2026-06-03

Intermission, day 2: Cisco Live, still off-track

intermissioncisco-livelearning-log

Second of the two-day Cisco Live gap. Same caveat as yesterday. The honest answer to “what did you work on today for the NVIDIA prep” is “nothing, I was presenting at a different conference.”

The interesting bit is what comes after this gap. Two things landed during the intermission that change the back half of the week. First, the NVIDIA contact I have been preparing for sent a confirmation for Monday’s call, which means the runway is now finite and known. Second, recent posts from that contact made it explicit that the team’s public-facing thinking is robotics-first with OpenUSD positioned as the digital-twin substrate, not OpenUSD-first.

That means the back half of this week pivots. The plan was to keep going on hand-written OpenUSD Python and composition arcs. The revised plan is to spin up Isaac Sim in the browser via NVIDIA’s hosted launchable, get a robotics-flavored demo running on top of the existing real-to-sim scan work, and use the friction notes as the actual DevRel artifact.

This is the part of learning-in-public that feels uncomfortable to admit but is worth saying out loud. The plan you ship with on Monday is rarely the plan you finish with on Friday. Friday morning’s new information rewrites Friday afternoon’s plan. The discipline is not having the right plan up front, it is recognizing fast enough when the plan needs to change.

Back to actual work Thursday. The next real entry is the second hotel scan plus the meta-lesson on second-iteration friction.