Hi. My name is Daniel Miessler and I'm an AI and cybersecurity researcher and founder in the San Francisco Bay Area. I've been building and writing online since 1999.
Right now I'm excited, but also worried, about humanity's future.
We've been trained for centuries to build our identities around work we don't really care about: our title, our place in a hierarchy, and our usefulness to someone else's company. And now AI is about to let most of those companies bring most knowledge work in-house.
That will be extraordinarily disruptive, but it's also an opportunity, because being a corporate drone was never a peak human experience. So many of us spend our best hours, days, and years dreading Monday and sitting in pointless meetings. It's no way to live, and it never has been.
I'm excited about AI's potential to give people a more meaningful alternative: one based on who they are and what they're trying to achieve.
So that's what I do. I help people (and companies) articulate and pursue their Ideal State. That's the loop above. LifeOS runs it for a person, and Vector runs it for a company.
- Using AI to upgrade people, so everyone gets the starting conditions the lucky already have. That's the whole mission, really.
- Security, which I did for a long time and still can't stop thinking about. Once you've watched enough systems fail, you build everything differently.
- Publishing what I think as I think it: around 3,000 essays since 1999 and a weekly show. I think better out loud, and it's the best way I know to find the people who care about the same things.
Everything below is that loop, run on something.
LifeOS
The AI harness that runs the loop: it captures what you ultimately want and carries that intent into every task, then checks the output against it. Formerly PAI.
TELOS
A framework for writing down deep context about what matters to you, so an AI can actually work toward it.
Daemon
A personal API framework from the Universal Daemonization idea in my 2016 book. Mine is live at daemon.danielmiessler.com.
Substrate
A framework for understanding what humans are, what we want, and what progress actually means.
TheAlgorithm
The general problem-solving loop itself, written down: verifiable Ideal State Criteria and the climb toward them.
SecLists
The security tester's companion: wordlists and payloads.
Fabric
An open-source framework for augmenting humans using AI, built on crowdsourced prompts for extracting wisdom, analyzing security, summarizing content, and hundreds more.
More projects (earlier and experimental work)
| Project | What It Does |
|---|---|
| RobotsDisallowed | The most interesting robots.txt disallowed directories |
| Ladder | A system for autonomous creation and optimization (concept, not active) |
| Frames | Positive and negative mental frames |
| ExtractWisdom | The original extract-wisdom prompt project (now in Fabric) |
| yt | Pull transcripts from YouTube |
| athi | An AI threat modeling framework for policymakers |
| Caparser | Find apps sending sensitive data unencrypted |
Each of these is the loop pointed at one question, and a few of them are live systems that keep running on their own.
|
samsaid.ai A neutral archive of what Sam Altman has said dariosaid.ai A neutral archive of what Dario Amodei has said danielsaid.ai My own positions, held to the same standard eternalquestions.ai The semi-unsolvable questions humanity keeps asking therealinternetofthings.ai My 2016 book on where IoT was actually headed experimentsinfiction.io A public lab for AI-assisted fiction, with the human share of every story disclosed aiharnesses.ai The AI coding harnesses compared: vendor claims and what people actually say |
shouldwecontrolopensource.ai A structured debate on controlling open-source AI aiunderstands.ai 18 mysteries that test whether AI genuinely understands defineitintext.ai If you can define the work in text, AI can do it thevulnequation.ai Interactive economics of vulnerability discovery highentropycontent.io Why high-entropy content wins oursafe.ai Secure AI For Everyone, a practical AI-safety taxonomy ispalantirtrustworthy.io An open inquiry into Palantir, with both steelman cases and every source public permanentstandardtime.io The case for permanent Standard Time, with the counter-argument at full strength |
daemon.danielmiessler.com My live personal daemon: my current state and work as an API thesurface.ai An AI-curated intelligence feed usdebt.io The US national debt as living data vulnerabilitydata.io CVE and patch data as a living dataset newarkcacrime.org Civic crime intelligence for my hometown usstats.io Over 100 US indicators back to the 1970s, straight from primary government data |
The pieces that carry the current thesis:
From Prompt Engineering to Intent Engineering, I Think AGI Just Happened, The Three Components of Becoming AI Antifragile, The Coming Divide: AI-Native or Left Behind, A Unified Theory For How AI Will Affect Jobs, and The Most Durable Human Value.
Latest posts (auto-updated daily):
- The Bitter Lesson, Generalized
- Vibe Coding Isn't Using AI to Make Things
- Focus on Problems Instead of Companies
- The Hard Problem of Jealousy and the Easy Problem of Consciousness
- The Post-AI Security Floor
- Your Energy Level Changes How You See the World
- General Hill-Climbing: The One Main Loop in AI
All 3,000 or so essays are at danielmiessler.com/blog.
- A Conversation with Alan Blount of Google Cloud
- Introducing CORE
- A Conversation With Jonathan Kreiner
- A Conversation With Prajakta Jagdale
- Reverse Engineering with AI
The show is Unsupervised Learning on YouTube.
Blog, Newsletter, YouTube, X, LinkedIn, Books, All repos