SoylentNews is people
SoylentNews
Uber is freezing hiring for software engineers and product managers across its US and Canadian workforce, the company acknowledged to Bloomberg on Friday. The shift was reported by Yahoo earlier in the day. The freeze does not apply to Uber's autonomous vehicle and freight shipping divisions.
The news comes a day after Uber reported second quarter operating losses of $5.4 billion—a new record for the company. That figure exaggerates Uber's quarterly burn rate because it includes more than $4 billion in one-time charges related to Uber's initial public offering. Still, excluding IPO-related charges still leaves around $1.2 billion in operating losses, worse than the $1 billion the firm lost in the first quarter.
Uber recently laid off 400 marketing workers. According to Yahoo, Uber employees are worried that this could be a prelude to broader cuts as the company's struggles to stem its losses.
(Score: 3, Informative) by Ethanol-fueled on Sunday August 11 2019, @01:06AM (8 children)
As an industry insider, I will tell you this: Self-driving is big, especially for the sensors that make it happen. The method that most interested parties are using is radars (I've posted better details here before) but there are a lot of issues that are still being worked out related to those radars. I can't go into details, but what I can tell you is that Boston Dynamics, my employer, is working on the small details that Uber and others use as band-aid fixes to prevent reflections from the car metal and reject sidelobes.
When you think self-driving, and even see my own employer's robotic success, you think LIDAR or computer vision. The problem with LIDAR is its dependence on moving parts, which are a point of failure; and the problem with computer vision is that it projects 3-D space into a 2-D plane, which makes confounding anomalies a lot more lethal.
As an industry insider, I will tell you this: being a first adopter with the latest model of iPhone can be inconvenient, being a first-adopter with self-driving cars will send you to your death. You wouldn't believe the discussions we all have laughing about how that Tesla steered its owner under a semi at highway speeds, or how that Uber car killed the pedestrian in Arizona.
(Score: 1, Insightful) by Anonymous Coward on Sunday August 11 2019, @02:32AM (3 children)
The tesla idea is not terribly far fetched if you stop and think about it. I have 2 eyes. It is basically 2 spherical 2d arrays mapped over time into a fading memory array with continuous input that my brain makes sense of. If my neck does not hurt I can move my eyes effectively 180 degrees vertical and horizontal with fine motor control at about 100 degrees.
So I think to get it to work correctly we would have to get rid of the idea of frame rate, sample rate, bits per pixel, and refresh. Computers are not analog though so it could be a lot of an issue. Which is where those odd errors come from at the root of it. The other techs will have similar issues of aliasing but in just different ways.
(Score: 0) by Anonymous Coward on Sunday August 11 2019, @04:23PM (2 children)
Yes and our eyes regularly deceive us. Why bother replicating what we have when we can add additional sensors to give a much better sense of what's going on around us?
The arrangement of eyes we have is to enable us to hunt things down and kill them, not to react to things behind us. And we do so only under certain lighting conditions.
(Score: 0) by Anonymous Coward on Sunday August 11 2019, @08:09PM (1 child)
Yes, but billions of people can drive a car and walk around. So it must work pretty good...
My point was aliasing. Which is created by these techs is something that needs to be worked around. My brain does not have aliasing. It has a cascading network that ignores things at random and is easy to trick. The hardware itself works. The network behind it is not reliable. Pretty much 99.9999999% of the computer sensors out there use sampling and have a alias to get it 'close enough' that we think it looks good. But the reality is there are gaps and jaggies everywhere. Those are the weak spots in the current systems being made. Working around those will create blind spots and weird anomalies.
Your point is you can enhance them. Which is true. But to simply discard the existing designs is a bit short sighted?
(Score: 2) by MostCynical on Monday August 12 2019, @03:53AM
Humans do drive, but we're actually not very good at it [wikipedia.org]
"I guess once you start doubting, there's no end to it." -Batou, Ghost in the Shell: Stand Alone Complex
(Score: 0) by Anonymous Coward on Sunday August 11 2019, @02:34AM (1 child)
> The problem with LIDAR is its dependence on moving parts,
I figured that everyone in the autonomous driving game was using expensive military/aircraft LIDARs to get things working. By the time they are ready to go into big production solid state LIDAR will be available at automotive prices (crazy cheap). Here's one player from a few months ago, I believe I've seen other press releases as well,
https://spectrum.ieee.org/cars-that-think/transportation/sensors/lumotive-says-its-got-a-solidstate-lidar-that-really-works [ieee.org]
(Score: 1) by Ethanol-fueled on Sunday August 11 2019, @02:59AM
We're using radars, not LIDARS. But hell, I suppose that theirs too is a workable solution. HahahhahahheeeeheeeehOOOOOOO!
(Score: 2) by darkfeline on Sunday August 11 2019, @06:41AM
The first adopters of stairs fell to their deaths. Hell, people still fall down stairs and die quite often. You would think we'd have worked out all the issues with stairs throughout the millennia. As it turns out, living is quite dangerous; the death rate is 100%.
Maybe self-driving cars still have issues. But Waymo has driven first adopters around for eight months and I haven't heard of any of them dying, yet. Maybe one of them will get T-boned by a drunk truck driver before the end of the year, but I'm not really seeing any glaring reasons to FUD self-driving cars.
Join the SDF Public Access UNIX System today!
(Score: 3, Insightful) by Anonymous Coward on Sunday August 11 2019, @02:37PM
You're not an "industry insider". You're about 12 years old. And from your posts, I'd say there's a high probability you are intellectually retarded or otherwise mentally challenged.