they will raise billions in capital but will never show a positive cash flow

10/4/2017 1:06:44 PM

https://www.npr.org/2018/08/21/639646651/watch-self-driving-cars-need-to-learn-how-humans-drive

8/25/2018 11:05:20 AM

https://gizmodo.com/tesla-autopilot-malfunction-caused-crash-that-killed-ap-1834453661

5/1/2019 10:58:12 PM

I don't believe in a future.

https://blog.piekniewski.info/2018/05/28/ai-winter-is-well-on-its-way/

Quote :

"But this narrative begins to crack. And as I predicted in my older post, the place where the cracks are most visible is autonomous driving - an actual application of the technology in the real world."

5/2/2019 9:51:52 AM

he makes some good points (aside from the ridiculous tweets/day metric as a measure of deep learning advances)

this stuff is very hard of course. but the industry has undeniably made a lot of progress, and I still believe we'll get there, even if it requires a detour away from current methods or plays out in the form of incremental improvements from today's technology rather than a giant leap to full autonomy. this feels like a 90-90 rule thing - the first 90 percent of the code accounts for the first 90 percent of the development time. The remaining 10 percent of the code accounts for the other 90 percent of the development time

5/3/2019 10:22:50 AM

Quote :

"Within the next 5-10 years, I hope to transition my career to pursuing this vision."

Almost 9 years from then, is this your career now?

5/3/2019 3:54:03 PM

I think Ken went into education software.

5/3/2019 4:28:27 PM

Thanks, but Mr. Warner may not have wanted that made public.

5/3/2019 5:00:40 PM

Dude has a twitter feed.

5/3/2019 5:14:00 PM

Quote :

"2018/05/28"

5/3/2019 6:20:32 PM

I started the journey a while back, but I don't work in that industry at the moment despite a few attempts to transition to it

In 2011 I took two online classes from Stanford professors, Intro to AI and Machine Learning, both of which enrolled tens of thousands of students from all over the world. The success of those two courses led the instructors to spin them out into new startups, Udacity and Coursera, respectively.

Udacity's first course was Artificial Intelligence for Robotics, with a specific emphasis on self-driving cars, which was amazing considering the founder and professor Sebastian Thrun was the lead on Stanford's winning entry in the 2005 DARPA Challenge and by this time in 2012 was Google's head of research for their self-driving vehicle technology program. I was very active in the course, answering questions on the forums and whatnot, completed the course with highest distinction and was recruited by the staff to start writing class notes for the second cohort of students. After the class completed I interviewed with Udacity for a job, but ended up moving to NYC to work for Kaplan Test Prep instead and for the next few years focused on that career for the most part, while keeping an eye on the developments in autonomous vehicle tech.

In late 2016 Udacity introduced a new Nanodegree program in Self-Driving Cars, to which I applied and was accepted. One of the key points for me was by now most car companies had started paying attention to self-driving tech and had begun investing in it as well. And Udacity had started to partner with those companies to form a hiring network through its Nanodegree program, so if I did well in the course, I would be reasonably sure of finding a job in the field despite the fact that most of my career at Kaplan and before had been focused on building products on the web. Plus, the machine learning technology that had been driving most of the progress in that field had matured to the point where people other than PhD's could meaningfully contribute.

I started the course, but pretty much all the jobs at that point were still in Silicon Valley, and I wasn't interested in moving there at the time, for reasons, so I stopped the course and instead at the beginning of 2018 I took another job in NYC with a chatbot startup, so I could at least transition to getting some daily hands-on experience with machine learning technologies. But that company ended up being pretty trash, plus I ended up having to move back to NC last summer anyways, so I took a job with another education startup I had connections with and I could work remotely from NC.

so no, I have not yet accomplished this 5-10 year goal :-|

/blog

5/3/2019 7:37:14 PM

https://www.redhat.com/en/open-source-stories/road-to-ai

5/4/2019 9:16:23 AM

Automated cars will never be full proof. I've watched a lot of Air Disasters episodes on Smithsonian Channel lately. Multiple episodes were about the same model of Air Bus that failed. It was pretty much designed to fly itself, but there were still crashes. Nothing is 100% fail safe. At least with an airplane at 35k feet you have time to react during an unplanned pitch/altitude change. You don't though when your car is going around a curve at 70 mph in the mountains. At that point you drive off the cliff and die.

5/4/2019 12:44:47 PM

Shut up idiot

5/4/2019 2:09:16 PM

B4C raises a good point. cars will always have some limit to their cargo capacity. if you put enough stuff in even the biggest car it will eventually get full, whether or not it requires a human driver. no automobile is full proof.

5/4/2019 2:56:10 PM

I believe in a future where I drive my own car and run into driverless cars for fun.

5/4/2019 3:12:18 PM

Quote :

"Automated cars will never be full proof."

They just have to be better than people. It's a low bar.

5/6/2019 11:23:05 AM

^ they actually have to be better in every situation and every change from one situation to another.

https://mindmatters.ai/2018/10/self-driving-vehicles-are-just-around-the-corner/

Quote :

"The first problem begins when we switch the requirement for autonomy from handling a few identified driving situations to handling all of them. Drivers behave differently in different cities. Once, when I was in New York City, for example, the drivers just spontaneously decided to organize themselves into five lanes on a four-lane road. My years in programming have taught me that the number of real-world situations that can arise is significantly greater than the number of situations I might forecast just by thinking about a problem."

[Edited on May 6, 2019 at 11:35 PM. Reason : A]

5/6/2019 11:35:41 PM

Feel deep learning and neural networks were expected to be the breakthrough, but turns out we probably need a Manhattan project level thing to get get level 5. Don't think deep learning and statistics is going to do it.

You're basically trusting your life to the photoshop magic wand tool at this point. Figure out the things you do automatically in anticipation of other vehicles, seeing a slowdown in front of the car in front of you by the wind shield. Tires turning, turn signals, people turning the right turn signal then cutting across and turning left, stickers on stop signs.... You can see and react and adapt. None of the self driving cars can do any of that nor are they close to doing it.

[Edited on May 6, 2019 at 11:43 PM. Reason : A]

5/6/2019 11:39:55 PM

still just around the corner?

3/23/2021 12:44:12 PM

I no longer believe in a future

3/23/2021 12:51:26 PM

Cruise is now testing driverless rides in Raleigh.

https://twitter.com/Cruise/status/1694052381770805494

8/22/2023 3:23:22 PM

but if cars become automated and have super sensors and stuff, how will we run over the bad guys and crash cars into them like in the movies?

8/22/2023 4:37:47 PM

Hopefully they're training the car AIs on a ton of '80s action movies to cover that

8/22/2023 5:23:57 PM

^^^deserves its own thread so people can post instances where it gets confused in normal circumstances

8/22/2023 7:18:32 PM