The AI panorama: one scroll, five zones 63 nodes, from compute at the riverbed all the way to rules outside the city. This is not a knowledge map but a position map : every node is tagged with “from which score on you really need to understand it”. First turn on “only level N-x” and the nodes above you fade out — those are what you take on next. Drag to pan, scroll to zoom, click a node for one line.
Only: All L1 Listen L2 Wade L3 Handle L4 Make L5 Orchestrate L6 Set the board
Find Zoom in Zoom out Reset 100% ← drag to pan, or scroll horizontally/vertically; street, city and wild lie to the right →
River · compute & data Base supply: the model's grain and horsepower Bridge · models & training Connection layer: turning compute into capability Street · applications & products Market layer: capability turned into something usable City · organisation & process City gate layer: it counts only once it enters the organisation Wild · rules & society Outside the city: the feasible region is drawn here GPU clusters GPU clusters: the bricks of compute, stacked one by one. Memory bandwidth Memory bandwidth: models run slow because this corridor is narrow. Interconnect Interconnect: whether ten thousand cards act as one depends on this line. Chip process Chip process: how much compute fits in the same area, and so the cost curve. Data-centre power Data-centre power: a rack's wattage often becomes the bottleneck before the algorithm does. Cooling Cooling: the other half of compute cost hides in air and liquid. Optical modules Optical modules: the pins that stitch a thousand machines into one big computer. Training corpus Training corpus: the model's grain; its source decides its taste and bias. Data cleaning Data cleaning: dirty data doesn't make a model dumber, it makes it more confidently wrong. Synthetic data Synthetic data: feeding models with models — cheap, but it amplifies its own flaws. Labeling & human review Labeling & human review: the most expensive layer of labour, and the draft of “what counts as right”. Copyright & licensing Copyright & licensing: whether a corpus may be used is a legal riverbed, not a technical question. Vector store Vector store: the warehouse that lets a model “recall”; the ceiling on retrieval quality. Data feedback Data feedback: real production questions returning to the training set — that is a closed loop. Transformer Transformer: the bridge that made attention parallelisable. Attention Attention: within one sentence, who should look at whom — the mechanism of comprehension. Parameter scale Parameter scale: one dimension of capability, not all of it; past a point the gain slows. Tokenizer Tokenizer: the model's “characters” are not yours; many oddities start here. Pre-training Pre-training: one huge upfront investment that decides how wide the base is. Supervised fine-tuning Supervised fine-tuning: teaching it to speak in human format, not to think. RLHF RLHF: making it “likeable”, and teaching it to pander. Preference optimisation Preference optimisation: a cheaper alignment route, at the price of extremes. Distillation Distillation: pressing a large model's craft into a small one — cheap, but details get lost. Quantisation Quantisation: trading precision for speed and VRAM; most deployments rely on it. Inference engine Inference engine: on the same card, how fast it runs is half decided here. Context window Context window: the range it can “see” right now, not its memory. Mixture of experts Mixture of experts: waking only the experts that need to work, saving compute. Multimodal alignment Multimodal alignment: images, audio and text landing in one semantic space. Benchmarks Benchmarks: the score you trust may just be a question it has practised. Chat assistant Chat assistant: the most widespread entry point, and the most misunderstood one. Agent framework Agent framework: scaffolding for its hands and feet, not for its brain. Retrieval augmentation Retrieval augmentation: look it up before answering — the first gate against hallucination. Tool calls Tool calls: issuing invoices, querying databases, placing orders — that is real work. Workflow orchestration Workflow orchestration: stringing steps into a production line that can stop on failure. Coding assistant Coding assistant: the first scenario fully absorbed, because correctness is compilable. Image generation Image generation: the double battlefield of aesthetics and copyright. Voice & realtime Voice & realtime: conversation starts only below the human patience line of latency. Video generation Video generation: cost and consistency still fight on this layer. Digital humans Digital humans: the skin matured first, the brain followed later. Browser automation Browser automation: the old world without APIs gets connected through here. Low-code building Low-code building: lets the business build its own tools — and a pile of orphan systems. Private deployment Private deployment: data stays home, capability takes a discount. A hard trade. Cost & billing Cost & billing: per call, per volume, per seat — pricing is product design. Human–machine division Human–machine division: which step to hand over and which must stay is a design question. Task ledger Task ledger: collaboration without a ledger is not collaboration. Acceptance criteria Acceptance criteria: work you can't write criteria for should not go to an Agent. Permissions & audit Permissions & audit: what it may touch and what it touched must be retained. Knowledge-base governance Knowledge-base governance: a dirty library makes even a strong model confidently wrong. Shared definitions Shared definitions: same word, different meaning is the most expensive tax in human–machine work. Agent headcount Agent headcount: how many, which roles, reporting to whom — the org question comes before the tech. On-call & patrol On-call & patrol: who watches at night decides whether the system can sleep. Review & codify Review & codify: one mistake must buy one rule, or you will make it twice. ROI ROI: whether the saved labour can be counted decides whether this continues. Regulation & compliance Regulation & compliance: the feasible region's frame is often drawn outside technology. Copyright ownership Copyright ownership: who owns the output is still unsettled ground. Privacy & data Privacy & data: training, inference and retention — each has its own mines. Deepfakes Deepfakes: the rise in cost of trust is a bill society pays. Employment structure Employment structure: not “how many jobs vanish” but “which stretch gets replaced”. Compute geopolitics Compute geopolitics: where the cards come from decides who trains the next generation. Open vs closed source Open vs closed source: openness buys diffusion, closing buys control — both have a price. Standards & interoperability Standards & interoperability: capability only really flows when interfaces unify. Energy cost Energy cost: every inference sits on a unit of electricity. Trust & liability Trust & liability: who answers when it breaks decides how far this can go.
What each of the five zones governs 14 nodes
River · compute & data Base supply: the model's grain and horsepower
15 nodes
Bridge · models & training Connection layer: turning compute into capability
14 nodes
Street · applications & products Market layer: capability turned into something usable
10 nodes
City · organisation & process City gate layer: it counts only once it enters the organisation
10 nodes
Wild · rules & society Outside the city: the feasible region is drawn here
How to read this scroll ① Turn on “only level N-x” first and accept the batch you should know; don't grab everything. ② Then switch the filter off and look at the whole map: you will find that downstream problems are often answered upstream — a street that won't run is usually a bridge (model capability) or a river (data) issue. ③ Finally click three to five nodes and read their “one line” aloud. If you stumble, you only memorised the name.
This map is a scroll rather than a flow chart because real AI is not a pyramid — it is a whole market street with upstream and downstream: the water in the river decides how much the bridge can bear, whether the bridge connects decides how big the street can trade, the city gate decides what may enter, and the law outside decides how far the city can be drawn. Measure my level → Six levels