DOMAIN:INNOVATION — THOUGHT_LEADERS¶
OWNER: joshua
UPDATED: 2026-03-24
PURPOSE: define the people, organizations, publications, and events Joshua follows to stay informed on AI, engineering, and technology trends relevant to GE
SCOPE: curated list of high-signal sources — quality over quantity, relevance to GE over general interest
RULE: this is not a social media follow list — every entry must have a clear reason Joshua follows them and how their output maps to GE's operations
LEADERS:AI_COMPANIES_AND_RESEARCHERS¶
ANTHROPIC¶
WHY: GE's primary LLM provider — every Anthropic move directly affects GE's operations
KEY_PEOPLE:
- Dario Amodei (CEO) — strategic direction, safety philosophy, scaling plans, public essays on AI trajectory
- Daniela Amodei (President) — business strategy, enterprise positioning, pricing philosophy
- Amanda Askell (character/personality research) — directly relevant to GE's agent identity system, how to make agents behave consistently
- Chris Olah (mechanistic interpretability) — understanding why models fail, relevant for GE's anti-LLM quality pipeline
- Jan Leike (alignment team, ex-OpenAI) — safety research, alignment techniques that affect model behavior
WHAT_TO_WATCH: blog posts (anthropic.com/news), API changelog, model cards, research papers, Dario's essays
GE_RELEVANCE: model releases, pricing, API changes, safety research that affects agent behavior, constitutional AI patterns (GE has its own constitution)
FOLLOW: anthropic.com/news (RSS), @AnthropicAI (X/Bluesky), Claude API changelog
OPENAI¶
WHY: GE's secondary LLM provider — competitive dynamics with Anthropic affect GE's multi-provider strategy
KEY_PEOPLE:
- Sam Altman (CEO) — strategic direction, product announcements, AGI timeline claims
- Andrej Karpathy (ex-OpenAI, independent) — deep technical insights on LLM training, tokenization, architecture
- Logan Kilpatrick (ex-developer relations) — API strategy, developer experience direction
- Greg Brockman (President, on leave) — technical vision
WHAT_TO_WATCH: openai.com/blog, API changelog, developer forum announcements, Sam Altman blog/X posts
GE_RELEVANCE: model releases, pricing changes, API evolution, competitive positioning vs Anthropic, function calling improvements
FOLLOW: openai.com/blog (RSS), @OpenAI (X), developer forum
GOOGLE_DEEPMIND¶
WHY: GE's tertiary LLM provider, massive context windows, multimodal capabilities
KEY_PEOPLE:
- Demis Hassabis (CEO, DeepMind) — research direction, AGI perspective, strategic vision
- Jeff Dean (Chief Scientist, Google AI) — infrastructure-scale AI, system design
- Oriol Vinyals (VP Research) — sequence-to-sequence, language modeling foundations
WHAT_TO_WATCH: deepmind.google/blog, Google AI blog, Gemini API changelog, Google I/O keynotes
GE_RELEVANCE: Gemini model updates, context window innovations, multimodal capabilities, Google ecosystem integration
FOLLOW: deepmind.google (RSS), Google AI blog (RSS), @GoogleDeepMind (X)
META_AI¶
WHY: largest open-source model provider — Llama models are the baseline for open-source evaluation
KEY_PEOPLE:
- Yann LeCun (Chief AI Scientist) — provocative takes on AI architecture, strong opinions on open-source
- Mark Zuckerberg (CEO) — Meta's AI strategy, open-source commitment (or not)
WHAT_TO_WATCH: ai.meta.com blog, Llama releases, open-source model benchmarks
GE_RELEVANCE: open-source model quality (relevant for cost optimization / self-hosting future), training methodology insights
FOLLOW: ai.meta.com (RSS), @ylecun (X)
LEADERS:AI_ENGINEERING_AND_TOOLING¶
ANDREJ_KARPATHY¶
WHY: arguably the best AI educator alive — explains complex LLM concepts in accessible ways, builds practical tools
WHAT_HE_COVERS: LLM internals (tokenization, attention, training), AI engineering best practices, practical tutorials
GE_RELEVANCE: deep understanding of how LLMs work helps GE optimize agent prompts, understand failure modes, and evaluate new models
NOTABLE_WORK: "Let's build GPT" video series, nanoGPT, Eureka Labs, essays on LLM tokenization pitfalls
FOLLOW: @karpathy (X), YouTube channel, GitHub (karpathy)
SIMON_WILLISON¶
WHY: the most prolific and practical voice on LLM tooling — builds and documents real tools, no hype
WHAT_HE_COVERS: LLM tool building, prompt engineering, data analysis with LLMs, AI ethics, practical applications
GE_RELEVANCE: his tooling insights directly apply to GE's agent infrastructure, his blog is a masterclass in practical AI engineering
NOTABLE_WORK: datasette, llm CLI tool, shot-scraper, extensive blog at simonwillison.net
FOLLOW: simonwillison.net (RSS), @simonw (X/Bluesky/Mastodon)
FREQUENCY: posts almost daily — high volume but consistently high quality
SWYX (Shawn Wang)¶
WHY: coined "AI Engineer" as a discipline, runs AI Engineer Summit conference, deeply connected to the practitioner community
WHAT_HE_COVERS: AI engineering as a discipline, agentic patterns, LLM application architecture, industry trends
GE_RELEVANCE: his framework for thinking about AI engineering maturity applies to GE's agent system, community pulse on what works in production
NOTABLE_WORK: "The Rise of the AI Engineer" essay, Latent Space podcast, AI Engineer Summit conference
FOLLOW: @swyx (X), Latent Space podcast, latent.space blog
NOTE: more strategic/industry-level than Karpathy's technical depth — complementary
HARRISON_CHASE¶
WHY: creator of LangChain/LangGraph — shapes the multi-agent framework discourse
WHAT_HE_COVERS: agent frameworks, state management, tool use patterns, multi-agent orchestration
GE_RELEVANCE: LangGraph patterns inform GE's orchestration design even though GE does not use LangGraph — the concepts transfer
NOTABLE_WORK: LangChain (most popular LLM framework), LangGraph, LangSmith (observability)
FOLLOW: @hwchase17 (X), LangChain blog, GitHub
CAVEAT: naturally biased toward LangChain ecosystem — filter for patterns, not product recommendations
LEADERS:WEB_DEVELOPMENT_AND_INFRASTRUCTURE¶
GUILLERMO_RAUCH¶
WHY: CEO of Vercel, creator of Next.js — GE's frontend stack decisions are influenced by Vercel's direction
WHAT_HE_COVERS: Next.js roadmap, React Server Components strategy, edge computing, AI integration (v0)
GE_RELEVANCE: Next.js is GE's frontend framework — Guillermo's vision shapes the tool GE's agents use daily
NOTABLE_WORK: Next.js, Vercel, v0 (AI UI generation), Socket.io
FOLLOW: @rauchg (X), vercel.com/blog, Next.js blog
MATT_POCOCK¶
WHY: the TypeScript educator — makes complex type system concepts accessible, widely followed by TS developers
WHAT_HE_COVERS: TypeScript advanced patterns, generic types, type inference, Zod integration, practical TS tips
GE_RELEVANCE: GE's entire codebase is TypeScript strict mode — Matt's patterns directly improve agent-generated code quality
NOTABLE_WORK: Total TypeScript course, ts-reset library, TypeScript tips on X
FOLLOW: @maaborern (X), totaltypescript.com
NOTE: his content is excellent for updating agent system prompts with TypeScript best practices
KENT_C_DODDS¶
WHY: influential React educator, testing advocate, practical full-stack patterns
WHAT_HE_COVERS: React patterns, testing strategies, Remix/React Router, full-stack TypeScript
GE_RELEVANCE: testing philosophy aligns with GE's quality pipeline, React patterns inform frontend agent knowledge
NOTABLE_WORK: Testing Library (React Testing Library), Epic React, epicweb.dev
FOLLOW: @kentcdodds (X/Bluesky), kentcdodds.com
TANNER_LINSLEY¶
WHY: creator of TanStack (React Query, React Table, TanStack Router) — widely used in React ecosystem
WHAT_HE_COVERS: data fetching patterns, state management, routing, React ecosystem tooling
GE_RELEVANCE: TanStack libraries are common in client projects, React Query patterns are standard for data fetching
NOTABLE_WORK: TanStack Query, TanStack Table, TanStack Router, TanStack Form
FOLLOW: @tanaborern (X), tanstack.com
THEO_BROWNE (t3dotgg)¶
WHY: prolific content creator on TypeScript full-stack development, strong opinions, large following
WHAT_HE_COVERS: TypeScript ecosystem, tRPC, Next.js patterns, developer tooling, AI in development
GE_RELEVANCE: community pulse on TypeScript ecosystem, early signal on tooling trends
NOTABLE_WORK: T3 Stack (Next.js + tRPC + Tailwind + Prisma), Uploadthing, Ping.gg
FOLLOW: @t3dotgg (X/YouTube)
CAVEAT: strong opinions, not always aligned with GE's choices (he favors tRPC, GE uses Hono RPC) — useful for perspective
LEADERS:DEVOPS_AND_INFRASTRUCTURE¶
KELSEY_HIGHTOWER¶
WHY: legendary Kubernetes advocate, clear communicator, practical infrastructure patterns
WHAT_HE_COVERS: Kubernetes, cloud-native patterns, platform engineering, developer experience
GE_RELEVANCE: k3s is GE's orchestration layer — Kelsey's patterns inform infrastructure decisions
NOTABLE_WORK: "Kubernetes The Hard Way", Google Cloud advocacy, conference talks
FOLLOW: @kelseyhightower (X), conference talk recordings
NOTE: retired from Google but still influential in the community
MITCHELL_HASHIMOTO¶
WHY: co-founder of HashiCorp (Terraform, Vault) — GE uses both products
WHAT_HE_COVERS: infrastructure-as-code philosophy, secrets management, distributed systems
GE_RELEVANCE: Vault is GE's secrets management, Terraform is GE's IaC — Mitchell's vision shapes these tools
NOTABLE_WORK: HashiCorp (Terraform, Vault, Consul, Nomad), Ghostty terminal
FOLLOW: @mitchellh (X), hashicorp.com/blog
NOTE: HashiCorp acquired by IBM in 2024 — watch for strategic changes
LEADERS:ORGANIZATIONS_AND_LABS¶
THOUGHTWORKS¶
WHY: invented the Technology Radar format that GE's radar is based on — quarterly radar is essential reading
WHAT_THEY_PUBLISH: Technology Radar (quarterly), technical books (Building Microservices, etc.), engineering blog
GE_RELEVANCE: radar methodology, engineering culture patterns, technology evaluation frameworks
FOLLOW: thoughtworks.com/radar (quarterly), thoughtworks.com/insights
VERCEL¶
WHY: maintains Next.js, shadcn/ui, v0 — three pillars of GE's frontend stack
WHAT_THEY_PUBLISH: Next.js releases, framework blog posts, AI tool updates (v0)
GE_RELEVANCE: every Next.js release directly affects GE's frontend agents
FOLLOW: vercel.com/blog, nextjs.org/blog
DRIZZLE_TEAM¶
WHY: maintains Drizzle ORM — GE's database access layer
WHAT_THEY_PUBLISH: release notes, migration guides, new feature announcements
GE_RELEVANCE: every Drizzle update directly affects GE's backend agents and data layer
FOLLOW: orm.drizzle.team/blog, GitHub releases, @draborenzzleorm (X)
LEADERS:NEWSLETTERS¶
TIER_1 (must-read, weekly)¶
TLDR_AI: daily AI news digest — broad coverage, good signal-to-noise for breaking announcements
THE_BATCH (Andrew Ng): weekly AI newsletter — research-oriented, good for understanding capability trajectories
LATENT_SPACE (Swyx): AI engineering focused — practical, production-oriented, interviews with practitioners
SIMON_WILLISON_BLOG: technically a blog but functions as a newsletter — daily LLM tooling insights
TIER_2 (useful, weekly)¶
BYTES (ui.dev): JavaScript/TypeScript ecosystem — entertaining, covers relevant framework updates
NODE_WEEKLY: Node.js ecosystem updates — relevant for GE's runtime
REACT_STATUS: React ecosystem — relevant for GE's frontend stack
JAVASCRIPT_WEEKLY: broader JS coverage — catches things the specialized newsletters miss
TIER_3 (strategic, monthly)¶
THOUGHTWORKS_RADAR: quarterly technology radar — strategic technology positioning
PRAGMATIC_ENGINEER (Gergely Orosz): engineering management and industry — strategic context
STRATECHERY (Ben Thompson): tech industry strategy — macro context for AI industry dynamics
LEADERS:PODCASTS¶
TIER_1 (regular listening)¶
LATENT_SPACE: Swyx + Alessio Fanelli — AI engineering interviews, technical depth, practitioner focus
FREQUENCY: weekly
GE_RELEVANCE: production AI engineering patterns, framework comparisons, industry insider perspectives
PRACTICAL_AI: Chris Benson + Daniel Whitenack — practical AI applications, accessible explanations
FREQUENCY: weekly
GE_RELEVANCE: grounded perspective on what actually works in production AI
TIER_2 (selective listening)¶
LEX_FRIDMAN_PODCAST: long-form interviews with AI researchers — deep dives when relevant guests appear
FREQUENCY: weekly (listen selectively — only when guest is relevant)
GE_RELEVANCE: deep technical interviews with Anthropic/OpenAI/Google researchers
SYNTAX_FM: Wes Bos + Scott Tolinski — web development — covers TypeScript, React, Next.js ecosystem
FREQUENCY: weekly
GE_RELEVANCE: web development trends, tooling updates, developer sentiment
CHANGELOG: open-source and developer tools — broad coverage of the tools GE uses
FREQUENCY: weekly
GE_RELEVANCE: tooling ecosystem awareness, project sustainability signals
LEADERS:CONFERENCES¶
TIER_1 (essential — attend or watch recordings)¶
AI_ENGINEER_SUMMIT:
WHAT: dedicated to AI engineering as a discipline — the most relevant conference for GE's domain
WHEN: annually (usually San Francisco)
WHY: cutting-edge agentic patterns, multi-agent architectures, production AI deployment, tool launches
GE_RELEVANCE: direct competitor intelligence, pattern discovery, networking
WATCH_FOR: keynotes, workshop materials, demo showcases
NEURIPS (Conference on Neural Information Processing Systems):
WHAT: premier ML/AI research conference
WHEN: annually (December)
WHY: frontier research that becomes production technology 12-18 months later
GE_RELEVANCE: early signal on model capabilities, training techniques, evaluation methods
WATCH_FOR: best papers, workshop papers on agents/tool use/code generation
TIER_2 (relevant — watch selected talks)¶
NEXT_JS_CONF (Vercel):
WHAT: Next.js ecosystem conference
WHEN: annually
WHY: Next.js is GE's frontend framework — roadmap announcements, new features, best practices
GE_RELEVANCE: framework direction, breaking changes preview, ecosystem updates
WATCH_FOR: keynote (roadmap), migration guides, performance talks
REACT_CONF:
WHAT: official React conference (Meta)
WHEN: annually
WHY: React is GE's UI library — core team direction, new patterns
GE_RELEVANCE: React evolution, new APIs, deprecations, ecosystem direction
WATCH_FOR: core team talks, React Compiler updates, Server Components evolution
KUBECON (CNCF):
WHAT: Kubernetes and cloud-native ecosystem conference
WHEN: twice annually (EU + NA)
WHY: k3s is GE's orchestration — Kubernetes roadmap and ecosystem
GE_RELEVANCE: Kubernetes direction, security updates, operational patterns
WATCH_FOR: k3s/Rancher talks, security talks, edge computing talks
TIER_3 (awareness — skim summaries)¶
WWDC (Apple):
WHAT: Apple's developer conference
WHEN: annually (June)
WHY: iOS development capability (hilde agent), Apple Silicon impact on local AI inference
GE_RELEVANCE: iOS SDK changes, on-device AI capabilities, App Store policy updates
WATCH_FOR: iOS SDK breaking changes, AI/ML framework updates, privacy changes
GOOGLE_I_O:
WHAT: Google's developer conference
WHEN: annually (May)
WHY: Gemini updates, Android, web platform features, Google Cloud
GE_RELEVANCE: Gemini model announcements, web platform features, Firebase/Cloud updates
WATCH_FOR: Gemini keynote, web platform talks, AI tool launches
LEADERS:GITHUB_REPOS_TO_WATCH¶
CRITICAL (watch releases)¶
- anthropics/claude-code — GE's execution substrate
- drizzle-team/drizzle-orm — GE's ORM
- honojs/hono — GE's web framework
- vercel/next.js — GE's frontend framework
- tailwindlabs/tailwindcss — GE's CSS framework
- shadcn-ui/ui — GE's component library
IMPORTANT (watch releases)¶
- vitest-dev/vitest — GE's test runner
- microsoft/playwright — GE's e2e testing
- stryker-mutator/stryker-js — GE's mutation testing
- hashicorp/vault — GE's secrets management
- hashicorp/terraform — GE's IaC
MONITOR (watch for trends)¶
- langchain-ai/langgraph — agentic patterns
- crewai/crewai — agent framework
- microsoft/autogen — agent framework
- modelcontextprotocol/servers — MCP ecosystem
- ollama/ollama — local model running (future relevance)
LEADERS:INFORMATION_PROCESSING¶
DAILY_ROUTINE¶
MORNING: scan TLDR AI, check Anthropic/OpenAI/Google blogs, review GitHub release notifications for critical repos
MIDDAY: check X/Bluesky for trending AI engineering topics, review npm advisories
EVENING: compile daily briefing for Dirk-Jan (see tech-radar-methodology.md for briefing format)
WEEKLY_ROUTINE¶
MONDAY: review newsletter backlog (Latent Space, Bytes, Node Weekly, React Status)
WEDNESDAY: deep dive into one topic from the week's signals (write assessment if warranted)
FRIDAY: update tech radar positions if any movements detected during the week
QUARTERLY_ROUTINE¶
FULL_RADAR_REVIEW: score all tracked technologies, update quadrant positions
CONFERENCE_REVIEW: watch/read key talks from quarter's conferences
STRATEGIC_BRIEF: write quarterly technology landscape summary for Dirk-Jan
LEADER_LIST_UPDATE: review this page — add new relevant voices, remove inactive ones
SIGNAL_QUALITY_FRAMEWORK¶
PRIMARY_SOURCE: official announcements, release notes, changelogs (highest trust)
SECONDARY_SOURCE: recognized experts (Karpathy, Simon Willison, etc.) commenting on primary sources
TERTIARY_SOURCE: community discussion, benchmark results, user reports
NOISE: speculation, rumors, engagement-farming takes, "hot takes" without evidence
RULE: never escalate to Dirk-Jan based on tertiary/noise sources alone — verify with primary or secondary source first