In this interview, Qasar Younis discusses the future of autonomous vehicle technology, contrasting Tesla's driver‑assisted Full Self‑Driving (FSD) approach with Waymo's fully driver‑less system. He explains that Tesla’s FSD still requires a human to monitor the road, while Waymo operates without a driver but is limited to specific geographic zones such as Phoenix and parts of San Francisco. Younis highlights the broader industry question of which model—geographically constrained, fully driverless services or personal autonomous cars owned by individuals—will dominate. He notes that personal car ownership will persist for decades due to vehicle lifespans, meaning a generational turnover before fleets become fully autonomous. The conversation shifts to the current stage of the industry: autonomous driving is no longer a research problem but an engineering and cost challenge, with the focus now on making the technology cheap enough for mass adoption. Younis also addresses societal concerns, stating that fears about government surveillance are misplaced; the technology itself does not prevent surveillance, and misconceptions about autonomy’s capabilities are similar to early misunderstandings of large language models. He draws parallels to historical technology rollouts—automobiles, traffic signs, and regulations—to illustrate that widespread adoption takes time, even as modern tech evolves faster.
Chapters
You are in the autonomous vehicle space and recently Tesla has been known for ramping up their autonomous vehicle availability with full self driving. I wanted to ask what is your view of the future technologically? What should be, what should we be excited and afraid of? Because this technology obviously enables a lot of cool things to happen. But do you envision a future where maybe this limits freedom or enables government surveillance, that sort of thing limits.
If the government wants to surveil you, they can. It's not the self driving car that's preventing them. The heart of fear is a misunderstanding, right? Once you start to learn, learn a subject well enough there is. And it's like when people get afraid of like large language models and you know, once you actually work with them, you realize their limitations in autonomy. Just, just a 2 cents there. The general autonomy ecosystem is falling into two broad categories. There's gross generalizations, but it's the Waymo way and the Tesla way. Tesla 12.3 incredible product. But it is not full self driving in the sense of it literally cannot operate without a human. You need to have a human there, eyes on the road, hands on the wheels. That is required. That's not the case with Waymo. You can have a driver out. The constraints roughly are Waymo is geographically constrained and they are over the next five, 10 years are looking to see how can we expand that geography as fast as possible to get the most addressable market. Right now you have to be in Phoenix or some parts of SF to experience Tesla. On the other hand, again gross generalizations or the OEM path, the passenger car, personal private ownership of the vehicle is the, the, the highway system works the link. You can go everywhere but you just can't do everything. And you have to be in the loop. The human has to be in the loop like you know, giving commands. So the fund the trillion dollar question is what is going to be the prevalent way? And I think you can make argument that's extremely heat. I think you make arguments both ways. I think personal car ownership is not going away anytime soon. When you buy a car people, a car's lifetime tends to be 15 years. So even if every car was an EV that was autonomous, it would still take a generation for all the old cars to actually come out of the system. So that's just the reality of the situation. In terms of risk. I mean I think we're now at a part in the industry for if we were having this conversation 2017, when we started the company, there was a big question like will self driving happen? It's definitely going to happen. It's no longer a research and development problem. It's moved over to the engineering side. Engineering is the full line is. It's a cost problem. It's an engineering cost. How do you get this to be as cheap as possible? So it's generally available. By the way, if you go back in technology, mobile and a bunch of other technologies, they appear the same way. Even automobiles are not generally available immediately. Mercedes invents the car in 1876. The first stop sign in the Octagon form is 1930 and then the first time it's the Department of transportation is 1960. So this is a long time for these things to emerge. You can say, well today technology grows faster, et cetera, but it's going to take some time for autonomy to get over there. But it is no longer a research and development problem. Problem like that's solved now we just got to make it cheap.
