McCord: Government AI predictive maintenance claims are largely 'false goods'
Cameron McCord · Sequoia Leads $75M Series B Into Nominal | Alfred Lin Joins Board · Sourcery with Molly O'Shea · Jun 12, 2025 · at 1:04:08
Cameron McCord, CEO and founder of Nominal, discusses the limitations of AI integration in aerospace and defense hardware testing.
0:00 / 0:25exact quote · 25.9s
▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark
“On the government side specifically, I think there's been a lot of like false goods that have been sold, predictive maintenance, prognostic, you know, type things that when you actually get down to it, like the data, the telemetry and the sensor data and the logs, like it's so unstructured and so messy and so disorganized that it's not really possible to actually, you know, do proper, you know, machine learning or just train inference in general on, on that data.”
quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →
More from Cameron McCord
McCord: Not enough defense program cuts for startups to win big programs
“When I look at the sort of overall budgetary picture, like there's just not, there's not a enough. Of these big program cuts, I think for all of the amazing new companies building out there to like really get the big program or record wins that they want over …”
McCord: Traditional primes do things no defense startup can do
“The primes play a like really key role in national defense. They have resources, they have capabilities. They can do things frankly, that. You know, and say essentially no other startup can do. And so we need to treat them accordingly in that way.”
McCord: Legacy Tools Like MATLAB and National Instruments Fail Modern Hardware Testing
“Scenario A, where you would use really old outdated software. Things like win plot people know, but a lot of MATLAB has used Python analysis, Python scripts, some of the national instruments, you know, tools, things like that to do modern test engineering. And…”
McCord: Software tools like Datadog and Grafana fail hardware testing
“Or you know, have modern technologies that were generally built in a software only world with like no consideration for hardware limitations or constraints. And so you might have new technologies like people use Tableau, right? Or like Looker or Grafana or Dat…”
McCord: Every deployed vehicle and drone is fundamentally a test asset
“Every asset, every drone, every plane, every, ah, you know, ground vehicle, every UUV is a test asset. Fundamentally, it's generating, you know, telemetry sensor data logs, just like you would a hardware in the loop, you know, test stand, or a test cell, or a …”
McCord: DoD is compressing hardware development lifecycles to 24 months
“The DoD is like shifting the way that it's thinking about a lot of these larger programs, trying to get away from 10, 12, 20 year program development life cycles. Like let's build the next exquisite, you know, bomber or exquisite plane or submarine. Right. Tho…”