what are all different areas of ai people are working on
everyone is talking about AI. Elon Musk, our governments, Pope Leo, your neighbour.
that alone tells you how big this change is — and it is happening right now.
— listing and tracking down the sub-areas —
0. investing in AI
the money that makes everything else possible. shapes what gets built and what doesn't.
— venture capital: a16z, Sequoia, Khosla, Benchmark, general catalyst
— corporate bets: Microsoft ($13B in OpenAI), Google, Amazon, Apple
— sovereign wealth: UAE, Saudi Arabia, Singapore — nations buying into the future
— Wall Street: Nvidia stock, AI ETFs, analyst coverage, public market bets
— government funding: DARPA, NSF, national AI initiatives
— startup ecosystem: YCombinator, accelerators, seed rounds
1. fueling AI
the physical world that keeps AI running. often invisible, always essential.
— energy: data centre power demands, nuclear revival, solar farms for compute
— water: cooling infrastructure, water consumption of large training runs
— real estate: hyperscale data centre construction, land acquisition
— annotation manpower: the human labellers behind every model — Scale AI, Remotasks
— RLHF workforce: contractors in Kenya, Philippines, India doing feedback work
— rare earth & supply chain: minerals for chips, geopolitics of hardware
2. making & understanding AI
building the capability and trying to understand what we've built.
— foundation model labs: OpenAI, Anthropic, Google DeepMind, Meta AI, xAI, Mistral
— hardware & chips: Nvidia, AMD, TSMC, Google TPUs, custom silicon
— cloud infrastructure: AWS, Azure, Google Cloud, CoreWeave
— data: datasets, labelling, synthetic data — Scale AI, Appen
— tooling & frameworks: PyTorch, JAX, Hugging Face, vLLM, LangChain
— open source models: Llama, Mistral, community ecosystem
— interpretability: mechanistic understanding of what's inside models
— alignment & safety: MIRI, ARC, Redwood Research, DeepMind safety
— benchmarking & evaluation: how do we even measure capability
— ML theory: the mathematics underneath all of it
— science communication: Karpathy, newsletters, journalists making it legible
4. deploying AI
the layer between labs and end users. building products on top of models.
— AI-native consumer apps: Perplexity, Midjourney, Character.ai, ElevenLabs
— coding tools: Cursor, GitHub Copilot, Replit
— enterprise AI: Microsoft Copilot, Salesforce, ServiceNow
— vertical SaaS: Harvey (law), Abridge (medicine), Glean (enterprise search)
— agent frameworks: autonomous pipelines, workflow automation
— API layer: the middleware between labs and builders
— edge AI compute: running models on-device, not in the cloud.
Apple Neural Engine, Qualcomm AI, Samsung, Rabbit r1, Frame glasses.
lower latency, privacy-preserving, works offline. a big and underrated frontier.
5. using AI
every domain of human work, being transformed. this is the granular tracker.
— software engineering
— creative writing & journalism
— visual art & design
— music & audio
— film & video
— medicine: diagnosis, drug discovery, genomics, radiology
— mathematics: theorem proving, research assistance
— science: physics, chemistry, biology, materials
— robotics & physical AI
— autonomous driving
— space exploration
— education
— law
— finance & trading
— agriculture
— manufacturing
— language & translation
— defence & cyber war: AI-guided weapons, autonomous drones, cyber offence & defence,
battlefield intelligence, Palantir, DARPA, nation-state AI arms race
— gaming & interactive media: NPC intelligence, procedural worlds, game design co-pilots,
AI dungeon masters, real-time character behaviour, Inworld AI, Ubisoft Neo NPC
6. governing AI
not against it — trying to shape how it develops and who it serves.
— national policy: US executive orders, China AI regulations, UK approach
— supranational: EU AI Act, UN AI governance bodies
— standards bodies: NIST, ISO AI standards
— policy think tanks: CSET, GovAI, Centre for AI Safety
— international coordination: AI Safety Summits, Bletchley Declaration
— corporate governance: internal ethics boards, responsible AI teams
7. resisting AI
people who want to slow it down or stop specific uses. legitimate concerns.
— labour & unions: writers, actors, illustrators — SAG-AFTRA, WGA strikes
— artists & copyright: lawsuits, visual artists, Getty Images
— religious & ethical voices: Pope Leo, faith communities
— academic critics: Timnit Gebru, Gary Marcus, Emily Bender
— pause & slowdown movements: Future of Life Institute letter
— anti-surveillance: facial recognition bans, biometric data fights
— environmental: AI energy and water consumption concerns
8. second order AI impact
not what AI does — what AI does to us. the ripple effects on how we live.
— psychology: attention, cognition, dependency, the outsourcing of thinking
— identity: what does it mean to be human when machines can do what we do
— relationships: AI companions, loneliness, parasocial bonds with models
— work & meaning: if AI does the work, what do people do with their days
— culture: who makes art, who owns stories, what is authenticity now
— education & childhood: kids growing up with AI tutors. what does that produce
— inequality: AI amplifying the gap between those who use it well and those who don't
— political: deepfakes, information warfare, trust in institutions eroding
— spiritual: is this a tool or something more. where does consciousness fit