VC
Value Add VC
โšกHomePulseโšกHelpful Apps๐Ÿ“Blog
โ† Value Add PulseFUNDINGPhysical AI

The Engineer Who Made VLC Run Smoothly Is Now Doing It for Robots

A veteran engineer behind the buttery playback of the free VLC media player is applying the same low-level performance obsession to robotics, tackling the unglamorous software layer that makes machines move smoothly and reliably. It's a bet that the bottleneck in physical AI is execution, not just intelligence.

VLC media player
Origin
Robotics performance
Focus
Real-time control software
Layer
Physical AI
Theme
TC
Trace Cohen
Early-stage VC & angel ยท Founder, New York Venture Partners
June 19, 2026
1 min read
ShareXLinkedInEmail
THE RUNDOWN
1

Robotics' hardest problems are increasingly in real-time software and performance, not just models

2

Deep systems-engineering talent migrating into robotics is a leading signal for the category

3

Smooth, reliable low-level control is the unsexy moat that separates demos from products

4

It reflects capital and talent flowing toward physical AI's execution layer

TC
The VC Read ยท Trace's TakeTrace Cohen

The robotics conversation is dominated by models, but the people who've actually shipped hard real-time software know the bottleneck is execution -- latency, control loops, the ugly edge cases that turn a demo into a product. So when a generational systems engineer leaves consumer software for robots, I read it as a signal about where the scarce, defensible value sits. Most robotics startups will nail the brains and faceplant on the boring performance layer. Back the teams who treat that layer as the moat, not an afterthought.

๐Ÿค– AI Landscape โ†’๐Ÿ’ฐ Funding Tracker โ†’

An engineer renowned for making the open-source VLC media player handle video playback smoothly across countless devices is now turning that performance expertise to robotics -- the low-level software that governs how machines perceive, decide and move in real time. The throughline is a craft most people overlook: squeezing reliability and speed out of constrained hardware.

The move highlights a shift in where robotics' hardest problems actually live. As foundation models get better at high-level reasoning, the binding constraint increasingly becomes execution -- the unglamorous real-time control loops, latency budgets and edge cases that determine whether a robot is a smooth product or a jerky demo. That's a systems-engineering problem as much as an AI one.

โ€œThe move highlights a shift in where robotics' hardest problems actually live.โ€

For investors tracking physical AI, deep low-level talent migrating into robotics is a meaningful leading indicator. The category's winners will need both the brains and the boring, exacting performance layer underneath -- and the people who can build that layer are scarce. When a generational systems engineer picks robotics, it's worth noting where they think the value is going.

ShareXLinkedInEmail

Originally reported by TechCrunch. Analysis and editorial commentary by Value Add Pulse.

โ† Back to Pulse

THE WIRE in your inboxโ€” Tech, startup & VC news with Trace's take. Free, no spam.

Read Next

FUNDING$100M launch round

Neo Launches With $100M to Secure Agentic AI Software

Three SentinelOne veterans launched Neo out of stealth with $100 million from a16z and Bessemer to secure the AI agents enterprises are now running against production systems.

FUNDING$30M Series A

Natural Raises $30M to Build Payments Rails for AI Agents

A 193-day-old startup raised $30 million from Forerunner to build payments infrastructure for AI agents, positioning directly against Stripe.

FUNDING~$6.2M JPYC investment

AZCOM Maruwa Pays Amazon Japan's Truckers in Stablecoin

AZ-COM Maruwa, one of Amazon Japan's primary logistics distributors, will pay roughly 2,300 partner trucking companies in JPYC, a yen-pegged stablecoin, marking the first large-scale corporate contractor payroll rollout of a stablecoin in Japan.

@Trace_Cohenยทt@nyvp.com