Field guide / primary domain
36
sources in this field
Updated July 16, 2026
Current thesis
The shortest path to orientation.
AI-accelerated learning operates through structured loops. NotebookLM enables radically compressed cycles via prompt sequences; Socratic prompting (asking rather than directing) improves output by forcing deeper reasoning. Domain-specific prompt libraries for market research, consulting, and competitive intelligence unlock productivity; AI now generates consulting-grade deliverables (McKinsey-style slides with data visualizations). Open-source GitHub repos are the primary educational institution for AI practitioners, with stars as quality filters. The newest frontier is the self-improving knowledge base: a weekend setup (folder structure + schema) becomes a compounding asset when query answers feed back into the corpus. Monthly health checks flag contradictions and gaps; claude-smart converts repeated mistakes into reusable rules. Running AI prep prompts before 1:1s (5-minute routine) converts "flying blind" into structured conversation. Karpathy-style git wikis can grow to 2.3GB+ before forcing migration to SQLite—plan early. NotebookLM podcasts paired with .md files create dual-format artifacts. GitHub repos represent community-validated curriculum; high stars signal better quality than credentials. Dramatic improvements in coding, math, and research stem from verifiable rewards (unit tests) and concentrated B2B value. Refero's 2,000 DESIGN.md files address why AI agents make ugly UIs—exposure-as-training works better than fine-tuning. LLMs prefer Markdown; convert files before querying for better extraction and token efficiency.
Evidence board
Claims worth carrying forward01OpenAI Forward Deployed Engineers earn $280K compensation. The role is not evaluated through LeetCode-style interviews, suggesting the assessment tests practical AI deployment and customer problem-solving skills rather than algorithmic coding.
02The existence of a 'Forward Deployed Engineer' role at OpenAI—distinct from core engineering—signals that AI companies are building dedicated customer-facing technical deployment teams as a separate high-paying career track.
03A 5-minute AI prep routine before 1:1s—running 5 structured prompts—replaces ad-hoc agenda glancing and produces noticeably sharper conversations. The pattern: prepare prompts in advance, run them consistently before every session, not just high-stakes ones.
04The post references 5 specific prompts run before every 1:1 but does not disclose them in this excerpt—the substantive content (the actual prompts) is missing. The value claim (5 min, consistent improvement) is present; the reusable artifact is not.
05River's core design constraint: it only works in public Slack channels and declines DMs. Every conversation is searchable company-wide. This single architectural decision—not the model—is what transforms an AI coding tool into an institutional learning system.
06River's PR merge rate improved from 36% to 77% over two months with no model change or retraining. The gain came entirely from employees watching River work, identifying failure points, and writing corrections into River's skills and memory. Collective human feedback beat model switching.
07River opened 1,870 PRs in one week in Shopify's main monorepo; ~1 in 8 merged PRs last week were River-authored. 5,938 employees worked with River across 4,450 Slack channels in 30 days. These are the scale numbers for a mature enterprise coding agent deployment.
08Lütke's thesis: the risk of AI isn't that it does the work—it's that it does the work in private windows and nobody else learns. When agent interactions are public, prompt patterns, investigation templates, and domain skills spread laterally without any training program or manager.
Adjacent fields
Latest evidence
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.
Khairallah AL-Awady
A detailed 12-month curriculum for breaking into AI engineering without a CS degree, structured around six phases: Python fundamentals, LLM API mastery, RAG systems, agents, evaluation/deployment, and job search. The core argument is that a portfolio of three shipped projects carries more weight than credentials for most AI engineering roles, with salaries ranging from ~$120K entry to $200K+ with experience.
Nico Bailon
Nico Bailon built the open-source Claude/Codex agent skill 'grill-for-unknowns,' merging Matt Pocock's grilling/grill-with-docs skills with Thariq's four-quadrant 'Finding Your Unknowns' framework (known knowns, known unknowns, unknown knowns, unknown unknowns). The skill and its source article detail concrete techniques—blindspot passes, brainstorm/prototype cycles, interviews, source-code references, implementation notes, and post-implementation quizzes—for reaching shared understanding with an AI coding agent before and during implementation, producing durable artifacts (ADRs, launch packets, CONTEXT.md) instead of leaving alignment buried in chat.
Boris Cherny
> Something I have been thinking about: in the past, the best engineers I knew spent a lot of time automating their work in various ways. Better vim/emacs automations, writing lint rules to catch repeat code issues, building up a suite of e2e tests so they don't need to smoke test the app manually.
Matt Van Horn
> /last7days of Fable 5 and GPT-5.6: What Thousands of Upvotes Say Actually Works > > For the first time, the two newest frontier models are live at the same time, and the window is one week old. Call it the /last7days. I pointed [/last30days](https://x.com/slashlast30days) at [Claude Fable 5](https
Samantha Trimble
> another codex hack: for internal facing comms, make it read the last 3 weeks of slacks from someone at your company you trust and admire. ask it to understand what he/she pushes back on, what piques their interest, what they ignore, how they communicate with their team, etc. create a skill that yo