Hello all! CEO of Voyage and former YC founder here. Would love to answer any questions I can about autonomous taxis. I'm obviously very passionate about the area. It feels like dev time is only accelerating (vs. linear progression), so I believe we'll see self-driving cars much faster than the skeptics think.
We're also hiring. I don't care about your credential, just that you love robotics or machine learning. Ping me! jobs@voyage.auto.
Hi there, I own a manufacturing company in Arizona, and we make (eseentially) brackets, and I would like to make brackets for your retrofit kits. I actually wrote a business plan out 3 years ago based on the likely idea that someone would build a self drive system but would need a partner to figure out how to make the bracketry to bolt the parts on.
The way I see it, the 1 billion vehicle fleet on the road now could be upgraded with a standard suite of sensors and actuators, but they would need unique vehicle specific brackets to bolt the parts on. We can help with that part! Most of our vehicles we cover are current model year, back to about 2000, and we work with the steering and suspension on late model vehicles everyday.
When will your product be available to test? My personal vehicle is a 09 Tahoe Hybrid, with electric steering and up to 28 mph electric only, purchased specifically to hack.
greg@traxda.com if you want to discuss
we make lift kits for pickup trucks, as well as LED lightbar brackets and air horn brackets, but it is really just making hundreds/thousands of parts to bolt into existing holes and make the truck do something a little different. We are in the SEMA Tech Transfer program, so we can get OEM CAD files, and we make everything in our robotically powered, vertically integrated operation in Tucson :)
So you guys literally milked ideas and and code in the guise of teaching the students only to spin up another business. I loved the whole MOOC idea until it was ambushed and hijacked by people running all manner of fluffy positions.
The whole system is now bent on profiteering with marketing staff fabricating fancy outcomes and guarantees to lure people to sign up for the paid version of the course.
Especially Udacity and Coursera. Edx not so much even though their programs are more rigorous!
> So you guys literally milked ideas and and code in the guise of teaching the students only to spin up another business. I loved the whole MOOC idea until it was ambushed and hijacked by people running all manner of fluffy positions.
As someone enrolled in the course, I don't think this is a very fair judgement. The TAs genuiunely seem there to help. I'm actually happy that Oliver decided to do this and wouldn't mind if more did it, more employment opportunities!
> The whole system is now bent on profiteering with marketing staff fabricating fancy outcomes and guarantees to lure people to sign up for the paid version of the course.
Oh, you mean like traditional universities? $2400 doesn't look so bad when you compare it to grad school somewhere that would teach this stuff.
Not at all! The open source self-driving car project is still open sourc and hasn't been commercialized in any way. The spinout involves no student-written code. Hope that helps.
Most likely they were just passionate about self-driving cars then, and are passionate about self-driving cars now. It doesn't mean the two were connected.
What are you (or anyone in the space) doing to advance the public's perception of the viability of self-driving cars?
Whenever I speak to laypeople about self-driving cars their mind immediately shoots to the 1 million+ ways that things can go terribly wrong, or even slightly wrong. Therefore, they have no interest in self-driving cars.
It'd be helpful if we could see when self-driving cars get there.
Like crash-safety tests soothe people's nerves, a battery of self-driving tests -- visual, visceral -- would silence objections. Something to point to... something tangible. A way to score the playing field.
Because between Google, Tesla, Uber, Cruise, Voyage, and whoever else, it's insanely difficult to see through the fog.
> Whenever I speak to laypeople about self driving cars their mind immediately shoots to the 1 million+ ways that things can go terribly wrong, or even slightly wrong.
When I raved about self driving cars to my wife, she said to me: look, Google Maps is messing up on a continuous basis (GPS errors, software errors, bad routes, mapping inconsistencies, underestimated ETA). If they can't make a decent GPS app, how could they make a self driving car?
I had no way to counter that. It's true, GPS is an old tech purposefully dumbed down for non-military applications. But how can we convince people of progress when it is so lacking?
A self-driving car, as proposed by Waymo and many others, in no way relies just on GPS. These beasts have fusion of many sensors with built in redundancy.
Technology improves and evolves. If it didn't you wouldn't be so reliant on Google Maps. We convince people of progress by a safe track record. Google maps may fail occasionally but more often than not it gets you to your destination at the estimated time.
Yes self-driving cars can't be "more often than not" safe. But starting with smaller areas and showing the difference it makes on people's lives and the environment can be a good start.
Sensors quality does have something to do with it, but regardless, you won't be able to get the accuracy and consistency from GPS you need for localizing a self driving vehicle.
Unless you get a DGPS unit (super expensive and require 3k subcription from satellites) you get north of 1m accuracy.
Great question! I think the biggest help would be statistical significance. We (as in the tech community) realize that machines can exceed humans in efficiency and safety, but the data that this is unarguably true does not yet exist. Waymo drives 1 million miles per year, but this is obviously dwarfed by humans driving.
A self-driving car company needs to be doing hundreds of millions of physical miles in order to quieten these fears, and it's only a matter of time before this is happening. This is one of the reason I think autonomous taxis are the perfect vehicle (pun intended) to bring autonomy to the world, since the business gets stronger the more miles you do.
Thanks for answering Oliver. I understand the statistical point of view (as a member of the tech community). However, I think that argument is falling upon deaf ears across the general population. You still get a "yeah, but..." with an uneasy shrug.
People don't like the idea of losing agency when things go sideways. While people may agree with the premise of a computer being better in extraordinary situations, it's hard getting over an unknown unknown. They don't trust the system. They trust themselves.
Elon Musk fundamentally understands that you must "capture the imagination" of the general public to advance a field. [1] He's done it consistently through SpaceX and Tesla, and I witnessed it first-hand on two separate occasions with two very different people.
For years, I would tell my father -- an older American muscle-car loving guy -- about Tesla and their cars. All the logical arguments. But his heart was set upon the Dodge Hellcat: one of the fastest street-legal cars ever released. That was, until the day the Tesla Model S beat the Dodge Hellcat in a drag race. From then on, he started talking about Tesla, calling me when he saw one, and talking with the owners. Tesla captured his imagination with a visual, visceral experience.
And so it was with my girlfriend -- a young, 20-something, not really interested in cars. I'd point out Tesla cars while we were driving and tell her all about them. The logic behind electric cars didn't interest her. Until one day, she called me, screaming into the phone, "Babe! Babe! I just saw this car... it... the doors, I don't know, they like went up like a bird! But it was an SUV!" She was over the moon excited. To which I responded, "That's the Tesla Model X." From then on, her dream car is the Model X. Tesla captured her imagination with a visual, visceral experience.
My gut feeling is that the company to win will be the one with statistical significance and a way to capture the public's imagination.
P.S. This is all assuming that you're trying to win as a consumer company. If your mission is just to succeed as a supplier, then disregard.
But you didn't do your dad any favors with the Tesla vs Dodge Hellcat comparison. The speed of the Tesla is really a parlor trick. I'd bet on the Hellcat every time I needed to do more than a few laps on any decent race track.
Thing is, I didn't show my dad the comparison (I didn't even know it existed until he told me). He came across an article independently and watched the video for himself. He's also much more concerned with acceleration -- the feeling of being pinned back in your seat -- than distance. So the quarter mile was enough for him.
and if you wanted a little history to back up your thoughts, Henry Ford had the 999 racecar before the Ford Motor Company was founded, and Soichiro Honda was crazy about racing as well and almost bankrupted the company in pursuit of it (some dispute the closeness of it). Both felt that racing captured the hearts and minds of the public, and gave the newspapers something to write about.
Assuming that the tech behind the company are all sound and working as intended, how does the company differentiate itself from competitors (both local and overseas) like Lyft (which is still dependent on humans) and others like Uber/Otto, Didi and a few others looking to add autonomous vehicle into their product offerings?
We're hiring for machine learning & robotics engineers. We like pre-mapping and LIDAR, and we like machine learning and cameras. We're focussed on the Bay Area right now, mainly since our team is so small.
I will be staying on as an advisor to Udacity, primarily to help with the self-driving car curriculum!
With a self-driving taxi what kind of challenges and opportunities do you see in terms of the passenger's request and ride experience (i.e. in the BI article there's a reference to using voice controls for music/navigation)
If you think of the things you'd ask a human Uber driver to do, we need to reproduce almost all of those digitally, whether through voice or just your mobile phone. Things like changing destinations, asking for ETAs, or turning up the AC.
Surprising indeed, but not entirely. Udacity co-founder Sebastian Thrun competed in the DARPA Grand Challenge, and led the development of Google's self-driving car.
They offer a nanodegree [0], so I'm not really surprised they want to utilise that expertise (lecturers, engineers, anybody that helped build the coursework) in more than one way.
We've been building an open source self-driving car alongside providing an amazing self-driving car curriculum. We've found that the two equate to one of the best educational experiences on the planet!
Will there be an employment link up between Voyager and the self-driving car nanodegree? And do you think it's likely that other spin offs will occur from Udacity related to other courses?
I work in higher education and I'm incredibly grateful for how Udacity are opening the eyes of my more traditional colleagues as to how learning can be different, relevant and exciting.
I'm currently enrolled in Udacity's self-driving car nanodegree, I can certainly say that the program thus far has been very fun and good at conveying knowledge. They've developed a curriculum that has a great cadence that keeps you moving forward, and the content is super interesting.
Congratulations to the team involved! In particular Oliver, Eric and Mac.
As a CarND student who enrolled because of the Open Source aspect I'd love to know if Voyage is planning to contribute open source software like Comma.ai and Polysync.
Can the billions being invested in this space earn an economic profit? It seems to me that multiple companies will be able to bring this technology to market in roughly the same timespan and that number will increase rapidly with time, driving commodification.
We're also hiring. I don't care about your credential, just that you love robotics or machine learning. Ping me! jobs@voyage.auto.