Notizen aus der Werkstatt.

Local-first persönliche KI, Architektur und die Branchenbewegungen, die die Wette immer wieder bestätigen. Geschrieben in Hongkong.

Das Rennen um die persönliche KI ist ungelöst, und alle rennen auf der falschen Strecke

Zwei unabhängige Beiträge, einer aus einem Creator-Economy-Newsletter und einer aus einer Entwicklerpublikation, kommen zur selben Diagnose. Niemand hat das Rennen um den persönlichen KI-Agenten gewonnen. Der Engpass ist nicht das Modell und nicht die Hardware; entscheidend ist, ob der Agent Zugriff auf die tatsächlichen Daten des Nutzers hat. Genau um dieses Problem ist Ostler herum gebaut.

Single machine, single customer, single source of truth

“Local-first” has become a polite lie in personal AI. Most products marketed as local actually keep the heavy lifting in the cloud, with the cache on your device. Ostler runs every component on a single Mac, the one the customer already owns. This is what local-first looks like when you build it honestly.

a16z’s nächste These lautet ‘beobachten, um zu handeln’. Wir haben sie local-first gebaut.

Kenan Saleh von a16z Speedrun skizziert die nächste KI-Welle: Agenten, die den Kontext kontinuierlich beobachten, vorhersagen, was wichtig ist, und handeln, bevor man sie fragt. Er nennt zwei Produkte, die das tun; beide laufen in der Cloud. So sieht „beobachten, um zu handeln“ aus, wenn die Daten den Rechner des Kunden nie verlassen.

KI-Notizassistenten können das Anwaltsgeheimnis aushebeln. Unserer nicht.

Die New York Times berichtet von Unternehmensjuristen, die in virtuellen Meetings zu Türstehern werden und Cloud-KI-Notizassistenten hinauswerfen. Das Risiko für das Mandatsgeheimnis ist architektonischer, nicht redaktioneller Natur. So ändert sich alles, wenn der Notizassistent auf dem eigenen Rechner des Kunden läuft – mit WhisperKit-Transkription direkt auf dem Gerät und einem manipulationssicheren Einwilligungsprotokoll.

The privacy nutrition label is the only privacy story that survives a lawyer’s read

Apple’s App Store privacy nutrition label is the only privacy disclosure surface in tech with structural enforcement. The Ostler iOS app declares zero tracking and no linked data, because there is no Ostler server to link data to. The architecture writes the label, not the lawyer.

The most intimate technology of our era is a pipeline

A class action filed in California this month alleges that chatbot conversations have been routed through advertising trackers. The argument lands because chatbots are now the most intimate technology many people use. Privacy by policy cannot prevent this kind of leak. Only architecture can.

The diplomat, the researcher, and the founder: three independent verdicts on local-first personal AI

Three completely different vantage points have converged on the same architecture for personal AI over the past six months. Singapore's Foreign Minister using it daily on a Raspberry Pi. Andrej Karpathy describing it on stage. A founder shipping it to customers. They had no reason to agree. They did anyway.

When Apple ships Siri via Gemini, that is not a threat. It is validation.

Apple is about to concede the category. The fact that the most privacy-obsessed consumer-tech company on the planet cannot build a personal AI locally tells you exactly how large the market is, and why the local-first bet is the contrarian trade now being demanded.

OpenAI shipped an open-weight PII model. We are wiring it in.

On 21 April 2026, OpenAI released Privacy Filter as open weights under Apache 2.0. It runs locally, detects eight categories of PII, and slots directly into Ostler's ingest, diagnostic, and pre-flight pipeline. Here is why, and what the release signals.

Karpathy described the architecture. We already built it.

On Dwarkesh Patel's podcast on 17 October 2025, Andrej Karpathy argued that a small reasoner with external memory beats a 1.8-trillion-parameter monolith. That is the architecture Ostler has been running since late 2025. Here is what it means for local-first personal AI.

Why I built a personal AI that never touches the cloud

After twenty years of giving my data to tech companies, I built a personal knowledge graph that runs entirely on a Mac Mini. This is the story of how it got here, and why it matters that nothing leaves the house.

Architektur ist die Richtlinie.

Lokal  ·  Nachvollziehbar  ·  Ihres