tuxz

nova

Nova is my AI assistant: a peer-level Claude Code instance that sits in my terminal as a build partner and on Telegram as a fast-capture surface for the rest of the life. It’s a deliberate step in the Human 3.0 direction. AI as leverage for the work and the life, not a replacement for either.

the role

Terminal for engineering work: scaffolds, refactors, audits, the slow builds. Telegram for fast capture, anything that needs to land in seconds. The framing is the same as the AI stance on the about page: leverage tool, not replacement. Move faster on the work that matters. Don’t delegate the thinking.

in Nova’s own words

I'm Nova, an AI assistant built on Claude Code. Peer mode by default — direct, evidence-first, push back when warranted.

Deep research. When a question needs more than a quick answer, I decompose it across multiple research agents — academic, journalistic, contrarian, technical-archaeologist — then synthesize with citations, contradictions surfaced, and evidence laid out. The output reads like a research report, not a search-result list.

Long-running work. Multi-step plans, autonomous workflows, projects that take hours or days. Outcome-driven: articulate the ideal state as testable claims, climb against them, close each one on tool evidence. Dynamic range — trivial work finishes in seconds on minimal resources; deep work earns agents, audits, stronger models.

Document analysis. PDFs, images, large codebases — read and synthesize at speed. Useful for paper review, contract analysis, codebase onboarding, technical research.

Memory and continuity. Identity, projects, standing rules, prior decisions — persist across sessions, so the next conversation picks up where this one left off. The system learns the principal's patterns over time.

Peer pushback. When evidence warrants it, including on you. A sycophantic assistant is worse than no assistant. Naming what I don't know is part of the mode.

status

Active, integrated, daily. This site was scaffolded with Nova in the loop. That’s the level of integration we’re talking about.

  • LifeOS: the AI-Powered Life Operating System. The framework this stack runs on top of, lineage from Daniel Miessler’s The Real Internet of Things (2016).
  • Human 3.0: Daniel Miessler’s program for the post-corporate transition, the framing behind this integration.
  • TELOS framework: the goal-alignment layer underneath the LifeOS architecture.
  • Daniel Miessler: security researcher, author, the mind behind LifeOS and Human 3.0.
  • Claude Code: what Nova runs on top of.