🤫husshhussh
🤫husshhusshOnePuppy
Reference design · Systems · edge AI · scheduling

When should an agent leave the device?

The local-first compute thesis treats placement as a human decision as well as a scheduler decision: stay local when latency or privacy makes that right; propose a temporary burst only when the task earns it. It is a design target, not a shipped control plane in this repository.

Read the companion essayAll open problems
Problem statement

The hard question.

Jointly optimize latency, energy, memory pressure, cost, and consent while giving the owner a reason they can understand and override.

The same person may have an on-device model, a Mac, a desk system, and a cloud account. A scheduler that only sees FLOPS cannot decide whether moving the work is acceptable.

Status

What is true today.

The local-first compute ladder and hu_ssh transport are published research/reference design. This repository does not contain the placement engine, provider, credential vault, burst API, job store, or macOS agent.

Primary source: Private Agent One heterogeneous-supercompute dossier

Constraints

The work is only useful if these survive.

  • The user’s instruction and data boundary outrank a throughput-only optimum.
  • A local failure is not automatically a cloud-worthy escalation.
  • Credentials, job lifecycle, and teardown need explicit trust boundaries.
  • Every proposed burst needs an owner-visible explanation.
Evaluation

How we would know.

Trace-driven simulator with reproducible device and workload assumptions.
Latency, energy, cost, and data-movement baselines.
A user study on whether placement explanations are understandable and actionable.
A first contribution

Start with something that can fail.

Build a placement simulator first. Compare resource-only policy, privacy-first policy, and owner-overridable policy on the same trace before provisioning any cloud resource.

Read it plainly, then make it better.

The companion essay is written for a broader technical audience. The repository and developer community are the places to turn a claim into a contribution.

Companion essayJoin Discord

One is a product of Hushh Technologies Corporation (brand: 🤫 “hussh”), an independent company. One runs on third-party silicon, systems, and cloud; platform names are used solely to describe where One software runs and imply no affiliation, endorsement, or sponsorship by those platforms. Our own go-to-market and bill-of-materials partner programs are real and actively in pursuit; we name a partner only once an agreement is executed.