domain · ai-native
Build for an AI-native operating model
How operating changes when the agent is on the team. Verification loops, evals, AI-mediated workflows, agent infrastructure.
Strongest claims
- Measure AI search on three layers: Presence, Readiness, Business Impact Aleyda Solis
- Run agent-first GTM as a three-stage flywheel with one named agent per job Yamini Rangan
- Agent-first content has six platform layers, access, discovery, capability, format, token, UX bridge Addy Osmani
- Agents are first-class product users; design for output reliability, not navigation Elena Verna
- Agents work when treated as a team, not a single super-tool Claire Vo
- A single seeded fake claim can self-confirm in AI Overviews Lily Ray
Adjacent domains
- engineering · 183 co-occurrences
- founder-operator · 52 co-occurrences
- product · 50 co-occurrences
- leadership · 42 co-occurrences
- pmm · 41 co-occurrences
- gtm · 35 co-occurrences
- future-of-work · 25 co-occurrences
- growth · 16 co-occurrences
Synthesis patterns in ai-native
- Agent-first GTM (rebuild, don't bolt-on)
- Agents are first-class users, design for output, not navigation
- Agents mapped 1:1 to JTBD with named human checkpoints
- AI code generation shifts the quality bottleneck, it does not remove it
- AI defensibility comes from non-AI moats
- Frontier AI capability is now constrained by deliberate policy, not technical limits
- Build for the next model, not the current one
- Context, not capability, is the bottleneck
- The Economic Turing Test, outcomes pricing, agent labor, revenue per employee
- Evals are data analysis, single judge, binary rubrics, error analysis first
- Execution is becoming free; judgement is the part that doesn't compress
- Generalists with taste, shipping end-to-end
- LLM-as-OS, post-training as moat
- Make the implicit explicit
- The harness layer is where control concentrates above the model
- Research preview, frontier programs, seasons not roadmaps
- Sales is an engineered system, not individual art
- Strip what does not compound: AI abundance makes judgment the scarce resource
- Substrate runs the loop; humans run alignment and taste
- Verification, not execution, is the irreplaceable human job
411 insights in ai-native
- 10-80-10: human direction, AI execution, human polish · Arvid Kahl
- An automation that works 95% of the time is not an automation · Cat Wu
- Advisor-tool replaces ensemble-of-3 stability hacks at near-Sonnet rates · Cat Wu
- AEO is a GTM capability, not an SEO experiment · Maja Voje
- Measure AI search on three layers: Presence, Readiness, Business Impact · Aleyda Solis
- Keep agents on ask-for-approval until you know their failure modes · Ethan Mollick
- Agents differ from chatbots because they give the AI a computer to use · Ethan Mollick
- Run agent-first GTM as a three-stage flywheel with one named agent per job · Yamini Rangan
- Agent-washing is the buyer-confusion surface PMM should attack · Gartner
- Agent-first content has six platform layers, access, discovery, capability, format, token, UX bridge · Addy Osmani
- Agents are first-class product users; design for output reliability, not navigation · Elena Verna
- Agents work when treated as a team, not a single super-tool · Claire Vo
- At agent scale, security governance must shift from humans approving humans to automated agent-identity and policy systems · Jonathan Jaffe
- Coding agents erased the cost advantage of cross-platform mobile frameworks · Mustafa Ali
- AI agent swarms self-organize without management overhead because they lack the organizational pathologies management was built to solve. · Ethan Mollick
- AI adoption creates an asymmetric information problem where wins are public and costs are private, requiring deliberate feedback loop design · Charity Majors
- AI is the co-pilot, not the autopilot, automate the 80% you understand, hold the 20% that needs taste · Ben Tossell
- A 7-step AI prompt chain beats one-shot resume rewrites for job search · Nishchal Dua
- AI cold-call compliance follows the prospect's residence, not the seller's HQ · Gartner
- AI replaces PMMs whose job was launch emails, recap docs, and slide tweaks · Sachin Jha
- The number one enemy of innovation is efficiency, AI is collapsing the moats specialization built · Leah Tharin
- AI has crossed the threshold to something indistinguishable from judgment and taste, winners will know what to build, not how · Matt Shumer
- AI vendors cannot correct inflated customer productivity claims without destroying those customers' credibility · Nikhil Suresh
- A single seeded fake claim can self-confirm in AI Overviews · Lily Ray
- Per-employee AI tool spending caps anchored to compensation signal an implicit enterprise ROI model, and are more sustainable than consumption competitions · Simon Willison
- An AI-queryable codebase knowledge graph compresses domain onboarding from months to weeks, enabling generalists to contribute to specialist code · Ahmad Al-Dahle
- Small post-trained models outperform frontier models on narrow latency-sensitive tasks, making multi-model routing the default production architecture · Ahmad Al-Dahle
- Feature-organized companies will struggle as AI ships features autonomously; outcome-organized teams set the direction AI cannot replace · Ahmad Al-Dahle
- Production AI deployments require a separate eval suite per use case, sampled from real production queries, not shared benchmarks · Ahmad Al-Dahle
- Every LLM rewrite changes meaning because the model lacks the author's mental model of what the words are meant to communicate · Sophie Alpert
- "We're using AI" is not a business strategy, defensibility comes from domain expertise, customer relationships, and data, not from the model layer · Sam Altman
- When intelligence is abundant, taste, judgment, relationships, and the ability to identify what is worth doing become the scarce resources · Sam Altman
- The singularity is one smooth curve, vertical looking forward, flat looking backward, never the disruptive shock people expect · Sam Altman
- Third-party evaluators embedded in AI development are the only mechanism that makes pacing commitments verifiable · Dario Amodei
- Building production ML systems at scale · Anand Karunan
- Claude Code subagents run in isolated context windows with model routing, enabling cost-optimized bulk delegation without polluting the orchestrator session · Anthropic
- Audience first, Community second, Product last, and AI makes the inversion much faster · Greg Isenberg
- A conversation-context classifier grants or blocks agent tool calls per call rather than per tool type · Thariq Shihipar
- AI agents that run bug investigations in parallel lower the cost of each shipped defect, raising the rational shipping threshold · Thorsten Ball
- Remote agent environments have displaced the local development setup for a real shipping team · Thorsten Ball
- Engineers who dismiss AI-generated code as low-quality apply the same professional bias that causes domain experts to over-intervene within their specialty · Thorsten Ball
- Minimizing time-to-feedback is the single most valuable practice when building new software · Thorsten Ball
- As AI handles coding reliably, most software bugs shift from implementation errors to specification errors · Thorsten Ball
- Battle cards become workflow primitives, not Notion pages · Gartner
- Production evals must measure UX friction behaviors that damage retention, not just task accuracy · Cat Wu
- Being crisp and concise is underrated. The lack of it signals something more concerning · Chantal Cox
- Appoint one trusted-taste expert as the eval benevolent dictator, committees stall the loop · Hamel Husain
- Data foundation determines AI agent value delivery. · Bogdan Tyskyy
- When the agent isn't doing what you want, fix the context, not the model · Sherwin Wu
- When frontier model pricing rises rather than falls, harness quality becomes the primary investment · Drew Breunig
- Frontier labs are baking native harness preferences into model weights, not layering them on top · Drew Breunig
- AI can generate 180,000 lines of production code in three weeks, but resource management and architecture still require active human babysitting · Rick Brewster
- Build for the model six months out, not the one that ships today · Boris Cherny
- Build for the model six months out, the current model will eat your scaffolding · Sherwin Wu
- Build products at the edge of what does not yet work · Cat Wu
- Domain experts who publicly predict AI catastrophic risks outside their actual domain abuse the trust their expertise earned, causing fear to propagate faster than any rebuttal can follow · Bryan Cantrill
- In AI products, capability overhang is the central growth problem · Amole Naik
- AI capability is not evenly distributed, it spikes where labs have data, rewards, and verification loops · Andrej Karpathy
- Automate the four stages of a growth experiment; keep humans on alignment · Amol Avasare
- Continuous Calibration, Continuous Development (CCCD) is the operating loop for AI products · Aishwarya Naresh Reganti