Not exactly in-depth reporting here, but I wouldn't expect much more of Vogue. In all honesty, I'm not surprised that restaurants you have to make reservations with want to know everything about you before you walk in - our world's economy is all based on personal information; the "customer" is the most valuable product. Heck, if I'm to be perfectly honest, I would not be surprised if there was a restaurant out there that ran reverse image searches on every guest; with the speed of it now and the huge number of images indexed, you can find out more than ever before.
However, my takeaway here is pretty simple: avoid booking a reservation over a website. Doubly so if they want private data, and triply on the same account from which you booked another table. Then again, I'm a bit of a nut and I like my private information to stay private.
That might be true. Their preference for a personalized experience should not allow the establishment to impact on my privacy. It should be strongly opt-in, not "avoid and never patronize this restaurant to opt-out".
It might not be, although privacy laws in the EU tend to be stricter than those in the US. "Busted" could simply refer to this practice being outed to customers who might not like it.
I think people would be incredibly surprised how common this is. Heck, I've seen random billboards at Chicago bus stops have facial recognition cameras.
I predict this market to basically become more or less pervasive within 5-10 years.
I worked in digital signage advertising (fancy schmancy name for those screens in malls that show ads). Facial recognition was spreading throughout the industry when I was involved a few years ago. The cameras count the number of people in front of the sign, how long they look, and try to guess demographic info (age, gender, race). Everything we hate about online advertising... yeah these people want to make it part of physical life too.
I would be interested to know how to spot cameras on devices such as these (presumably, they aren’t as obvious as a laptop camera). I checked the link, but didn’t go further than the first page.
The worst thing about that is that the billboard that you're talking about is for car insurance. Theres a camera built into the display and the animation tracks where your face is looking.
I talked to a company that was building some kind of hardware to transcribe in real time what conversations people were having at bus stops (the round trip to the server was not fast enough).
They said the buyers used it for targeting ads, after the facial recognition stuff was considered not enough.
This article sounds pretty extreme. "They've read your biography" to personalize your dining experience? I think very, very few places would bother with something like that. Maybe if you are such a regular that you know all the people there by name, or if you go to a very expensive restaurant where cost is not an issue.
This is an important point that the title elides: there's still a good amount of manual work required here which means that it's only really worth the effort for fine dining. As the article mentions, they already did things like this with index cards in the past.
It smells like this approach opens up a sort of attack vector. Eg, to wreck someone's experience by getting some false information injected into a target's social media streams in order to be picked up by these information harvesters.
True, but in Android's case, the implementation seems to be flawed:
In our lab environment we observed that in addition to periodic global MAC addressed probe requests, we were able to force the transmission of additional such probes for all Android devices. First, anytime the user simply turned on the screen, a set of global probe requests were transmitted. An active user, in effect, renders randomization moot, eliminating the privacy countermeasure all together.
The only model researchers found to do randomization correctly was the Cat S60
Amusingly enough, that's probably more because of an oversight by the manufacturer (who is probably just copying a reference design) --- they don't bother writing a "fixed MAC" and just leave the NVRAM blank, meaning the device doesn't actually have a MAC and generates a random one every time you turn WiFi on and off.
That's how it works on the Mediatek platforms that I'm more familiar with, and you'll find this is the case for a lot of the cheap Chinese unbranded phones. It's irritating when you need to connect to networks using MAC filtering, but I'm almost tempted to respond to all the "my MAC keeps changing" threads on XDA with "that's a feature, not a bug." ;-)
Not green is disconnected from any access point. The radio is still active, and listens for SSID broadcast, and advertises a MAC address, for possible connection handshakes from access points that are listening and active, but do not advertise presence and require users to request by name.
Not green AND CROSSED OUT means the radio is actually turned off, and is supposedly neither listening nor broadcasting, if state of the settings applet is somehow more trustworthy than the state of the control center’s slide-out drawer widget.
Note that other non-apple handsets might just be doing whatever, unpredictably.
Euclid and other companies are setup to provide this as a service to retailers, though device fingerprinting became harder when iOS starting randomizing MAC addresses
Right now, it's high-end restaurants in New York City doing this. But they did that by hand in the past. The "headwaiter who knows everybody" was a big deal once.
The big change will come when this filters down to chain restaurants. Which way will Starbucks go on this? They already have an affinity card and know your WiFi information. "You haven't visited a Starbucks in 72 hours. Are you OK?"
Starbucks has been pushing gameified star bonuses and events from my understanding. It effectively does "Hey would you like some stars" on everyones phone. Also they give bonuses and stars for food you don't eat to encourage you to try the other food.
It works quite well (at least on me). Responses usually go from "well, I was already feeling like one of those so might as well get some points" to "why would I want a Frappachino when it's so cold out"? I assume their algorithms are under heavy development and are basically random A/B tests at this point, perhaps primed with some basic data and stars added as an enticement.
In the future, I could see it evolving into something where it sees what the weather is, looks at what websites I'm visiting (via data brokers), analyzes historical trends, extrapolates my likely mood and finally uses all of this to figure out what sugary comfort drink I'm mostly likely to bite at. Perhaps I should take comfort in how utterly predictable I am and how easily open to suggestion...
I wonder if there could be an interesting ML classifier using information like this. Crawl a customers social media, track their behavior in the restaurant. Surely some interesting patterns will emerge.
Be sue to read though to the last paragraph, where the story becomes a bit darker (though not unexpected to HN readers):
They’ve read your biography . . .
“We have had guests come in and we’ve read their biography and read maybe they had a really rough last year or maybe they lost someone in their life recently,” Meyer says. “That may sound extreme, but all of those pieces of information are going to help us do a better job of customizing the hug when they come in. ..."
> "We know more about our guests than we ever have before—you cannot have enough information. Now we have no excuse not to customize the experience [...]"
Maybe some customers do not want a custom experience? There might be advantages, but I guess that some customers pick places because of "the atmosphere, the general food and the waiters there" - so a too targeted experience could also be harmful.
No, what’s really going on is the franchise vs. mom & pop version of corporate espionage.
Normal people with no skin in the game, beyond what’s for lunch, don’t behave this way.
There’s a small-ish subculture of afficionados and foodies, but I’ve never met anyone that obsesses over dinner to that degree. Usually it’s a byproduct of a phase they breifly went through, when they were on a kick about something.
The rare exception being the occassional young romantic couple, obsessing over the courtship rituals of wooing their hot date. This is less about the food than the date, and again, it’s yet another small-ish subculture, since 90% of everyone is undateable. [0,1]
Nice times when customisation and personalisation were things that the user did. Reading other comments here, I guess we need some ad block software not only on our browsers, but also on our faces.
Disclosure: I am the Director of Engineering at Tock. We power the reservation system at Eleven Madison Park which Will and Daniel own.
The notion of a user's past history with rich data has been something high end restaurants have been trying to manage either via spreadsheets or via custom crm implementations. What we have done is to make that easier through product design and machine intelligence to bring it to more down market restaurants.
It unlocks really great experiences for guests, avoids the need to talk about your allergies every time you walk into a restaurant and makes sure it is accurate and upto date.
c) Improved revenue comes with increased visit and spend, that happens only if the customers get a great experience.
The practice of taking notes is something that every restaurant is doing. What we make sure is that they capture only the necessary information. Every customer controls their public information through their profile page and can choose to remove it if needed.
Lots of businesses are looking for ways to do lots of unethical or undesirable things - that doesn’t somehow absolve you of responsibility for enabling them.
Typing up notes is far from unethical or undesirable. Our responsibility is to make sure the product focuses on the right things like dietary restrictions or visit history, to reward for loyalty and provide a great experience
It's going to be like eating at L'Idiot in Steve Martin's LA Story. There was a screen showing tidbits of live data about celebrities as they entered - how much their last film grossed, their sexual orientation, etc.
"You think with a financial statement like this you can have the duck?!"
59 comments
[ 3.1 ms ] story [ 108 ms ] threadHowever, my takeaway here is pretty simple: avoid booking a reservation over a website. Doubly so if they want private data, and triply on the same account from which you booked another table. Then again, I'm a bit of a nut and I like my private information to stay private.
> A crashed advertisement reveals the code of the facial recognition system used by a pizza shop in Oslo...
Source: https://twitter.com/gamblelee/status/862307447276544000
> Slang Placed under arrest. http://www.thefreedictionary.com/busted
maybe that deosn't apply in this context.
I predict this market to basically become more or less pervasive within 5-10 years.
Example of what I'm talking about: https://signbox.tv/digital-signage-products/digital-signage-...
I imagine as awareness increases you'll see things like remote cameras zoomed in/etc.
To be fair the one I saw I only knew had a camera - I assumed it was for facial recognition, but I can't think of any other plausible reason for it.
They said the buyers used it for targeting ads, after the facial recognition stuff was considered not enough.
In our lab environment we observed that in addition to periodic global MAC addressed probe requests, we were able to force the transmission of additional such probes for all Android devices. First, anytime the user simply turned on the screen, a set of global probe requests were transmitted. An active user, in effect, renders randomization moot, eliminating the privacy countermeasure all together.
https://arstechnica.com/information-technology/2017/03/shiel...
Amusingly enough, that's probably more because of an oversight by the manufacturer (who is probably just copying a reference design) --- they don't bother writing a "fixed MAC" and just leave the NVRAM blank, meaning the device doesn't actually have a MAC and generates a random one every time you turn WiFi on and off.
That's how it works on the Mediatek platforms that I'm more familiar with, and you'll find this is the case for a lot of the cheap Chinese unbranded phones. It's irritating when you need to connect to networks using MAC filtering, but I'm almost tempted to respond to all the "my MAC keeps changing" threads on XDA with "that's a feature, not a bug." ;-)
By "off", do you mean when the iPhone WiFi setting is set to off (not green)? How does that work?
Not green AND CROSSED OUT means the radio is actually turned off, and is supposedly neither listening nor broadcasting, if state of the settings applet is somehow more trustworthy than the state of the control center’s slide-out drawer widget.
Note that other non-apple handsets might just be doing whatever, unpredictably.
The big change will come when this filters down to chain restaurants. Which way will Starbucks go on this? They already have an affinity card and know your WiFi information. "You haven't visited a Starbucks in 72 hours. Are you OK?"
In the future, I could see it evolving into something where it sees what the weather is, looks at what websites I'm visiting (via data brokers), analyzes historical trends, extrapolates my likely mood and finally uses all of this to figure out what sugary comfort drink I'm mostly likely to bite at. Perhaps I should take comfort in how utterly predictable I am and how easily open to suggestion...
On a similar note, if you haven't seen it; https://www.ted.com/talks/zeynep_tufekci_we_re_building_a_dy...
They’ve read your biography . . .
“We have had guests come in and we’ve read their biography and read maybe they had a really rough last year or maybe they lost someone in their life recently,” Meyer says. “That may sound extreme, but all of those pieces of information are going to help us do a better job of customizing the hug when they come in. ..."
Maybe some customers do not want a custom experience? There might be advantages, but I guess that some customers pick places because of "the atmosphere, the general food and the waiters there" - so a too targeted experience could also be harmful.
I must be massively out of touch with modern culture, because I've never once done that.
Normal people with no skin in the game, beyond what’s for lunch, don’t behave this way.
There’s a small-ish subculture of afficionados and foodies, but I’ve never met anyone that obsesses over dinner to that degree. Usually it’s a byproduct of a phase they breifly went through, when they were on a kick about something.
The rare exception being the occassional young romantic couple, obsessing over the courtship rituals of wooing their hot date. This is less about the food than the date, and again, it’s yet another small-ish subculture, since 90% of everyone is undateable. [0,1]
[0] http://www.pkmeco.com/seinfeld/wink.htm
[1] https://youtu.be/C-a64OwOYqU
Non-Seinfield non-Youtube citation needed.
---
I have become one of those "Citation needed" people. Argh.
https://en.wikipedia.org/wiki/Drifting_(motorsport)
(hey, hacker news, fuck your downvotes)(and by the way, I finally figured out what that warning light means, by googling TPMS! Thanks, Batman!)
The notion of a user's past history with rich data has been something high end restaurants have been trying to manage either via spreadsheets or via custom crm implementations. What we have done is to make that easier through product design and machine intelligence to bring it to more down market restaurants.
It unlocks really great experiences for guests, avoids the need to talk about your allergies every time you walk into a restaurant and makes sure it is accurate and upto date.
a) you spy on your guests
b) you cross reference with other data sets without your guests knowing
c) you do this for improved revenue
I'll make sure I never eat at restaurants using Tock. Good look seting up proper opt-in once you serve EU guests after May 2018.
b)Yes, enriching the data is a key part of our offering but we are very transparent in our data practices and take this job very seriously.
https://www.exploretock.com/terms https://www.exploretock.com/privacy
c) Improved revenue comes with increased visit and spend, that happens only if the customers get a great experience.
The practice of taking notes is something that every restaurant is doing. What we make sure is that they capture only the necessary information. Every customer controls their public information through their profile page and can choose to remove it if needed.
"You think with a financial statement like this you can have the duck?!"