The habit that makes AI worth using

Ernie Svenson · email · 2026-09-11 Reader 🖥️

practical · mixed — Opens with a first-person narrative (the China-solar tweet) and closes with an explicit prescriptive checklist ('A simple habit to build'). The narrative is not decorative — it is the entire evidentiary basis for the argument, which is otherwise asserted rather than demonstrated. The text argues for adopting a specific behavior, not for understanding a phenomenon.

Structure

Key terms

• validate
  A specific, minimal prompting act — pasting a claim into a chatbot and typing an instruction to check it — that returns a verdict plus cited sources, distinct from asking the AI to explain, summarize, or elaborate.
• the real skill AI rewards
  The compound capacity to (a) notice a claim is worth checking and (b) evaluate the resulting answer for its sourcing — treated as more decisive for AI's value than any prompting technique.
• jumping-off point
  The AI's validation answer is a provisional starting position for further inquiry, not a terminal, fully trustworthy verdict — the reader retains responsibility for going deeper if the topic matters.

Propositions

• explicitA single short prompt asking a chatbot to validate a suspicious claim can, in about ten seconds, catch a factual error that would otherwise be accepted as true — in this case, conflating installed solar generating capacity with actual electricity produced.conf: high
• explicitMost people who encounter a surprising or important claim do not check it — they either believe it outright or scroll past without a second thought — because they are too reactive, distracted, or accustomed to treating a single unverified statement as final.conf: high
• explicitManaging one's own attention and knowing how to evaluate information, once slowed down enough to look at it, are real, teachable skills in the age of AI — and both matter more than any prompting technique.conf: high
  extendsThe Best Prompting Advice Came From a Writing Teacher · p7 — Both pieces, from the same author, devalue prompting technique in favor of a more fundamental human skill as the real driver of AI value — clarity of intent in the earlier piece, attention-management plus evaluation-of-information here — extending the same 'it's not the prompt trick' thesis to a new use case (fact-checking claims rather than generating text).
• implicitThe bottleneck to benefiting from AI's fact-checking ability is not technical capability but human initiation — whether a person decides, in the moment a claim surprises them, to actually check it.conf: high
• explicitWhen validating a claim with AI, the answer's cited sources matter more than its bottom-line verdict, and the verdict should be treated as provisional rather than as settling the question.conf: high
• implicitThe chatbot's validation answer is itself only provisionally trustworthy — worth checking for sources rather than accepted on the AI's authority alone.conf: medium

Arguments

Narratives

• The author sees a tweet claiming China now produces more solar power than every U.S. power plant combined; instead of believing or ignoring it, he pastes it into ChatGPT with the instruction "Validate this," and within ten seconds learns the claim conflates installed capacity with actual electricity produced, with sourced figures provided.
  → Grounds the entire piece — it is the sole piece of evidence for the claim that AI-assisted validation is fast, effective, and catches a specific, non-obvious category of error (capacity vs. output), and it motivates the prescriptive habit that follows.

Questions addressed

• ✅ What separates people who get real value from AI from those who don't, specifically with respect to catching bad information?
  Not prompting skill, but the habit of noticing a surprising/important claim and choosing to check it, plus the capacity to evaluate the resulting answer.
• ✅ Concretely, how should a person validate a claim using AI?
  Copy the claim, paste it into a chatbot, ask it to validate the claim, read the answer for its sources rather than just its verdict, and treat that as a jumping-off point for further digging if the topic matters.
• ⭕ How should a reader handle the case where the AI's own validation is wrong, incomplete, or itself a fabrication (hallucinated sources, motivated framing, or an out-of-date model of the facts)?

Gaps

• incompleteThe habit rests entirely on trusting the AI's own validation, but the piece never addresses what happens when that validation is itself wrong — hallucinated, subtly biased by training data, or simply mistaken with a confident tone. The "ten seconds" framing that makes the habit attractive is the same framing that discourages actually clicking through to primary sources when the AI's citations aren't immediately convenient. A habit built to catch bad information needs a second-order check for the case where the checker itself fails silently.
  Read next: Research on LLM-based fact-checking accuracy and calibration; work on adversarial or politically-contested claims designed to survive a single validation pass; the corpus's existing material on hallucination and on the capacity needed to recognize wrong AI output.
• incompleteThe single example generalizes cleanly because it is a quantifiable, well-documented statistic (solar capacity vs. output) with an unambiguous correct answer and clear public sources. The piece extends the habit to "a claim" in general without addressing contested, values-laden, rapidly-evolving, or novel claims where sources themselves disagree, are sparse, or are already AI-generated — cases where a confident ten-second validation could produce false confidence rather than correction.
  Read next: Literature distinguishing fact-checkable empirical claims from contested/interpretive claims; studies of chatbot reliability specifically on breaking-news or actively-disputed topics.
• incompleteThe essay treats attention as the scarce resource the reader must manage, yet its prescription — validate every claim that surprises you or seems important — is itself an added attentional demand with no discussion of how to triage which claims are worth the ten seconds versus which aren't, or how this habit competes with other calls on the same finite attention.
  Read next: Work on attention as a hijackable/depletable resource and on triage heuristics for information verification under time constraints.

Related documents

• The Best Prompting Advice Came From a Writing Teacher
• How Coders Can Survive—and Thrive—in a ChatGPT World
  LLM Reliability Failures Prompting as Communication
• LLMs Reward Expertise
  Capacity to Verify AI Output Prompting as Communication
• The real skill behind good AI prompts is older than computers
  Attention as a Hijackable Resource Prompting as Communication