Story
OpenAI is becoming a deployment company
The harder phase of the AI cycle is no longer model access. It is institutional adoption.
Work, craft, and institutions
People at Work
How people learn, organize, build, and make decisions as intelligent systems enter the workplace.
Latest from People at Work
90 published works
Story
Anthropic’s usage data shows where automation actually lands
The Economic Index makes the labor question more specific—and more uneven.
Story
Google I/O moved the interface from answers to action
Gemini’s expansion is less about a single model than a new layer across Google’s products.
Story
Apple’s AI strategy is becoming a developer story
At WWDC, the consequential shift was the opening of system intelligence to apps.
Story
Figma wants code to remain design material
Config’s code layers are an attempt to narrow the gap without pretending it can disappear.
Story
NVIDIA’s next market is the AI factory
GTC’s language of tokens, racks, and power turns AI into an industrial infrastructure story.
Story
NeurIPS is tightening the social contract of AI research
Reproducibility, disclosure, and responsible research are moving closer to the center of conference operations.
Story
The summer AI calendar has become a product map
ICML and SIGGRAPH now sit alongside developer conferences as launch surfaces for models and tools.
Story
Y Combinator is adjusting Demo Day to an agent-first cohort
The startup batch is becoming a live index of where software creation has become cheap—and where distribution has not.
Story
Cursor is turning the code editor into a control surface
The product’s real ambition is not faster completion. It is supervised software production.
Story
Mistral’s European advantage is operational, not symbolic
The Paris lab is competing on deployability, model choice, and regional trust.
Story
Perplexity is testing the boundaries of the answer engine
A search product becomes a media question when it decides how sources are seen, credited, and replaced.
Story
Anduril’s software model is remaking the defense contractor
Lattice turns hardware programs into a shared autonomy and command layer.
Story
Figure is making the humanoid robot a manufacturing argument
The near-term test is not resemblance to a person. It is reliable work inside an existing line.
Story
Runway is building a studio around the model
Generative video becomes a media company when tools, production, and distribution meet.
Story
Midjourney remains the strange independent lab in the image race
Its product culture offers a counterexample to both hyperscale platforms and enterprise AI.
Story
ElevenLabs is turning voice into infrastructure
Synthetic speech now spans narration, localization, accessibility, and agent interfaces.
Story
Lovable’s rise makes product judgment the scarce input
When an interface can produce working software, deciding what deserves to exist becomes the harder work.
Story
Linear is defending calm software in an agentic moment
The issue tracker’s restraint is being tested by a future of machine-created work.
Story
Notion’s AI layer is becoming part of the workspace grammar
The company’s advantage is context—but context also creates new expectations for permission and trust.
Story
Vercel is competing to host the agent-made web
Deployment speed matters more when software can be generated continuously.
Story
Replit’s agent makes the browser an entire software shop
The promise is broad access; the operational question is how people understand what was built for them.
Story
Databricks is folding models into the data estate
Enterprise AI is becoming a governance and lineage problem before it becomes a model contest.
Story
Hugging Face is the public square of open machine learning
The Hub is simultaneously a repository, a labor market, and a governance experiment.
Story
Cohere is betting that enterprise AI needs a different product culture
Private deployment, retrieval, and organizational control are the product—not afterthoughts.
Story
Thinking Machines Lab is organizing around collaboration
The new lab’s public thesis places human control and customization at the center of advanced AI.
Story
Safe Superintelligence is narrowing the company to one problem
A singular mission can clarify priorities—and postpone the questions that deployment usually forces.
Story
Meta’s open-model strategy is also a distribution strategy
Llama’s reach depends on developers, clouds, devices, and a large ecosystem carrying it outward.
Story
Microsoft is making Copilot an organizational layer
The enterprise advantage is access to work context; the governance burden follows the same path.
Story
AMD’s AI opportunity depends on software catching up
Hardware competition is now measured by the developer path around the accelerator.
Story
Arm is moving from mobile inheritance to AI infrastructure
Its architecture now connects phones, edge systems, and data-center efficiency.
Story
Qualcomm is treating the edge as an AI market
On-device intelligence makes latency, privacy, and power consumption product choices.
Story
CoreWeave’s rise reveals the new cloud stack
Specialized GPU capacity is becoming a market between chipmakers and model companies.
Story
Cloudflare wants agents to run at the edge
The network is becoming a place to execute software, enforce policy, and mediate model traffic.
Story
Supabase turned Postgres into a product movement
The open-source backend shows how a familiar primitive can become a modern developer experience.
Story
Neon is separating the database from the machine
Serverless Postgres makes storage, compute, and branchable environments behave more like software.
Story
Weights & Biases is moving from experiments to agent traces
Observability is expanding from training runs to the behavior of production AI systems.
Story
Harvey’s legal AI is a test of professional accountability
In high-stakes work, speed only matters if review, provenance, and confidentiality remain visible.
Story
Sierra is making the customer-service agent a company representative
An autonomous interface carries policy, tone, authority, and escalation rules into every conversation.
Story
Cognition is selling the idea of a software teammate
Devin turns the benchmark debate into a question about delegation and review.
Story
Canva is broadening design without dissolving the craft
Generative tools make production easier while increasing the value of systems, taste, and governance.
Story
Shopify is treating AI fluency as operational infrastructure
The commerce platform’s internal culture makes a strong claim about how quickly work norms can change.
Story
Stripe’s AI opportunity runs through economic infrastructure
Payments data, billing logic, and internet businesses give the company a different view of agentic commerce.
Story
The Bay Area’s AI geography is tightening again
Labs, capital, compute buyers, and talent are clustering even as the products become globally distributed.
Story
The AI hiring page is becoming a product roadmap
Open roles reveal where companies expect deployment to become difficult.
Story
Open-source AI is becoming an operations choice
The model license is only the beginning; deployment requires teams to own more of the system.
Story
AI policy is moving closer to the data center
Rules about models increasingly meet questions about energy, chips, export controls, and physical capacity.
Story
The next developer stack includes evaluation by default
Prompting is giving way to traces, datasets, tests, permissions, and rollback.
Story
Creative AI companies are learning to speak in workflows
The model is less visible when it becomes one step in editing, collaboration, rights, and delivery.
Story
Why the frontier lab is becoming an institution
As AI systems spread, research companies acquire the responsibilities of platforms, employers, and policy actors.
Feature
The stack beneath the intelligence
A journey from the prompt to the power grid reveals the industrial system inside every AI answer.
Feature
The people who decide whether an agent is ready
Evaluation teams are becoming the quiet editorial layer inside autonomous software.
Feature
Paris builds an AI identity of its own
Mistral’s rise is part of a larger argument about language, sovereignty, talent, and European infrastructure.
Feature
Inside the new design-to-code continuum
Figma, Vercel, Lovable, and Cursor are converging on the same unfinished object.
Feature
The defense startup becomes a systems company
Anduril’s rise shows how software, hardware, manufacturing, and procurement are being recombined.
Feature
What a humanoid robot has to learn before the demo ends
The impressive movement is only a small part of reliable physical work.
Feature
The independent labs shaping generative culture
Midjourney, Runway, and ElevenLabs are building aesthetics as well as infrastructure.
Feature
San Francisco’s second AI city
The current cycle is less about a neighborhood than a network of labs, meetups, houses, and compute buyers.
Feature
The company page as an accountability surface
In the AI economy, products change faster than institutions can explain themselves.
Feature
The long road from research paper to everyday tool
A model becomes a product through a chain of packaging decisions that rarely appear in the benchmark.
Feature
When the workspace starts making decisions
Notion, Microsoft, and Glean are turning organizational context into an active software layer.
Feature
The new economics of making software
Agents compress implementation, but the total cost of ownership moves into review, operations, and maintenance.
Explainer
What does an AI agent actually do?
A practical guide to planning, tools, memory, permissions, and human review.
Explainer
What is a context window—and why is it not memory?
The text a model can see is different from what a product remembers about a person or organization.
Explainer
How does a GPU turn electricity into tokens?
A plain-language route through matrix operations, memory bandwidth, networking, and inference.
Explainer
What makes an AI model open?
Weights, code, data, licenses, and documentation can be open in different combinations.
Explainer
How are AI systems evaluated before release?
Benchmarks are only one layer in a release process that also needs scenarios, experts, and monitoring.
Explainer
Why do AI models hallucinate?
Fluent generation is not the same operation as checking a claim against the world.
Explainer
What is inference—and why does it cost so much?
Training creates a model; inference pays the bill each time someone uses it.
Explainer
How does retrieval-augmented generation work?
RAG gives a model selected documents at answer time, but it does not guarantee that the answer is correct.
Explainer
What changes when software can use tools?
The moment a model can send, buy, edit, or deploy, interface design becomes authorization design.
Explainer
What is a world model?
The term covers systems that learn useful representations of how environments behave over time.
Research
Where people use Claude at work
A reading of Anthropic’s Economic Index, occupation by occupation.
Research
The public evidence for longer-running agents
What task duration reveals about reliability, supervision, and the changing unit of AI work.
Research
A field guide to frontier model documentation
We compared the public artifacts used to explain capability, safety, and release decisions.
Research
The geography of the current AI company
Company headquarters and event calendars show a dense Bay Area core with important European and global nodes.
Research
The agent toolchain, mapped
Editors, runtimes, model providers, observability systems, databases, and deployment platforms now form a recognizable stack.
Research
How AI companies describe safety
A taxonomy of public commitments across labs, infrastructure providers, and application companies.
Research
The hidden work in a generated video
A workflow analysis of prompting, selection, editing, sound, rights, and delivery.
Research
What startup homepages promise in the agent era
A language study of autonomy, speed, teammates, and the disappearance of implementation.
Research
The compute market between NVIDIA and the model lab
Specialized clouds are creating a distinct layer for capacity, scheduling, and systems integration.
Research
An index of AI conference disclosure rules
NeurIPS and ICML are turning tool use, reproducibility, and research ethics into explicit publication operations.
Profile
Fei-Fei Li and the unfinished map of spatial intelligence
From ImageNet to world models, Li’s work keeps returning AI to the richness of the visible world.
Profile
Jensen Huang and the industrialization of computing
NVIDIA’s founder turned a graphics processor into the organizing component of an AI economy.
Profile
Dario Amodei and the institution built around scaling
Anthropic’s CEO has made capability, risk, and organizational design parts of the same public argument.
Profile
Dylan Field and the shared object
Figma’s enduring idea is that design improves when the work becomes visible to everyone around it.
Profile
Arthur Mensch and Europe’s frontier-model wager
Mistral’s co-founder is building a Paris lab around efficiency, model choice, and enterprise deployment.
Profile
Aravind Srinivas and the answer-engine bet
Perplexity’s founder is testing whether synthesis can become the default interface to the web.
Profile
Palmer Luckey and the software-defined contractor
Anduril’s founder returned from consumer hardware with a theory of autonomy, manufacturing, and defense procurement.
Profile
Patrick Collison and the infrastructure view of progress
Stripe’s co-founder has spent a decade treating payments as one constraint in a larger internet economy.