Sep 10, 2025 · 1h 4m · big-technology
Booz Allen CTO: AI in Government, Autonomous Driving, Quantum's Promise
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of the Big Technology Podcast, host Alex Kantrowitz interviews Booz Allen Hamilton CTO Bill Vass to explore the transformation of federal technology, covering generative AI deployments, government contracting reform, autonomous robotics, quantum computing, and strategic lessons from Silicon Valley.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Alex holds 26.5% of the talking time here. How this is scored →
Alex as informed peer 4.3 Guest teaching 6.1 Guest disagreement 3.0 Alex pushing back 2.5
speaking balance: gold is Alex, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 13:18 Vass rejects narrative of backwards government technology
Vass forcefully rejects the host's framing that government tech is universally obsolete and stuck in the 1990s, challenging Alex directly by pointing out that GPS and Mars Rover tech are federal innovations.
Hardest push from Alex ▶ 26:08 Kantrowitz refuses calculator analogy using Microsoft researchAlex refuses Vass's simple rhetorical comparison between LLMs and calculators, citing specific cognitive research on critical thinking atrophy and navigational dependency.
Biggest teaching moment ▶ 50:00 Quantum Hamiltonian simulation capability comparisonVass completely re-educates the host on quantum utility by calculating that simulating ammonia molecules would take all classical computers longer than the universe's lifetime, but only three minutes on an error-corrected quantum computer.
Alex holds their own ▶ 12:20 Kantrowitz details firsthand Capitol Hill CRM dysfunctionAlex demonstrates concrete domain knowledge by recounting his firsthand experience with outdated congressional CRM systems like IQ to challenge Vass on government tech modernization speed.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Alex as informed peer | Guest teaching | Guest disagreement | Alex pushing back | Why |
|---|---|---|---|---|---|---|
| Government Redundancy and Stovepiped Agency Systems | 3 | 5 | 3 | 2 | Alex asks about government system redundancy based on his experience in NYC quasi-governmental work. Vass explains the reality of stovepiped agencies and quotes Bezos's philosophy on redundancy before consolidation. | |
| DOGE Initiatives and Contracting Model Reforms | 4 | 6 | 2 | 2 | Alex asks about DOGE's centralization efforts, prompting Vass to educate him on the mechanics and trade-offs of firm fixed-price versus cost-plus/time-and-materials contracts. | |
| Data Centralization Feasibility and Archival Preservation | 3 | 7 | 4 | 3 | Alex brings up the sensationalized story of veterans' records stored in caves, and Vass directly corrects the premise by detailing federal archival legal requirements and optical character recognition formats. | |
| Dispelling Government Legacy Tech Myths and Political Friction | 5 | 7 | 5 | 4 | Alex challenges the guest with the perception that government technology runs on obsolete systems like Windows 95, citing his Capitol Hill experience. Vass counters with foundational government inventions like GPS and DARPA architectures, clarifying that legacy tech exists in private industry as well. | |
| Practical Deployments of Generative AI Across Federal Agencies | 4 | 6 | 2 | 1 | Alex prompts Vass to detail how LLMs improve government workflows. Vass delivers an extensive breakdown of real deployments, from Llama on the ISS to claims processing at the VA and end-to-end 2D vision navigation. | |
| Evaluating AI Reliability, Cognitive Reliance, and Hallucinations | 6 | 5 | 5 | 6 | Alex pushes back hard using arguments from AI critics regarding brain atrophy and high stakes on space missions, rejecting the simple calculator analogy with Microsoft research. Vass defends augmentation via direct manual citations and contextual search. | |
| Vision for AI Public Services and Computing Cost Constraints | 4 | 5 | 2 | 2 | Alex asks about the aspirational vision for AI in public services and flags major Saudi sovereign AI infrastructure investments. Vass outlines compute costs and ROI metrics before reflecting on his time launching AWS in the region. | |
| Applying Amazon Leadership Principles to Government Operations | 4 | 5 | 2 | 2 | Alex asks how Amazon's leadership principles map onto federal leadership. Vass analyzes bias for action, customer obsession, and dive deep in the context of federal technology management. | |
| Autonomous Driving Simulation and Real-World Edge Cases | 5 | 6 | 3 | 3 | Alex asks about self-driving timelines, comparing Waymo's approach with Tesla's full self-driving. Vass gives an expert breakdown of NVIDIA Omniverse synthetic simulations and edge cases like human gesture signaling. | |
| Humanoid Robotics and US-China Technological Competition | 4 | 6 | 4 | 2 | Alex raises the conventional wisdom that the US is lagging China in robotics following viral race clips. Vass rejects the premise of US inferiority while emphasizing China's long-term state-directed investment as a genuine pacing threat. | |
| Quantum Computing Frontiers, Error Correction, and Post-Quantum Security | 4 | 8 | 2 | 2 | Alex brings up Jensen Huang's skepticism regarding quantum timelines, leading Vass to deliver an in-depth masterclass on physical versus logical error-corrected qubits, Hamiltonian molecular simulations, and post-quantum encryption deadlines. | |
| Strategic Lessons from Sun Microsystems and Managing Tech Transitions | 5 | 7 | 2 | 1 | Alex asks about the historical cautionary tale of Sun Microsystems and its lasting lessons. Vass provides a first-hand post-mortem detailing Sun's failure to navigate the capital-to-cloud transition and hesitation to open-source Solaris early. |
Statements from this episode (27)
Booz Allen Hamilton employs 22,000 engineers and 3,000 generative AI experts
“Booz Allen used to be a business consulting company, and they sold that off in 2008, and now they have 22,000 engineers, about 3000 AI, Gen AI experts, and about 8600 cyber experts.”
Jeff Bezos believed two redundant systems were better than zero at Amazon
“Jeff Bezos at Amazon used to have this saying that two is better than zero. So he, we would have redundant systems at Amazon, but then work to consolidate them over time.”
Intelligence agencies deliberately built opposite systems to avoid overlapping efforts
“When I used to work in the IC, one agency would sometimes do the opposite of the other agency just to avoid overlapping.”
DOGE's push for firm-fixed-price federal contracts benefits taxpayers and delivery
“And the shift back to outscum based that, that Doge is pushing is a really good thing. In my opinion, it's good for the taxpayers. It's good for delivery.”
Silicon Valley's ecosystem is built entirely upon government-funded foundational technologies
“I mean, look at all the things that came out of the government, integrated circuits, the internet, GPS you know, it goes on and on and on. Those came from government programs. All of Silicon Valley is built on top of it, right?”
The federal government will consolidate fragmented citizen payment and tax systems
“You know, I think there's a lot of places where citizen services could be much better. Through consolidation, making it easier to do your taxes, easier to make payments easier to get payments from the government those kinds of things. And I think you're going …”
US government purposefully archives records underground on paper to outlast tech cycles
“There are, there is the need for long-term storage at NARA and other places like that of data that is underground and that data is also stored purposefully on non-electronic formats, and the reason for that is we have legal requirements, but the government to …”
Bill Vass often sees more cutting-edge technology in government than Silicon Valley
“I see more cutting edge technology in the government often than I see in Silicon Valley and having been to Silicon Valley a lot as well.”
Bad technology decisions in business and government are always driven by politics
“The way I view this in corporate, it's very true. And in government is very true. Whenever you see a bad technology decision, it's always politics.”
Booz Allen deployed Meta's Llama to the ISS for zero-latency astronaut troubleshooting
“So we just put Lama on the international space station on the edge on satellites, right? So that, that enables the astronauts who are working on international space station to have Lama to chat with in space with no latency to determine when things go wrong, h…”
LLMs reduced Veterans Affairs claims research time from hours to seconds
“We have large language models helping the VA do claims processing. What used to take many hours for a person researching on the claim process happens in a few seconds through The use of a large language model.”
Amazon machine learning models achieved 98% accuracy identifying cancer from scans
“At Amazon, we did a lot of like cancer identification from MRIs and CAT scans. And, you know, the ML was about 98% accurate, which is tremendously good.”
Amazon successfully used manual-referencing LLMs to debug fulfillment centers without hallucinations
“We use large language models at Amazon to help debug things in our fulfillment centers. And it was very successful in those areas, but you still have, it references the manuals directly so you can avoid hallucination. You can see what it actually, you know, fo…”
Concerns that AI tools will cause human cognitive atrophy are overblown
“I think the, just like any tool you know, you can cut yourself with a knife in the kitchen. You know, you don't have to tear things apart with your hands. Right. It's, you know, oh geez, we've lost the talent of tearing things apart with our hands cause we've …”
High GPU training and inference costs raise ROI concerns for generative AI
“These models use a lot of GPUs. They are really expensive to train and they are really expensive during inference on today. So that's another area that, that we really question sometimes the ROI of some of these things because of the cost of all of it.”
Saudi Arabia is building significant tech brain trust to diversify from oil
“I think there's a tremendous amount of brain trust happening in Saudi Arabia and investment there in their movement to technology and their movement to you know Diversify their oil investments into other areas.”
The US government lacks sufficient customer obsession in delivering citizen services
“Customer obsession is one of my favorites at Amazon. I don't think the government is as customer obsessed as it should be, and they need to be thinking about that in citizen services. And I think that's an area that, that, that could be improved.”
Tesla Full Self-Driving cannot be fully trusted yet without risking trouble
“I have two Teslas and I play with a full self-driving all the time. It's entertaining, but I wouldn't trust it. Entirely, right? If you trust it, you're going to be in trouble. So it's not a hundred percent there yet.”
Real autonomous driving will start to emerge over the next five years
“I think you'll start to see real autonomous driving over the next five years. You know, maybe I'll be burned by that prediction.”
Crowdsourced data from millions of users gives Tesla a major autonomy lead
“Remember, Tesla has this advantage very much like the Echo devices at Amazon, where they're able to crowdsource training from the users. So basically, they're learning and training their model based on all the millions of people driving Teslas every day. That'…”
China is neither ahead of nor behind the US in robotics technology
“I don't think China is ahead, but I don't think they're behind.”
Bill Vass left AWS over concerns about government AI adoption trailing China
“One of the reasons I left at AWS and I loved being at AWS. I worked on 63 of the services there and built a lot of them myself, worked on quantum computing and robotics and a whole bunch of things is I was worried a little bit about government adoption of AI a…”
Achieving 1,000 error-corrected logical qubits will require roughly 7 million physical qubits
“The goal would be to get to a thousand error corrected qubits, right? But just from that perspective, that's going to be around seven million physical qubits to do that.”
The first 100 error-corrected qubits on fast architectures will arrive by 2028
“I think you'll see in 20, 27, 20, 28, the first hundred air created qubits on fast machines.”
Quantum computers will have enough qubits to break standard encryption around 2040
“In around 2040, we think there'll be enough qubits to start to break encryption, and secrets last longer than that.”