments

Elon Musk’s AI Moon Factory
Copied to clipboard!
(00:00:34)
  • Key Takeaway: Elon Musk reportedly told XAI employees the company needs a moon factory and a massive catapult to launch AI satellites into space.
  • Summary: Elon Musk proposed building a factory on the moon to construct AI satellites and using a massive catapult for launching them into space. This was framed as a pivot away from previous Mars-focused ambitions. The hosts reacted with humorous disbelief regarding the feasibility and sanity of the plan.
SaaS Apocalypse and Stock Sell-off
Copied to clipboard!
(00:01:27)
  • Key Takeaway: A significant sell-off is occurring in software-as-a-service (SaaS) stocks, driven by investor anxiety over AI’s potential to enable custom software creation, thus eroding the value of established business software providers.
  • Summary: Stocks for companies like Monday.com and Workday saw precipitous drops, reflecting investor concern that AI tools could allow businesses to build their own solutions, bypassing incumbent SaaS providers. The vulnerability is seen less in the technology itself and more in how AI enables new, leaner business models, such as outcome-based pricing replacing seat-based licensing. Companies relying on seat-based enterprise software face significant long-term risk if internal development becomes viable.
Viral Essay on AI Acceleration
Copied to clipboard!
(00:03:15)
  • Key Takeaway: The viral essay ‘Something Big Is Happening’ argues that AI is accelerating rapidly due to recursive self-improvement, suggesting software engineering could be fully automated within a year.
  • Summary: Matt Schumer’s essay explains why technical professionals are alarmed by AI’s rapid progress, particularly agentic coding capabilities. OpenAI claims GPT-5.3 Codex was instrumental in creating itself, suggesting development timelines are shrinking from months to weeks. Engineers note that AI agents are relentless and never tire, contributing to a perceived ’takeoff’ phase in development.
AI’s Impact on White-Collar Work
Copied to clipboard!
(00:08:01)
  • Key Takeaway: The primary threat of AI to established businesses is not necessarily replacing technology wholesale, but enabling small startups to achieve massive output with minimal staff, fundamentally changing business models.
  • Summary: The shift is predicted to move toward outcome-based pricing (e.g., paying per resolved customer service inquiry) rather than per-seat licensing, which endangers current SaaS structures. The hosts suggest that white-collar workers should get familiar with the tools and that political conversations about potential permanent job automation are necessary. The potential for furious backlash when job displacement becomes widespread is also anticipated.
AI in Romance Novel Production
Copied to clipboard!
(00:27:23)
  • Key Takeaway: Romance novelists are using AI tools like Pseudo-Write and Claude to generate hundreds of novels annually, often acting as ‘directors’ rather than authors, though AI struggles with emotional nuance and specific tropes.
  • Summary: Writers are using AI to produce books in a day by providing detailed outlines, subgenre specifications, and even lists of desired sexual kinks or unusual settings to avoid generic outputs. A common AI artifact noted was the phrase ‘whispers her name like a ragged prayer,’ which writers are now actively blocking. Major publishers are concerned about copyright issues since AI-generated work cannot be copyrighted, complicating acquisition strategies.
Spotify Prompted Playlists Feature
Copied to clipboard!
(00:51:06)
  • Key Takeaway: Spotify’s new ‘Prompted Playlists’ feature allows premium subscribers to use natural language prompts, leveraging ‘world knowledge’ beyond listening data to create highly customized, anti-slop music mixes.
  • Summary: This feature functions like a chatbot interface within Spotify, enabling users to request complex criteria, such as songs played frequently but not recently, or even playlists based on abstract concepts like ’eating fish tacos on the beach.’ The VP of Personalization confirmed that the system utilizes the user’s decade-long listening history if available. This tool is considered empowering because it allows users to actively direct discovery rather than passively following algorithmic ‘machine drift.’
Prompted Playlist Creation Demo
Copied to clipboard!
(00:56:44)
  • Key Takeaway: Spotify’s prompted playlists feature uses world knowledge to filter songs based on complex lyrical criteria, such as titles not appearing in the lyrics.
  • Summary: The demonstration involved creating a playlist of songs whose titles are not echoed in the lyrics, using examples like ‘Bohemian Rhapsody’ and ‘Day in the Life.’ This feature relies on world knowledge beyond basic song metadata, similar to an LLM. Users can set these playlists to update automatically on a chosen schedule.
AI Decodes Whale Sounds
Copied to clipboard!
(00:59:15)
  • Key Takeaway: Google’s Perch 2.0 model, trained on birdsong, successfully categorizes whale and dolphin vocalizations through transfer learning.
  • Summary: Google developed a bioacoustics foundation model called Perch 2.0 to interpret underwater audio from various marine mammals. Remarkably, the model was trained on birdsong, yet its general embeddings allowed it to accurately label and categorize underwater noises. This suggests that models generalized for one type of animal sound can effectively analyze related acoustic data from other species.
AI Transfer Learning Principles
Copied to clipboard!
(01:01:48)
  • Key Takeaway: The success of Perch 2.0 illustrates that principles seen in LLMs, where improved performance in one task transfers to related tasks, also apply to animal sound classification.
  • Summary: The researchers humorously titled their paper referencing ‘The Bitter Lesson,’ noting that making a model better at one task (like coding for LLMs, or birdsong classification for Perch 2.0) improves its ability to handle related tasks (like math or underwater acoustics). This project helps scientists classify unknown sounds, though full translation of whale speech remains a separate effort.
Sponsor Read: JIRA Product Discovery
Copied to clipboard!
(01:03:30)
  • Key Takeaway: JIRA Product Discovery centralizes feedback, prioritization, and roadmapping to connect product decisions directly to delivery within the JIRA ecosystem.
  • Summary: Product teams often struggle with scattered evidence and inconsistent frameworks when deciding what software to build. JIRA Product Discovery provides a single location for capturing feedback and building roadmaps that stakeholders can support. Because it is built on JIRA, decisions remain linked to the actual development process.
Sponsor Read: Invesco QQQ ETF
Copied to clipboard!
(01:04:06)
  • Key Takeaway: Invesco QQQ ETF offers investors access to innovation by tracking companies driving technological progress, from connectivity to AI microchips.
  • Summary: Over two decades, innovators have constantly rethought possibilities across technology sectors. The Invesco QQQ ETF allows investors to access this world of innovation through a single investment vehicle. Investors should review objectives, risks, and charges before investing, noting that tech sector investments carry higher volatility.
Sponsor Read: Chrome Enterprise
Copied to clipboard!
(01:04:39)
  • Key Takeaway: Chrome Enterprise provides proactive security controls, data loss prevention, and usage reports to safeguard company data within the browser.
  • Summary: The browser acts as the primary defense against online threats for businesses. Chrome Enterprise allows administrators to enforce policies, such as preventing data printing or pasting outside approved channels. It offers detailed reports on extension usage and automatic protections against phishing and malware.
Podcast Production Credits
Copied to clipboard!
(01:05:13)
  • Key Takeaway: The production team for ‘Hard Fork’ includes Whitney Jones, Rachel Cohn, and Viren Pavich, with original music by several artists.
  • Summary: Whitney Jones and Rachel Cohn produce the show, with editing handled by Viren Pavich and fact-checking by Caitlin Love. Katie McBurran engineered the show, and Jen Poyant serves as the executive producer. Viewers can watch the full episode on YouTube.
Sponsor Read: Raymour & Flanigan
Copied to clipboard!
(01:06:13)
  • Key Takeaway: Raymour & Flanagan is hosting a President’s Day Super Weekend sale offering savings up to 35% across furniture categories.
  • Summary: The sale includes deep discounts on sectionals, sofas, dining sets, and bedroom sets, available through Monday. Customers can also combine select mattress sale pricing with special financing, but only when shopping in stores.