If you've ever wondered what AI pioneers do during their morning commute, here's your answer: they're having lengthy conversations with chatbots. Andrew Ng, one of tech's most influential AI researchers, revealed at the Masters of Scale Summit 2025 that some of his most productive work happens while driving, using voice prompts to bounce ideas off AI models in real time.
Turning Drive Time Into Collaborative Thinking Sessions
"When I'm driving, I talk to AI quite a lot," Ng explained, noting that even his friends underestimate how much he relies on AI-assisted thinking. This isn't about asking Siri for directions or having a quick back-and-forth with ChatGPT. Ng treats these interactions as genuine collaboration sessions, cycling between multiple chatbots to leverage their different capabilities.
For coding challenges, he turns to Claude Code and OpenAI's Codex. For exploring broader conceptual territory, he switches to other models. The key, according to Ng, isn't just firing off a question and waiting for an answer.
"AI is very smart, but getting context in is difficult," he said. "A lot of it is not, 'Let me say some stuff, then give me ideas.'"
Why Long Conversations Beat Quick Prompts
The real value comes from extended, iterative exchanges where Ng guides the model through his thinking process and responds to its suggestions. It's less like querying a search engine and more like working through a problem with a colleague who needs some hand-holding but occasionally offers surprising insights.
When he reaches his destination, Ng asks the AI to summarize their entire conversation and forward it to his team. It's an efficient workflow that turns otherwise idle time into productive brainstorming sessions.
That said, Ng acknowledges there's also a place for the lazy approach. "It's sometimes faster to be lazy and dash off a quick, imprecise prompt and see what happens," he said earlier this year. "Most LLMs are smart enough to figure out that you want them to help."
The Bigger Picture: Where AI Is Headed
While Ng focuses on practical applications, the broader AI industry is grappling with fundamental questions about direction and safety. Leaders in the field are offering sharply different takes on what comes next.
On Tuesday, Ilya Sutskever, co-founder of OpenAI and Safe Superintelligence, argued that AI development has pivoted back toward research-focused progress. The reason? Model scaling has hit its limits. Simply making systems bigger no longer guarantees transformative improvements, according to Sutskever.
Analyst Dan Ives takes a more optimistic view, suggesting the AI market is still in its early growth phase. He points to strong momentum at major tech companies like Alphabet Inc. (GOOG) (GOOGL), Broadcom, Inc. (AVGO), and Apple Inc. (AAPL) as evidence that AI continues driving significant value for the industry's biggest players.
Then there's the existential risk debate. Mustafa Suleyman of Microsoft Corp. (MSFT) warned against pursuing autonomous superintelligence, calling it dangerous and misaligned with human values. Instead, he advocates for "humanist superintelligence" designed to amplify human judgment rather than replace it.
For now, though, Ng seems content to keep having his car conversations, iterating through ideas with AI models one commute at a time.