> Dooring people aside, what do you do if someone just leaves the door open when they leave their ride?!
Continue billing them for the ride and send an app notification or phone-call to their phone.
Other potential solutions: If the door is still not closed after n minutes, plead with passers-by, or offer a passing or nearby rider the chance to earn credit by closing the door.
Totally fair to be concerned about pervasive surveillance for the _potential_ of privacy violation. Not sure what to do about that.
That being said, just speaking with some knowledge of current state: the scans don't live forever. At this point, all the data they collect is way too big to store even for a short period. They'll only keep data in scenarios that are helpful for improving driving performance, which is a tiny subset.
Personally identifiable information is also redacted.
You should probably be more worried about what gmail knows about you than Waymo.
True; I should have said metadata and not just data since you're right that the volume of raw images would be too big to store indefinitely. It's way more feasible to process the raw images and store the inferences, like number of persons visible in last 5 seconds, or dates and times a person who looks like me has been seen by a Waymo while my particular Android phone is nearby, or dates and times they have seen [my OCRed car number plate].
> Video Clips captured by the LPR system will automatically be deleted after 30 days; although Images are deleted when no longer needed, the data obtained from the Images may be retained indefinitely. Should any information from the LPR Dashboard be needed to assist with a security or law enforcement matter, it may be retained indefinitely, in paper and electronic form, as part of the security file until it is determined it is no longer needed; in addition, it may be shared with local law enforcement who may retain it in accordance with their own retention policy.
If anyone can share a link to a similar IRL privacy policy for Waymo I would love to read it. The one on their website is conspicuously labeled Waymo Web Privacy Policy lol
That's still not really what I'm looking for. I am curious about a “what we keep, for how long” policy for the sensors on the outside of the cars like the Flock one I linked above.
Your second link does mention cameras and microphones outside the car but doesn't mention what they keep (full video? stills? LIDAR? RADAR?) or for how long:
“Waymo’s cameras also see the world in context, as a human would, but with a 360° field of view. Our high-resolution vision system can help us detect important things in the world around us like traffic lights and construction zones. Our systems are not designed to use this data to identify individual people.”
The “Our view on your privacy” section links to the same page as your first link, and that page's “What we keep” section is explicitly only about riders:
“We will retain information we associate with your Waymo account, such as name, email and trip history, while your account remains active.”
Right. I merely shared the closest thing that I could find, and as mentioned, the first link is specific to the ride-hailing service. Notably, in the second link, there is a reference to sharing information with law enforcement, but it's generally lacking in details.
It's Google we're talking about. In no way do I trust them to take pictures of me using their city-wide camera network and not use face/body recognition to keep track of where I go, for the purposes of targeting advertisement.
People are able to get their very boring suburban house that you can find pictures of the interior on zillow of blurred for years (indefinitely?) on google street view. If they were so cartoonishly evil they would not let you do that.
I agree in principle that a privately run company could use information in nefarious ways internally, and that barring any additional knowledge you should not trust them.
That being said, I have an anecdote as a former googler: the reality with Google though is very thoughtful and favorable for users if you ask Googlers who've worked on their software products. There are audit trails that can result in instant termination if it's determined that you accessed data without proper business justification. I've known an engineer who was fired for an insufficiently justified user lookup (and later re-hired when they did a deeper look -- hilariously they made this person go through orientation again).
And safeguards / approvals required to access data, so it's not just any joe shmoe who can access the data. Wanna use some data from another Google product for your Google product? You're SOL in most cases. Even accessing training data sourced from youtube videos was so difficult that people grumbled "if I were outside of Google at OpenAI or something I'd have an easier time getting hold of youtube videos -- I'd just scrape them."
This isn't to say any of this is a fair thing to make decisions on for most people, because companies change and welp how do you actually know they're doing the right thing? Imo stronger industry-wide regulations would actually help Google because they already built so much infra to support this stuff, and forcing everyone else to spend energy getting on their level would be a competitive advantage.
The impression I have, as an outsider, is that Google hoovers up all information available to them and uses it as input to various algorithms and ML models for targeted advertising. I'm sure individual Google employees are as thoughtful as you say, but I don't think the organization views itself as it users' "enemy" or as something which its uses should be protected from.
I'm not afraid of employees at Google or random Google divisions obtaining unauthorised access to information at me, it's not about that. I'm certain that there's very little data that the targeted advertising part of Google can't access.
Honest question - what's the harm in being targeted by ads? Is it just scrolling youtube more often than you should? Or is there a nefarious side that I'm failing to consider?
For me the thing I hate about location tracking and the ilk is primarily about its harmful externalities (e.g. put into use by gov't, abusive users, or by Google for anticompetitive reasons), not targeted advertising itself.
I find it disgusting that our society invests so much effort into manipulating everyone, companies spend billions on armies of psychologists, computer science experts and data centres whose only job is to manipulate people into buying things they don't need. Targeted advertising is even more disgusting than non-targeted advertising, because there you're trying to find an individual person's weaknesses for more effective manipulation. It's simply evil.
That, and the existence of targeted advertising incentivises collecting and correlating as much data about people all the time. If it wasn't for targeted ads, I'm sure that Google would've actually just used data from their city-wide surveillance networks for improving their cars (at least until a government would've asked for the data, which is also an issue). But with targeted ads in the mix, there's a huge incentive to collect it and correlate it with all the other data Google has, which is creepy.
I guess. But bringing new products to market requires distribution, and do you have a better way for people to crack that? Targeted advertising through say, Instagram, has enabled a lot of small businesses whom would otherwise struggle to aggregate demand.
So it's not like pure evil. In many cases there's a service being provided to match users to products they want / that don't suck.
> with targeted ads in the mix, there's a huge incentive to collect it and correlate it with all the other data Google has, which is creepy.
Strongly agree that in theory this shit can be used nefariously. That said, Google is far from the scariest of the bunch despite being the biggest. Telecom for example wants to deep inspect your network packets, and they can tell where you are physically today, anywhere in the country without even having cameras driving around 5 US cities.
Stronger regulations around data rights and privacy have been proven to work by the EU. I don't really see another solution apart from a legislative one.
> They gotta supplement mass transit for dense cities, not replace it.
Full agreement here. AVs are great for last-mile transit.
> horrible use of space until they become autonomous buses on dedicated bus lanes, or trains
This is where we disagree. The whole point of AV TaaS is that they can go where bus lanes and trains can't. Last mile transportation.
I also wouldn't say they do "jack shit" for traffic in the sense that they reduce the need for parking, and reduce accidents which are the source of a lot of unpredictable congestion.
Surely there are tradeoffs. They indirectly incentivize sprawl and taking more taxi rides overall. And I get the tire residue argument (especially since AV fleets are mostly electric with high torque generating more tire wear). But is tire noise really a fair complaint? They're just going where cars already go and tires are engineered pretty well to minimize noise...
Tire noise is still quite substantial - use the NIOSH noise meter app on your phone to compare the sound levels on a city block when an EV goes by compared to just bikes and people – and there’s a growing body of evidence that noise levels correlate with worse health and sleep for residents. EVs help, but it’s only partial.
I guess my question is - will Waymo really cause that much more tire noise vs today? Wherever car tires are making noise, I doubt Waymo is going to generate that much _additional_ noise.
They have nice clients (e.g. for MacOS, Tizen). Ofc headscale is a thing, but if you have a company, it's also nice to have someone to yell at if your mission-critical tailnet suddenly b0rks.
Imo they don't charge all that much relative to their value, depending on who you're asking.
Years ago at Google, the permissions system (OG ganpati, for the googlers) was having major performance issues related to OOM in Java. IIRC there was an epic thread where folks discussed potential resolutions including, this being Java, GC tuning. But in the end, string interning was what helped a ton to defer the scaling issues until something else could be done (mostly a rewrite, smh). I mostly remember this as being a shocking solution to me. Part of me imagined there'd be some fancy scalable solution instead of single-binary optimization, but at the end of the day sometimes you just gotta intern some strings.
Having a non-user replaceable battery is a really easy way to ensure a product stops working after 3-4 years though.
And the criticism is typically directed at companies like Apple, who does make things that last physically, but then force you to upgrade by way of battery.
Both can be true: planned obsolescence is real, but building things to last is difficult too.
IMO the durability problems in early generations of products tend to be "real", because there are still real engineering problems that aren't understood, and there isn't (generally) a super limited market. Once the engineering problems are solved and the market is fully saturated, there is suddenly an incentive to add planned obsolescence. I don't have any data to back up this claim though.
If you have a product that's been in the market for a while and it looks like it's meeting service life expectations you start looking at it trying to find ways to save money by substituting cheaper parts. You swap out metal gears for plastic gears, for instance.
If these parts have a shorter service life, but the service life is still longer than the warranty, then maybe that's a win in two ways for the manufacturer.
> You swap out metal gears for plastic gears, for instance.
Great, till the motor that drives the gears jams. When the gears are metal,the expensive part (the motor) is more likely to lose. When the gears are plastic, the motor survives and you need to replace the gears with nylon ones or 3D print your own.
The plastic gears may not always be designed as a sacrificial part, but most consumers unfairly dismiss the possibility immediately
This comes down to warranty too. If it fails during the warranty period, which one does the OEM want to pay to replace: the expensive motor, or the cheap gearing?
I think of it as a continuous feedback loop between engineering, finance, and QA that ultimately ends in a product being manufactured as inexpensively as possible without dying in the warranty period.
Wow that's super interesting! I've never heard of this, but the appeal is immediately obvious. Thanks for commenting, gonna have to do a Wikipedia binge.
You're never really forced to upgrade because of battery. If you don't want to get Apple to replace it (which though expensive is still much cheaper than a new phone), then you can take it to your local phone repair shop which will do it for not much more than the cost of a replacement battery.
> the screen is maliciously connected to the board with strong adhesive
That’s not necessarily malice. Using lots of glue makes the device stronger, and making glue that a) glues really well (if there’s as good as no bezel, how is the screen staying attached to the phone otherwise?), b) lasts for years in any climate and c) can be easily removed isn’t an easy problem.
So, how do you screw a thin piece of glass onto a phone that doesn’t have bevels to speak of in such a way that you can put it into your pocket for years, and push a finger on the center of the screen tens of thousands of times without breaking?
Also, if there’s room below the screen, the screen will bend more than when there isn’t, and that will affect longevity.
I’m not claiming using glue wasn’t done out of malice, just that we can’t say it is.
In any case, all it takes to repair a phone with a glued screen is a two face suction grip for about 20 dollars and an ordinary hair dryer.
The nasty part of a phone repair, I will admit that, is scraping off the glue gunk - I had to repair a Google Pixel once where the battery was dead, and during removing the glue on the display unit border I apparently managed to damage the seal between the OLED display and the glass, exposing the OLED to oxygen which led to eventual oxidization and a new display panel.
It not that easy with their glued in batteries on some Macbook Pros. You have to essentially use alcohol to remove the glue to replace the battery. Absolute PITA. They could have used 4 screws and it would be easy to replace.
Apple has a high profit margin on their products so I expect better. This isn’t a cheap laptop from a supermarket.
Agreed but on the other side it makes the manufacturing more complex - another plastic part and screws as well as the time needed compared to just gluing in the battery.
I suspect this is a classic example of corporate beancounting at work, even if it just a dollar or two per machine, at Apple's volume of millions of machines that's nothing to sneeze at.
To fix it, we need laws that require a certain repairability score for all devices sold. Then doing the "right thing" would be a KPI that competes with pure financial incentives.
> Agreed but on the other side it makes the manufacturing more complex - another plastic part and screws as well as the time needed compared to just gluing in the battery.
>
> I suspect this is a classic example of corporate beancounting at work, even if it just a dollar or two per machine, at Apple's volume of millions of machines that's nothing to sneeze at.
They make a high margin on each device and other manufacturers can manage it fine at similar price points. I believe it was deliberate, they back tracked after being highly criticised for it.
> To fix it, we need laws that require a certain repairability score for all devices sold. Then doing the "right thing" would be a KPI that competes with pure financial incentives.
If people are concerned about repairability they should seek out manufacturers that offer products where they have a good track record.
Laptops, tablets and phones are seen as partly consumable by the majority of people and they replace them every few years. I am not saying that it is right, I am just saying that is the reality. Also not every problem can be legislated away and if you make something a KPI it will be gamed.
It’s fairly easy to open. They designed it so a cheap and inexperienced worker in the Apple Store can replace the battery quickly and without issues.
They also made a massive improvement by designing an adhesive for the battery that detaches with electricity. So you no longer have to use pull tabs or heat.
Are Mac laptops still glued together? My 2013 MBP needed a new battery, which required replacing the following as one unit: battery, keyboard, top case, trackpad. The reason is that it was all one blob. (And then the charging circuit on the motherboard died, and I moved on to ThinkPads I can upgrade and deal with myself.)
My last iPhone’s battery lasted about 5.5 years before it needed to be replaced. Replacing it cost about $90 + tax at the Apple store. The bottom line is Apple products do last and if you need a new battery, you can get one.
> Support Ukraine enough so it can take all its territory (maybe minus Crimea). This may not be possible with weapons alone. This might require a NATO no-fly zone over Ukraine, which effectively makes us a combatant. I actually would support this path, but the downsides are all too obvious.
It _is_ possible with the combo of weapons and economic sanctions. Ukraine (had) more economic support, so it is both a war of blood and endurance, endurance which Ukraine could have by being a porcupine with next-gen weapons and non-stop bleeding for Russia until things start hitting the oligarch's pocketbooks too far and the people of Russia begin to demand regime change.
Russia is a glorified gas station. You can hurt it dramatically with economic sanctions because its whole GDP is concentrated into one sector.
Imo if Trump really cared about America's long term security, he would be trying to fund research into cheaper / easier to produce next-gen weapons for Ukraine to test for us, instead of complaining how much our HIMARS and tanks currently cost, firing top military brass, cozying up to a dictator, and trying to exploit a ravaged democratic ally.
Maybe. But Biden and Germany didn't push too hard on either because they feared Russian reaction. Now, with so many Ukrainian soldiers dead, it will be much harder.
Economic sanctions won't affect Russia as long as China is behind them. And if the West is against Russia, they will do anything to keep China as an ally.
Sanctions work! But their enforcement requires attention. Russians are draining their fund to the last drop and there is not much left, they cannot easily get a credit either. Unfortunately, these last weeks give some breathing room to Russians and their evasion schemes as people who look after sanctions are being fired.
> Biden and Germany didn't push too hard on either because they feared Russian reaction
And this was a mistake, one that they slowly tried to reverse once they realized Ukraine wasn't going to be a pushover especially with modern weaponry.
Really? Nobody uses TF but plenty of people use XLA, which is what you allude to saying that TPU in Pytorch is possible. That's arguably the most important piece; the sledgehammer that makes good performance and good devex compatible.
> I wish Google would compete with Nvidia directly and let me buy a TPU
I don't see the utility of buying _any_ GPUs. It's mainly a cost + availability optimization to own them yourself if you're doing a lot of continual training. Outside of the foundation model companies, most of us just use cloud services -- and I want those to be cheap and always use the latest thing.
Google has a highly optimized infrastructure to make sure unused resources (CPU, TPU, RAM, disk) are properly allocated. This means preemptible instances, highly co-located compute + data, etc. In theory everyone should win when you use ML hw through hyperscalers, because they collect on their structural advantages and your TCO is lower.
> Google could have Nvidia’s business and OpenAI’s
They arguably have a better _business_ on their hands but a worse _speculative outlook_ in the eyes of Mr. Market. Those are different things. There's an argument to be made that Google's valuation could be as high if they ran their public relations strategy as well as OpenAI / Nvidia / Tesla, which are all riding monumental hype.
Cloud prices are already driven down by the fact that you can deploy with Google's cloud offerings.
Also, if you weren't going to use Google's offerings you probably wouldn't be down to buy their HW either :)
FWIW your PTSD with Tensorflow is shared by everyone at Google, it was just unavoidable because Google needed ML to be performant and they started by sacrificing devex for performance whereas Pytorch went ergo-first, performance later. In retrospect a prescient move by Meta, but now making the high performance stuff ergonomic is proving out for Google via XLA and JAX.
XLA is the compatibility layer. You train using Torch/Jax/whatever you want, and XLA ensures you can use TPU. There aren't many hurdles, most people train in the cloud anyway.
Dooring people aside, what do you do if someone just leaves the door open when they leave their ride?!