Lightspeed’s Raviraj Jain on Future of AI & Automation was originally published on Springboard.
We recently sat down with Lightspeed’s Raviraj Jain to talk about trends in the AI and Automation sectors. Here are some highlights from the talk.
In this fireside chat, he discusses:
- Lightspeed’s investments in the sectors
- Trends worth watching
- How to prepare for the careers of tomorrow
- Aurora Innovation – full-stack software driver for self-driving cars
- AnyVision – one of the world’s leading face recognition technology — uses computer vision for speed face recognition
- Dexterity Robotics – robots that work in an unstructured environment for logistics, warehousing, and supply chain operations
- Fortress IQ – computer vision for mapping business processes
PC Magazine says that “Aurora is the most important self-driving car company you’ve never heard of.” Why is that?
Two reasons: the stellar leadership team and the unique product offering.
1. Team with Tremendous Sector Experience
Aurora was founded by three self-driving car sector leaders.
- Chris Urmson has been in the industry for 15+ years; founding member & CTO of Waymo, the Google self-driving program. He was previously with DARPA at CMU.
- Drew Bagnell previously led a large part of Uber’s advanced technology group, which is the self-driving car unit. He’s also a professor at CMU in their Robotics Institute.
- Sterling Anderson led the launch of Tesla’s autopilot and Model X.
Together, these three people have experiences of working on the cutting edge systems of self-driving cars, actually shipping anonymous software, and – in the case of Sterling Anderson – leading teams across various iterations of self-driving cars.
2. Product: Intel Inside of Self-Driving Cars
Instead of building a full car, they are building a software driver. So think of them as the Intel Inside of self-driving cars, except its not hardware, it’s software. So any car they envision that actually drives by itself can someday have Aurora software in it. Most of the car companies do not have the expertise to actually build software and that’s where Chris & team are trying to supplement that.
There are a lot of other technical reasons why they have been able to accomplish a lot in such a short time. But largely in a more simplistic way it is driven by their experience and leadership in the space for many, many years.
What’s Arora’s core product offering right now?
In the self-driving space, most companies are still in research and development and they don’t have a product out yet. But the vision is to launch software that any car manufacturer can actually deploy in their cars to turn them into self-driving cars.
What are some of the implications of self-driving cars we may not have thought about yet?
A lot of times we think, “oh, well self-driving car means I go buy a car at the dealership that will drive me back home.” Well, that can happen. There’s a lot more that can happen.
- Change the phenomena of people actually buying cars. Only 4% of the time our cars are actually on the streets. 96% of the time, they’re in our garage. So car ownership is pretty inefficient these days. People will start using shared transportation a lot more. It just makes sense. If you think about your cost of Uber right now, 40-50% of that is actually the drivers cost. So if you take that away, suddenly transportation on demand becomes much, much cheaper. So that’s first, car ownership will go away. And people might subscribe to a car which will give you access to unlimited rides within 50 miles. So I think that’s the big shift that’s going to happen.
- Change how cities are designed. There is a pretty significant portion of our city that is basically parking lots. And if your car drives you and drops you somewhere you don’t necessarily need to have those parking lots. And that fundamentally changes a lot of things about the city. Makes up more spaces, there will be more greenery, there’s a lot more space for people to move around.
- Limited Organ donations. Globally there are more than a million people, unfortunately, who die from road accidents. In the world of self-driving car, the belief is that the accuracy will increase meaningfully as a result, road accidents will decline purposefully. Because people are gonna be safer, there could be a shortage of organs.
Why do you believe a fully autonomous future is 20-30 years away (and not sooner)?
It’s first a technological barrier and the second is a production barrier.
The technological barrier is, honestly, it’s a very, very complicated problem. Because safety is the highest priority, cutting-edge technology can work 99% of the time. But in order for self-driving cars to be mass transportation, it has to work 99 followed by 9 nines. That level of accuracy means that the car should be able to navigate every possible scenario, and that’s hard.
There are 3 key elements of self-driving:
- Perception – the car should be able to perceive what’s coming in
- Planning – the car should be able to plan their response in run time based on the perception
- Actuation – the planning should be able to actually actuate the car to make changes
The biggest challenges are in planning and perception still. Planning will tend to be the problem overall because cars are not as smart to be able to understand every single scenario. So I think those are the reasons, that’s why the best way to make, bring self-driving cars to market at scale is actually bounding the problem. Limiting the different kinds of things that can happen to a car, by fixing a specific route, or driving in a no-traffic zone or in feeling.
Learn more about Raviraj’s thoughts on self-driving cars, robotic dexterity, and how to prepare for future jobs here.
Springboard is an online school that teaches job-ready skills in fields such as Data Science and UX Design. Check out our Career Track programs —you’ll learn the skills and get the personalized guidance you need to land the job you want (or we’ll refund your tuition).
The post Lightspeed’s Raviraj Jain on Future of AI & Automation appeared first on Springboard Blog.