How to write stuff no one else can

Original Article ↗ Hacker News Discussion ↗

Writers and technologists are examining how to produce truly original work in an era where Google and large language models can surface and remix most public information. Many see an enduring advantage in human context, deep domain knowledge, and “off‑the‑beaten‑path” sources such as obscure books, private conversations, and interviews with overlooked experts, arguing that AI remains largely constrained by existing data. The conversation also touches on practical implications for careers and hiring, suggesting that those who do the legwork to find under-the-radar people and ideas can gain a meaningful edge in crowded fields.

Information advantage in the age of Google and LLMs

  • Several commenters agree the web and search engines dilute “information advantage,” but argue we are still far from fully interconnected.
  • The real bottleneck is search and discovery, not connectivity; useful, actionable people and information are hard to find.
  • Books, paywalled/scientific sources, and direct conversations are seen as key non‑web knowledge bases.
  • There’s concern that people share less openly (e.g., due to commercialization of open source).

Capabilities and limits of LLMs

  • Some view LLMs as advanced retrieval/remix tools; others argue they can apply known concepts to new, hypothetical domains.
  • Example: using AI to design legal systems in a magical world impressed some, but others note this is just “our world, plus magic” and heavily covered in fiction already.
  • Strong debate over “intelligence” and “emergence”:
    • One side: models are just tools, bounded by code and data, fundamentally historical and not truly understanding.
    • Other side: behavior arises from training data, not explicit programming; even simple ML models “learn” from data.
  • Practical complaints: hallucinated APIs/configs, misuse of libraries, and contamination of specialized internal knowledge with wrong public info.
  • Consensus: very useful as a fast, knowledgeable assistant, but unreliable for unsupervised deep or off‑beaten‑path work.

Finding unique sources and writing distinctive pieces

  • Many see the described tactics (digging into lesser‑known employees, LinkedIn deep dives, Spotify/podcast searches) as basically “good journalism,” but still valuable.
  • Some find the focus on “getting ahead of competitors” off‑putting; others note that in crowded newsletter markets, competitive edge is survival, not vanity.
  • Appreciation for showing the tactical, painstaking work behind seemingly simple writing; others dismiss it as trivial or “blogspam.”

Hiring, potential, and reading between the lines

  • Several apply the article’s research tactics to hiring: look past polished profiles and big brands to spot under‑the‑radar talent.
  • Short, inflated C‑titles in college‑age startups are debated:
    • Some see them as a nuanced potential signal if context is understood.
    • Others view early “C‑level” titles as a red flag or immaturity.
  • Bootcamps and LinkedIn optimization are said to dilute old heuristics like “codes in their spare time” or “has GitHub projects.”
  • Split views on heavily curated LinkedIn: for some a red flag; for others a rational, high‑leverage investment in visibility.

Leadership, confrontation, and culture in software

  • One thread argues that a few key hires change company trajectory mainly via leadership, not raw skill.
  • Claims that software lacks real leadership spark pushback: many report strong leaders and plenty of trend‑bucking behavior, especially outside big enterprise.
  • Disagreement over “strong leaders turn agreeableness down to zero”:
    • Critics say “maximum confrontation” is toxic and counterproductive; teams need cooperation.
    • Defenders clarify the point is situational: leaders must sometimes be willing to confront, defy, and protect teams at personal cost.
  • Broad agreement that good leadership combines domain competence, willingness to challenge, and the ability to build and protect teams.

Visibility, exposure, and lifestyle businesses

  • Commenters lament that those “doing the work” often don’t talk about it, leaving the field to master self‑promoters. Some argue exposure is inherently valuable; others question organizing society around it.
  • Debate over “four‑hour workweek”–style businesses:
    • Critics see them as either implying low consumption or relying on exploitation/market failures.
    • Supporters point to small, stable software products that require heavy upfront work but light ongoing maintenance; they distinguish this from over‑priced, easily undercut offerings.