Nathan Maxwell

Builder
I build software and run a gaming club.
What I'm working on, and why.

Problems

  1. P1.Data loss and cognitive overload when coaching multiple swimmers
  2. P2.AI agents don't carry context across sessions
  3. P3.Agents as assistants, not thinking partners

Missions

  1. M1.Build a context system that compounds with me, so my AI learns from every conversation
  2. M2.Use software and AI to multiply what I can do
  3. M3.Build software that lets coaches coach

Challenges

  1. C1.Balancing build time vs. distribution time
  2. C2.Signal vs. noise at scale in the inbox

Current Projects 5

Hydrolyze — Swim Analysis Platform
Hydrolyze is a swim analysis platform for coaching. The mission has not changed.
active · impact 420
Felt Faction
A 12-person gaming club where the casino is the draw and networking happens naturally.
active · impact 520
Parasym
iOS app that connects to a Morpheus M7 (or any BLE chest strap exposing R-R intervals) and correlates the realtime HR + HRV data with Apple HealthKit to produce personalized health insights. **Realtime workout mode (HRR + motion fusion) is fully deferred to v0.2.**
planning · impact 130
gbrain omnipresence rollout
Make my second brain available everywhere I work.
in-progress · impact 120
Games Room
A game-night companion app being productized from the Felt Faction program so other groups can run their own games nights.
active · impact 40

Current thinking 5

What I'm thinking about right now. Click any idea for the thesis — what the idea is, how it's structured, and where it sits in the network.

  • writing2026-07-28
    The previous casino pack design (V0.3 spec, 2026-07-12 architecture, V0.24 → V0.29 ship path) assumed the **host scans** every member's chip stack at the end of the night. The host's iPad became the central vision hub; the member's phone was a spectator.
  • writing2026-07-05 · refreshed 2026-07-28
    The Hermes dashboard wedge on 2026-07-05 surfaced a structural pattern that no single atom captures: **a long-running local service has at least three orthogonal failure modes, and no single detector catches all three.** A supervisor catches one. A watchdog catches a second. A forensic capture catches the third. The triad is the minimum viable operational discipline for any self-hosted service you actually depend on. Skipping any layer is a known-recurrence failure waiting for the next wedge.
  • atom2026-07-23 · refreshed 2026-07-24
    Operationalize the 'never fool yourself' rule by asking one question before any decision: 'Where's the weakest assumption in this?' If you can't identify it, you haven't thought hard enough. If you can identify it, you've found the exact place your reasoning is most likely to collapse and the place to stress-test before committing.
  • atom2026-07-23 · refreshed 2026-07-24
    By locating the cause of behavior in subjective experience rather than external events, NLP effectively promotes the internal representation to the role of primary reality. This inverts the usual assumption that facts come first and feelings follow. For practitioners, it means the most durable interventions target the model a person holds of a situation, not the situation itself.
  • atom2026-07-23 · refreshed 2026-07-24
    NLP rests on one deceptively simple premise: human behavior is driven not by objective reality but by subjective experience. If that's true, the implication is enormous—changing someone's inner representation of events changes their behavior, without changing the events themselves. Bandler and Grinder built an entire discipline of intervention around this single causal claim.
  • atom2026-07-23 · refreshed 2026-07-24
    A nerve is a glistening white cordlike bundle of fibres, surrounded by a sheath, that connects the nervous system with other parts of the body. The description evokes something almost rope-like — not the abstract 'wiring' metaphor we usually default to. Treating nerves as visible, physical cords reframes how we think about nerve damage, regeneration, and the tangibility of the nervous system.
  • atom2026-07-23 · refreshed 2026-07-24
    The reason children can pick up any human language is that unlearnable words have already been filtered out across generations. A language's learnability isn't a coincidence or a gift from designers — it's the sediment of countless small cullings. Difficulty in language learning reflects what survived the filter, not what was originally possible.
  • atom2026-07-23 · refreshed 2026-07-24
    From Absalom and Alcibiades in antiquity to Cicero, Nero, and Stalin in later centuries, narcissistic leaders recur with striking consistency across civilizations. The pattern is too persistent to be coincidence, suggesting power either attracts narcissistic personalities or amplifies existing traits. Narcissistic leadership isn't a modern aberration — it's a recurring archetype of the powerful.

Books 0 Recent (5)Most Impacted

📖 Currently Reading 0

Articles 6 RecentMost Impacted

Articles I've recently ingested or touched.

Podcasts 6 RecentMost Impacted

Podcasts I've recently listened to or touched.

Preferences

OS
macOS
Tools
Xcode wrangler bun tsx sqlite ripgrep git obsidian zotero gbrain hermes-agent
Communication
Direct, honest.