Field guide / primary domain
41
sources in this field
Updated August 24, 2026
Current thesis
The shortest path to orientation.
This synthesis records claims and practices from the cited sources; reported outcomes and product capabilities have not been independently verified.
AI-accelerated learning operates through structured loops and community-validated curricula. NotebookLM enables compressed cycles via prompt sequences; Socratic questioning improves output by forcing deeper reasoning. Domain-specific prompt libraries unlock consulting-grade deliverables. High-star GitHub repositories represent community-validated curriculum better than credentials. The Pyramid Principle (answer-first, MECE points, data proof) structures communication with measurable cognitive effects. SCQA framing and an ~10-minute pre-send workflow provide repeatable checks. Refero's 2,000 DESIGN.md files show exposure-as-training outperforms fine-tuning for UI quality. Ian Vanagas distinguishes sharply between 'writing with AI' (using it for research while authoring prose) and 'using AI to write' (delegating final text generation), arguing the latter leaves 'skeletons of slop'. His research stack—Exa, Hacker News, RFC repos, Semble—mirrors manual sourcing. He avoids AI summaries because compression loses unique ideas; he prefers quotes or source skimming. AI struggles as a tightening editor, reflecting back framing rather than cutting prose effectively. LLMs prefer Markdown; converting files before querying improves extraction and token efficiency.
Evidence board
Claims worth carrying forward01Ian Vanagas draws a hard line between 'writing with AI' (using it for research, questions, and feedback while writing all prose himself) and 'using AI to write' (having it generate final text)—arguing the latter leaves 'skeletons of slop' even after heavy editing.
02He built a Claude 'researcher skill' to find real, quotable, sourced examples because unconstrained prompts caused hallucinated plausible-sounding examples; explicitly demanding sources and quotes keeps the model honest.
03His research tool stack for sourcing high-quality (non-SEO-gamed) examples: Exa (agent-oriented search), Hacker News (target-audience signal), local PostHog repos/RFCs, PostHog Slack, and Semble (a link-network discovery tool)—mirroring the sources he'd use manually.
04Revealed preference: he never reads AI-generated summaries because compression loses the interesting/unique ideas he's actually looking for; he gets more from a couple of good quotes or skimming the source himself.
05AI is a poor editor for tightening prose: it reflects back whatever framing you give it (e.g., it recommended shortening an already-short intro on nearly every review) and struggles specifically with cutting/rewriting tighter, since it's better at adding than removing.
06He moved from Notion to Obsidian specifically to use Claude Code and link code context with writing, letting AI act as 'advanced in-document search' across thousands of words of notes/drafts instead of requiring an elaborate tagging/backlinking system.
07Central unresolved question (from his linked prior post): despite LLMs making developers ~55% faster and enabling viral 'built in 3 hours' coding projects, no equivalent surge of viral 'banger' blog posts has appeared—possibly because writing lacks code's reusable structure ('writing is sand, code is Lego') and writers haven't yet found productized AI workflows the way developers have.
08Case study cited as a research example: PostHog's Wizard agent cost $6.67/run, with a trivial 'conclude' step eating $1.47 due to ~140K tokens of carried context; splitting into fresh query() calls cut input tokens 89% but raised total cost, because Anthropic cache writes cost 12x more than cache reads—so naive context-clearing can backfire economically.
Adjacent fields
Latest evidence
Ian Vanagas
Ian Vanagas distinguishes 'writing with AI' (using AI for research, gap-finding, and fact-checking while keeping prose fully human) from 'using AI to write' (having it draft prose). He details a research skill stack for sourcing real examples, explains why AI is weak at summarization and editing/conciseness, and links this to a broader thesis on why AI hasn't produced 10x more 'banger' blog posts the way it has for code.
Forward Future Brian
A collection of high-effort /goal prompt templates designed to consume leftover Codex quota before a reset, spanning software cleanup, performance auditing, personal life planning, and skill-building.
Anish Jaitwar
A tweet references a 38-minute Sam Altman talk to Stanford students claiming ChatGPT usage no longer requires writing prompts, with the poster promising a follow-up guide. The post itself contains no substantive guide content, only a promotional teaser and quote.
Nav Toor
Barbara Minto's 1985 Pyramid Principle (answer-first, MECE-structured writing) is operationalized as a 4-prompt Claude workflow: rewrite messy drafts into pyramid form, generate an SCQA opening, audit supporting points for MECE, and stress-test the draft as a busy CEO would read it.
Peter Yang
Peter Yang previews an interview with Jason, an OpenAI DevEx engineer who authored the official Codex usage guidebook, covering a personal Codex system spanning cross-channel assistant tasks, session-to-skill conversion, and custom learning tools.
PULKIT WALIA
OpenAI pays $280K for Forward Deployed Engineers (FDEs). The hiring process is explicitly not LeetCode-style, implying a different interview loop focused on practical deployment and customer-facing AI work. No further detail on the actual loop is provided in the post content.
Dave Kline
Dave Kline describes a shift from ad-hoc 1:1 meetings to a structured 5-minute AI prep routine using five specific prompts, claiming it materially improves meeting quality and outcome clarity.
tobi lutke
Tobi Lütke describes River, Shopify's internal AI coding agent living in Slack, designed with one hard constraint: it only works in public channels, not DMs. This forces all AI-assisted work into the open, creating a 'Lehrwerkstatt' (teaching workshop) where knowledge spreads osmotically. In 30 days, 5,938 employees used River across 4,450 channels; it opened 1,870 PRs in one week (~1 in 8 merged PRs). Merge rate climbed from 36% to 77% over two months purely from collective human feedback, not model changes.
vas
A practitioner guide to the Forward Deployed Engineer (FDE) role in Applied AI — covering why AI companies need FDEs, the three-phase job (Audit, Evals, Deployment), decision rules for when to use agents vs. code vs. manual work, and a concrete 30-day learning roadmap for consultants, PMs, and software engineers entering the field.
Thariq
An Anthropic employee shares that colleagues use a specific practice (via a linked resource from Suzanne) to stay current with Claude's capabilities and understand ongoing work. The linked content is not available in the bookmark, leaving the actual method unresolved.