706 saved links, 2013–2026 (peak: 2023). The oldest tag in your vault — it predates your AI pivot by a decade, and you can watch the pivot happen inside it: the Forbes/Entrepreneur/Inc. self-improvement drip of the early years gives way, somewhere in 2024, to a single consuming question — what does a business even look like when software is nearly free and agents do the work? Recent saves are less “how to start a startup” and more “how to build an AI-native company,” which is exactly the altitude an Enterprise AI Architect needs to think at.
Related: Leadership & Management · Career · AI (General) · Productivity & Learning
The thesis you keep re-saving: sell work, not software
The through-line of your 2024–2026 saves. AI products stop being tools people use and start being labor businesses buy.
- Sarah Tavel: AI startups sell work, not software — the 2024 original; everything below is a variation on it.
- Jeffrey Paine: AI labor is not SaaS — “banger by @jpaine,” says your note; the SDLC as the enterprise vertical to double down on.
- Bain: will agentic AI disrupt SaaS? Five scenarios and Bain: AI becomes a modular business platform — the consultancy view, twice saved.
- IBM: agentic AI’s strategic ascent — operating-model redesign, 800 executives surveyed.
- SVPG: build vs. buy in the age of AI — the reminder that enterprises hide millions of lines of business logic behind “simple” solutions.
- Ethan Mollick: every OpenAI product update kills a bunch of startups — the platform-risk memo.
- The skeptics’ bench: “Signal v. Noise: Agentic AI is marketing” and “A Project Is Not a Bundle of Tasks” — Google Docs founder Steve Newman on why naive automation curves mislead. Keep these next to the Bain decks.
The builder economy
Software got cheap, and your saves track what that does to who gets to build.
- Elena Verna: the Mom-and-Pop SaaS era has arrived — your note: “how many good ideas never progressed because of the cost of software?”
- Lovable hits $500M annualized revenue, 1M new projects a week — “those ‘vibe coding is dead’ predictions are quite premature.”
- The rise of micro apps: non-developers writing apps instead of buying them — “just build things.”
- Steve Blank: your startup is probably dead on arrival — the sober counterweight: pre-2025 founders built stacks for a world where software was bespoke and expensive.
- How to learn AI by building real projects, not tutorials — ship a live MVP, iterate from feedback, learn only what blocks you. Basically your own transition strategy in article form.
- “I built an AI app in 4 days — here’s how” · John Rush builds a working app with Bolt without touching code · a “vibe coded” MVP-finder app in .NET 10 + HTMX — the proof-of-speed exhibits.
Founder playbooks & agent toolkits
The practical shelf: how to actually run a company on AI. Directly reusable for your agentic-AI projects.
- Anthropic’s Founders Playbook (PDF) — how to build a company with AI; the note you saved reframes its workflow automation as “AI employees that each do one thing.”
- OpenAI: building an AI-native engineering team (PDF) — agents across the SDLC, aimed at engineering leaders.
- Claude Code docs — saved for the startup idea attached: watch what power users do, then wrap it in an app for people without a terminal.
- The skills repos: AI agent skills for startup founders · reusable Markdown skills for AI marketing agents · prompts for product management tasks.
- Midday — all-in-one AI business assistant for freelancers and Eigent — the latter open-sourced after “Anthropic Claude Cowork just killed our startup product 😅.”
- Google Cloud: four steps for startups to build multi-agent systems — localhost to scalable deployment.
- Lessons from building ~300 agents across 5 startups — field notes, not theory.
The founder shelf — evergreen canon
The older saves that still earn their place.
- Chris Dixon: the idea maze — the oldest-dated item on this page, still the best filter for “is this idea actually new?”
- Tim Ferriss: fear-setting — the quarterly exercise behind “my biggest business and personal successes.”
- A dozen things learned from Don Valentine — “the money flows as a function of the story.”
- Wired on Googlenomics — Hal Varian’s architecture of Google’s business, saved on his retirement.
- How a16z became the masters of venture capital influence — the VC-as-media-company analysis.
- Hampton’s Founder’s Library — the essential-books-for-founders index.
- Oege de Moor of Semmle, interviewed — the founder behind GitHub’s code-scanning acquisition (and, later, much more).
- Taste and craft: Inside Linear — building with taste, craft, and focus and Dylan Field’s storytelling masterclass with Nilay Patel.
- Hamel Husain: “we don’t have any distribution” — audience-building for technical founders; relevant to your own public-profile work.
- The best business advice: burnout comes from not having enough wins — the mindset flip.
Industry watching — signals & skeptics
- The State of Solo Founding — over 1 in 3 startups now started solo; the structural consequence of everything above.
- Kevin Weil (OpenAI CPO) on how AI changes moats, skills, and startup playbooks and how Claude Code maintains product velocity — Lenny’s Podcast — product management in the agent era.
- The Google exec spending 20 hours a week experimenting with AI — “AI as product manager, me as architect” — a working preview of your target role.
- “The Great AI Deskilling has begun” — the workforce risk nobody budgets for.
- Ed Zitron: the era of the business idiot — your note: “aligns with the things I experienced during both of my tenures” at a large employer. The long, angry one.
- RedMonk: AI assistants are now organizational accelerants — the shift from selling to individuals to selling to teams.
- Armon Dadgar (HashiCorp) on AI-native DevOps — where your old world and new world shake hands.
- Namma Yatri’s /open page — an Indian Uber competitor publishing all its revenue and KPIs live; radical-transparency business model.
- Doug Turnbull: the hidden dangers of selling search solutions — success is subjective and the work invisible; applies word-for-word to selling RAG.
Opportunity radar
- YC’s Requests for Startups — the “year of AI agents” idea list — plus the AI Startup School.
- 1,500+ startups that just raised and are about to hire — the public list, with the raw spreadsheet.
- CS639: foundation models for young engineers and entrepreneurs and MIT 6.824 Distributed Systems — saved together from one thread: learn distributed systems before you build foundation-model companies. Your infrastructure background already checks that box.