Kevin Surace is a Silicon Valley inventor and technology executive associated with Appvance, a company focused on AI-driven software testing. He is also known for building an early voice-driven virtual assistant in the 1990s. In his appearance on The Tech Visionary Show, Kevin Surace talks about AI co-pilots, software testing, driverless cars, jobs and the limits that GPUs and electricity place on AI. This article summarizes what he said and what it means for engineering leaders.
You can watch the full conversation on YouTube. Everything below is attributed to Kevin as he stated it in the episode. His forecasts are his own opinions, and none of them is a guaranteed outcome.
Who Is Kevin Surace?
Kevin Surace describes himself through several roles: CEO, CTO, futurist and inventor. He is an inventor with a long patent record who has spent his career building products where software meets everyday problems. On the show he connects three threads: the first conversational virtual assistant he worked on, today's AI co-pilots, and where automation is heading.
The quick facts a reader searching for his name usually wants:
- Known for: early voice-activated virtual assistant work and AI-driven software testing.
- Company association: Appvance.
- Where to hear him: The Tech Visionary Show, on the Tech Visionary YouTube channel.
- Main themes: joy as a driver of innovation, AI co-pilots, software quality, jobs and productivity, and infrastructure limits.
Kevin Surace on Why Joy Equals Success
Kevin opens with a personal point: joy equals success. His definition of success has little to do with trophies or profit and more to do with the work itself and what fuels creativity.
"I wake up every day with joy," he says, and he separates joy from happiness. For him, joy is the knowledge that today offers another chance for something extraordinary to happen. He also warns about negativity and argues that work carries fulfillment however small the effort, because there is always some point of success to reach.
For technology leaders, the practical reading is that a positive mindset makes people more willing to take on new problems and attack them with more imagination. Technology teams meet new obstacles constantly, and that attitude affects how they respond.
Kevin Surace and the First Virtual Assistant
In the 1990s, long before Siri, Alexa and Google Assistant existed as products, Kevin was working on what he describes as the first virtual assistant. By his account, it did not start as a cool technology idea. It started with a real problem: people received email from around the world and could not read it while driving.
His framing was simple: he could move his mouth but could not hold or pick anything up. His team built "Mary," a conversational assistant that handled calls, emails and even faxes by voice. Phones of the era had small screens and limited features, so a voice interface was the practical answer. "Phones back then had such small screens you needed a voice user interface. Everything stemmed from that," he says.
Kevin presents this work as an early version of the ideas behind voice interfaces, AI-driven responses and task automation. He sees those ideas living on in the assistants people use today. The lesson he draws is that durable technology starts from a concrete problem, and safety and accessibility were part of that problem from the beginning.
Kevin Surace on AI Co-Pilots, Testing and Jobs
The co-pilot phase
Kevin places AI today in a co-pilot stage. Tools such as ChatGPT and GPT-4 mostly assist a human operator with specific tasks in writing, customer support and other areas, rather than acting fully on their own. "AI helps us shed the tasks we don't want to do," he says, and he expects people to spend more time on higher-level, more strategic work as machines absorb repetitive jobs.
What changes in software testing
Software testing is the area he knows best. Testers have traditionally written scripts, run them and chased bugs by hand, which takes a lot of time. Kevin argues AI can now do much of that work faster, and he says plainly that AI can perform these tasks "much faster and better" than people. He points out that testing employs a very large global workforce, which is why the shift matters.
Jobs: displacement and new roles
Kevin acknowledges the downside. Automation brings efficiency and lower costs, and it raises legitimate concerns about job displacement. He stays optimistic and points to precedent. The Industrial Revolution moved work away from agriculture, and the computer changed office work, yet each shift produced new roles elsewhere. His view is that AI will likely follow the same pattern. That is a historical argument, not a promise about any specific job market.
Kevin Surace's Prediction on Driverless Cars
On transportation, Kevin makes his boldest forecast: in about 50 years, a person behind the wheel may be prohibited in certain places, particularly dense cities with heavy traffic. He imagines driverless vehicles reducing accidents and time lost on the road. It is a prediction, and he offers it as one.
He points to the fact that Waymo already operates driverless taxi service in cities such as San Francisco, which makes the idea less abstract. He also respects people who love to drive. "There will be areas where you can drive to your heart's content," he says. He expects a point where many people will say, "I actually want the AI to drive," especially as reflexes slow with age.
Kevin Surace on Trust, Testing and the CrowdStrike Outage
Trust is where Kevin sounds most cautionary. As AI moves into daily operations, mistakes and failures have wider reach. He cites the CrowdStrike software update, which he says went out without adequate testing and caused widespread disruption.
His comment on the episode contrasted a small saving on quality assurance with a cost to the world that dwarfed it. His broader point is that skipping testing is a false economy. Without it, failures can spread to airlines, banks and other systems that people depend on.
His rule of thumb is blunt: "No company should be touching the kernel without a large amount of testing." He expects AI to improve over time, but he argues that companies need their own safeguards in the meantime. As AI takes on more responsibility, oversight matters more.
Kevin Surace on GPUs, Energy and the Productivity Question
Asked about the biggest challenge in AI, Kevin points to infrastructure.
- GPU supply. "There's not enough GPUs, which is a type of processor that's needed for AI, and they're very, very expensive. Most of them come from NVIDIA." In his view, dependence on one main supplier creates a bottleneck that limits how fast AI can scale as demand grows.
- Electricity. He says AI data centers need very large amounts of power. To make the point vividly, he says he would want a nuclear power plant next to each operation center. His underlying concern is that without gains in energy efficiency, AI growth could stall.
- Expectations. Kevin warns that businesses may overestimate how quickly AI will lift productivity. He notes that executives who have tried generative AI are now asking to measure productivity gains against what they spend, and he cautions against expecting immediate results.
His longer-term argument is economic. "The more productive a company becomes, the lower the cost of goods and services, which increases demand and creates more jobs." He also says, "We've been relying on technology to run our businesses for hundreds of years," and he treats electricity, the internet and now AI as the same kind of productivity step. This is his reasoning about how the economy tends to work, and the outcome depends on how companies and governments respond.
What Kevin Surace's Ideas Mean for Engineering Leaders
Three ideas from the conversation are worth acting on, and they match what we see working with engineering teams across Latin America.
- Treat quality assurance as part of AI delivery, not a cost to trim. Kevin's CrowdStrike example is a case for testing capacity that scales with release speed. Dedicated nearshore QA and automation engineers working in your time zone can run that testing alongside your developers.
- Start with a real problem. His virtual assistant began with drivers who could not read email. Before funding an AI initiative, write down the specific task it removes and how you will measure the change.
- Plan for infrastructure limits. GPU supply and power are budget and architecture questions. Cloud design, workload scheduling and model choice all affect how much compute a project needs.
Kevin Surace's message is to move forward with intention: keep testing new technology, confirm it solves a real problem, and remember that problem-solving and creativity are what automation depends on. If you are building AI or cloud capability and want engineers who can handle testing, cloud architecture and delivery in your time zone, talk to Bydrec about a nearshore team.
For more of Kevin Surace's thinking, watch the full episode and subscribe to the Tech Visionary YouTube channel for future conversations.
