Updates from the lab

LiveXYZ’s Spatial AI Digital Twin Navigator is out—something we’ve nursed through a bunch of Sim/Nexus iterations. Day-one goal was boring on purpose: can your VP understand what they’re looking at before we show off reflections on metal?

Pilots said “oh, that’s the valve” faster than on older builds, and customer-success scores beat anything they’d shipped in this line before. Still chewing on orchestration when two agents disagree—write-up coming when it’s not hand-wavy.

A client engineer cloned the repo, ran compose, and the twin actually opened. We wrote the checklist down so we wouldn’t forget what worked: README in a known place, env template committed, CI green, and a twelve-minute video for cold start plus the two ways it usually breaks.

If we can’t hit that bar, we don’t get to call the sprint “done.” Sounds uptight; saves an incredible amount of Slack archaeology.

We’re publishing an internal memo (summary here) on conflict resolution between orchestrated agents when recommendations diverge—especially in capex rehearsal flows. The short version: without explicit ownership of “break glass” decisions, humans tune out.

We’re testing UI patterns that force a single accountable choice with logged rationale, rather than averaging model outputs. More in Q1 lab notes.

Sensor streams, signed handoffs, and TMS integration now produce exportable bundles QA can hand auditors. Web3 showed up only where attestation mattered; the rest is boring, reliable integration.

Sense engineers spent serious time on clock skew and gateway failover—details that never make hero slides but decide whether logistics trusts the number.

A year-long engagement wrapped with elected officials and contractors using the same Sim-backed scenarios in public meetings—replay included. Pushback dropped when everyone watched the same closure logic twice.

Nexus handled ingest from legacy traffic APIs; Sim carried visualization and slider-driven what-ifs. Handoff included training for internal GIS staff.

We documented a rotation playbook after a certificate renewal nearly bricked a pilot fleet. Topics include staged rollout, dual-stack trust, and explicit “last known good” firmware paths.

Published for clients under support; redacted version coming to this news feed next month.

Data-flow and key-custody slides are now sprint artifacts alongside architecture and MVP scope. Security isn’t a week-four panic; it’s a parallel track with explicit owners.

Teams grumbled once, then admitted fewer fire drills before sign-off. We’re keeping it.

We ran private workshops helping teams name where their risk actually lived—often Nexus and Sense while the slide deck said “AI.” Outcomes were scoped sprint briefs, not vague transformation roadmaps.

Repeatable format; contact us if your group needs the same clarity pass.

We released a trimmed version of our internal starter: lint, test, build, container publish, and env sample. Opinionated on purpose—boring beats bespoke for handoff.

License is permissive; issues welcome from teams adopting the same path.

Crews pulled the same locked work package as the control room after sync; conflict resolution was explicit in Nexus. Incidents of “wrong drawing” dropped measurably in the pilot region.

We’re exploring lighter headset profiles for summer heat deployments—hardware lab notes to follow.

We open-sourced a harness comparing chunking strategies and embedding models on redacted customer-shaped data. Goal: kill hand-wavy “RAG accuracy” claims.

Used internally on fintech compliance pilots; contributed fixes welcome.

A Series A manufacturing startup needed investor-grade proof on defect detection. We shipped the detector fast; the remainder was scanners, timestamps, and alert routing—classic Nexus work.

They raised their next round with demo plus architecture deck we co-authored. Still friends with their eng lead.

We flew a small team to a quiet site with bad Wi‑Fi on purpose—good for empathy with field pilots. The output was the whiteboard that became Core, Sense, Nexus, Sim, Forge as shared vocabulary.

Also: unanimous vote to ban PowerPoint for internal demos; show the build or the scope doc.

We published a short technical note (PDF on request) mapping when chains add tamper evidence vs. when they’re theater. Spoiler: most logistics wins are signed logs + integration, not new tokens.

Used in med-logistics and industrial traceability conversations with auditors in the room.

We went public with a pretty simple pitch: if the problem matters, stack what it actually needs—and hand it off like adults. First pilots were messy factories and grumpy compliance queues.

Huge thanks to the early customers who hired us before we had a fancy website. You know who you are.