"All pictures you transfer to Memoto’s cloud service are stored encrypted. The pictures are only visible to you: only you can see them and only you can change them."
Is this real client side encryption, or is it Dropbox/Hushmail style encryption where they can do a little server side trickery in order to obtain your encryption key if the US government compels them to?
I'm just thinking how useful it would be to law enforcement if they could just specify a location, date and time and go through pictures taken there for evidence.
I hate this whole idea. It's inevitable that it will happen, but I hate it nonetheless. You wont be able to go outdoors without being constantly recorded by strangers. My only hope is that the popular services manage the technology such that we don't end up with databases that governments, or organisations can read private data from.
Our goal is that all your data will be encrypted and completely secure from the point that they leave your camera until you view them in the app/website/etc. The images would only be your to decrypt and we would have no access to your key, only you will be able to view your images, unless you share them ofc.
Our biggest concerns at the moment is partly how we can implement effective encryption/decryption that won't drain the battery life of your phone or decrease the user experience.
Another concern is what will happen if the user loses their password? If we have no possibility to access a users encrypted data then their data would be lost in this case. We would also need to decrypt and reencrypt all data if the user changes password, something that would be very time/battery consuming if done on the device, instead of our servers.
These concerns are something that we are evaluating solutions to as we speak, our goal is to make it impossible for anyone but you to access your private, aka. not shared, data.
If you have any suggestions we are of course happy to hear them.
The most important thing you need to do is to make your implementation completely transparent and publicly available. Data doesn't get more personal and private than the data you will be storing.
Simply saying that you encrypt data is not enough. People need to know who has the keys and who is able to get access to the keys if they need to.
Also, if the images can be viewed in a webpage, then they're not encrypted as far as I'm concerned. Even if you use client side javascript to handle the decryption, you can be compelled to modify that javascript for a particular user to obtain their key/password next time they log in.
"We would also need to decrypt and reencrypt all data if the user changes password"
This depends if you are using a key derived from the password. Or if you are using a randomly generated key that is encrypted using a key derived from the password. With the latter, you would only need to re-encrypt the randomly generated key. Not the entire data set.
I understand that usability is important for your business case. Just don't forget how sensitive the data is that you will be collecting and how ripe for abuse it is if you don't secure it fully.
I also think that you should open source your client software. If not, I'd be concerned that at some point you'll be forced to put back doors in it to get at peoples keys.
I think the best to hope for is the opposite: that all the databases are publicly available, so at least we have equal access, and high-ranking politicians are just as exposed as everyone else.
The world is changing. When it's this easy and cheap to record everything that happens in public and put it in a searchable database, people will do it. You can't stop the tide.
I still think the best to hope for is that only the individuals with the actual devices are able to look at the data. Can you imagine future potential employees being able to type your name into a web page, getting an image of you, and then through the magic of facial recognition, seeing a long list of geo tagged photos of you going about your normal daily life? Who you associate with, where you go, etc.
We shouldn't support this project or projects like this unless we get full assurances that the privacy issues have been tackled appropriately. At the moment, we don't know how the encryption is done, so as far as we should be concerned, there is none. No encryption.
As far as I'm concerned, with the knowledge we currently have of the project, if this project becomes big, this is worse than for example: The government installing CCTV on every street corner, which automatically uploads data to a central database that they can peruse at will. These cameras will be in and on private property as well.
>Can you imagine future potential employees being able to type your name into a web page, getting an image of you, and then through the magic of facial recognition, seeing a long list of geo tagged photos of you going about your normal daily life? Who you associate with, where you go, etc.
Easily, but what do you propose to do about it?
Creating such a database will become easy and cheap; the technology is inevitable, and there will always be a large userbase who don't care much about privacy concerns; in the "worst" case a large company could hire a bunch of people to walk around taking pictures of everywhere.
You could make taking the pictures illegal, but I don't like the implications of that one bit. Making aggregating them illegal would be very difficult to police and wouldn't stop governments. It probably wouldn't stop large corporations either (they'd do the processing offshore or find some other way around it).
The way I see it there's no way to stop the big corporations getting this ability, so the best thing we can do is level the playing field.
You don't have to be on your camera, if you're on half a dozen strangers cameras who are all publishing their video, and I have an image mosaic algorithm and face recognition.
Yes. I was describing the best case scenario. Which is that only the people who take images have access to them. They will still publish a small number of them of course, like people already do with normal cameras.
This is opposed to the worse case scenario where all images are automatically uploaded to a central database. If this is going to happen, that central database really needs to be encrypted such that only the user can access the images, not even the company that controls the database.
Yeah, but the point I was making, and that you missed, was that some fraction of people will voluntarily publish. They will set their photo album, or video feed, publicly readable. Because they want to.
If that's a large percentage of people, it's going to be hard to avoid being on somebody else's camera, even if you don't have your own.
And thus turning yet another thing into a waste of natural resources. It would be great if people figured out how to avoid zero-sum games as much as possible.
The life recorder is inevitable. I mean full video and audio. I've been talking about it for decades. The social and psychological shift is going to be huge. Are you self conscious when somebody is following you around taking pictures ? This thing will bug everybody out. The first adopters will be security and police (who already do it with their car based video). Then narcissists and people who want to have their own vanity tv show 247. And annoying tech dudes who wear it at parties.
many people will realize how boring their lives look from the outside and they will stop wearing it because it's depressing.
Anyway, the tech isn't the hard problem, it's the analysis, information overload and personal search engine industry. Many companies including my own are skating towards that puck.
As far as I know, no academic research or even rumored corporate research is going into useful, consumer-oriented (meaning able to provide meaningful value to the user with minimal curation and/or improving their quality of life), management of life-recording amounts of video. Most aren't even touching audio.
Audio recording is usually legally restricted under wiretapping laws. Video recording is even dicier, not just with laws varying wildly from state to state, but also with individual privacy and image rights laws coming into play. It's why all the researchers are sticking to photography, unless you're the military, when the rules don't apply to you.
The only researcher who's done any work, as far as I know, into handling that much video and audio is Deb Roy, with his timeworms visualizations, and that's only because it was his own household that he was able to clear an IRB. And it's still not useful by my previous definition.
I'm not yet interested in audio/video, but that's eventually where the puck will be (so that's where we should be skating to - if you're familiar with that metaphor).
But in fact I think what's externally visible about a person's day to day life isn't that interesting. Especially if he's staring at a computer all day. A lot of quantified self stuff I actually find quite dull and useless. I'd be more interested in seeing my blood sugar levels
Humans have an intense world of internal description and that's what I'm interested in working with. Describing your life is a way to actually frame your own feelings and sense of purpose.
My own life has been dramatically improved by the software I've written that really just tells me what the f•¶§∞ I've been doing and let's me visually compare that to what I want to be doing.
As I get rolling here turning into an actual product then I'll be more public about it. Definitely interested in talking with people though. I'm especially looking for datavisualization people.
"The analysis unit analyzes the log information collected from the information collection unit and decides at least one topic representing the user's daily life information."
Gordon Bell's testbed was outstripped by the possible data collection years ago, and he's admitted he rarely did anything with his archives after collecting what he did, serving more as a reference library for others.
That Samsung patent sure is interesting, though. The US filing of March 31, 2012 is interesting, too, as I started work on something similar in February (I see their foreign patents are from last year).
Cool, the slide deck was interesting. I saved your links.
The Samsung thing sounds like it could just choose a likely significant event and make it into a story. it could be extremely lame and unworthy of a patent. it could be entirely as combat move against apple.
Wouldn't it be easier to modify something like the already incredibly cheap (~$10) available key chain cameras? There are already quite sophisticated discussions on these out there http://www.chucklohr.com/808/index.shtml.
But this has a social app! This is a kickstarter project! This gives you the amazing opportunity to upload automatically every single moment of your life to a server on which you have absolutely no control! But they will be marked as private so totally safe! /irony
From a hardware standpoint, completely possible. The problem is trying to write that code for a non-standard architecture. These cheap Chinese processors often completely lack even minimal documentation. See this great post:
Depressing. This is like that movie with Robin Williams "The Final Cut". http://www.imdb.com/title/tt0364343/ A total privacy nightmare. Your government is gonna love it though.
Yeah, this is a really cool idea, but only if you have full control of your data. Ironic that they should mention "full control" as an argument in favor of storing the data on someone else's server.
You're only a subpoena away from having every intimate detail of your life revealed to the authorities, unless you manage to turn the camera off whenever something potentially suspicious happens (this is also called circumstantial evidence).
There's no way to have full control. You might get to have full control over photos your Memoto snapped, but I sure as hell can think of no way to get full control over photos of you snapped by someone else's Memoto.
The good thing about being with someone is that : being with someone ALONE. Not someone analyzing what I did with my wife one month ago, where, how, why.
The idea is not new, Justin started http://www.justin.tv/ recording everything in his life with a video camera.
I don't want anything I say or do being recorder by someone else(I know this camera only takes pictures, by now), and I don't want other people(smartphone manufacturer and the US government) to know what I did Thursday at 3:00 pm.
One of the worse things about living in a small village is all people controlling what others do, gossip here and there. In cities people were free of other people trying to control them, with cameras everywhere and web social services it is becoming rural village again.
It is big brother´s dream. In the future the government will use cameras on the street, face recognition and servers to track anybody at any time.
I remember a post from another thread saying that our experience of being able to go out anonymously in a crowd is basically a blip in human history. When you live in a village it's impossible to go anywhere (you still have privacy when you're actually in private) without all your neighbours knowing, and it'll be the same when this technology becomes widespread.
Many of our current daily experiences are a blip in human history. I'm not concerned about how old the experience of relative anonymity in public spaces is. It's a good thing, that we should strive to keep.
Ah, thanks. I misread. I thought the OP as complaining about wanting to be alone with someone, rather than analyzing what they did earlier with the person.
Looks cool... but... Does this have a long-term future? Soon (I hope) we will all be wearing glasses with cameras in and I wonder whether that will render this obsolete?
It has a "privacy button" that stops it from taking photos for a few minutes and records other information along with the images (like temperature, for example). Also, all of the images and data are stored locally on the device.
Another lifelogging camera, with better specs, about to hit the market (Nov 2012 according to the register/buy page) is Autographer, from the Oxford Metrics Group:
It costs 249 bucks for the later batches and 199 dollars for the early version. Expensive. I guess I'll have to wait for product evolution to make stuff cheaper.
On the other hand, the application for lifelogging is pretty long. An employed programmer(or indeed, anybody who makes money) should be able to justify the expense of 249 bucks no problem.
2000 pictures per day doesn't sound like much (a bit more than 2 pictures on average per minute) but this really adds up even at the resolution of this camera, they claim about 1.5 TB / camera / year. Uploading and storing that much data for a large number of users in a reliable and cost effective manner is a non-trivial exercise.
So the big question is what their monthly fee is and I think this project should disclose those (projected) fees and not just the price of the device because it could very well be that those monthly fees will be the large component.
I guess I'd rather have less photo quality and less space requirements. For this specific use case, I guess 3 Megapixel would be sufficient, and then save it not as jpeg but in jpeg2000 or another wavelet compression, at best with a solid compression rate.
The images could naturally degrade over time too. Ie, older photos are compressed more. Although the process of fetching an image, compressing it, and replacing it, might be more trouble than the storage saving is worth. Could keep every nth photo as full resolution maybe.
You probably just re-invented what the brain may be doing as our "memories fade": re-compressing (and culling) to free up space, with storage surprisingly expensive. :)
I'd rather see someone create really, really disruptive technology like a "personal privacy bubble" I can activate when I choose to. Something that will auto-blurr and auto-jam all video/audio. THAT is something people would like. "Want to film me? I opt out!"
Lol, people still buy adult books in stores? How about all those CCTV camera's? The camera's in stores can even tell if you're male/female and your general emotion (according to the lines in your face).
Question is: Do I even have the right not to be filmed? Somehow I think not... Apple did file a patent to disable iPhone video recording using lasers (to be used at concerts).
As cool as this is, I'll be interested once we have always on audio, video and GPS, with-- while we're day dreaming-- a post-process that analyses the data and reduces it to a day summary complete with transcribed conversations, things you saw and people you met.
You can assume the Memoto team is having an ongoing conversation with the (Swedish) authorities discussing privacy issues for a device that only saves images.
Adding sound and video moves the privacy concerns to a whole new level. Are we, as a society, even ready for such a device?
Also, you add a lot of complexity both in terms of usage and in terms of manufacturing and support.
Hi all, sorry to come late to this thread. I'm one of the founders of Memoto and a long term HNer. FWIW, Memoto would never have reached this stage without HN.
Of course we are very happy about the results of the Kickstarter campaign. Personally I am also very happy about the healthy concerns raised here on HN, to continue to inspire us to never settle for mediocrity. We have a great team of engineers working on both hardware and software. It is a challenge to handle such a long tech chain but we feel it is manageable with the team we have.
The geotagging of events, plus all the wonderful image recognition and categorizing (plus automatically judging photos as interesting or valuable) are the selling point of this camera. It'd be pretty hard to get shots of yourself though, maybe they could introduce some sort of opt in feature where other users of this kind of camera that were in a close by position at similar times would contribute to your feed? (Or did I just invent Color?)
I don't get it. I'm not going to wear this to work, partly because my boss would think I was a spy and partly because my work life simply isn't worth recording at that level of detail.
So then I take it with me places on the weekends. But, I already take a high quality camera with me most-places (courtesy of my smart-phone) and it takes pictures at EXACTLY the right moment, it doesn't miss the shot because of a poor-angle, bad light, bad composition or bad timing. At least when it does, it's my fault.
To me this seems way too niche to ever succeed. I expect I'll be wearing one though in a few years time :-)
There's some interesting psychological research that suggests "random" photos taken at fixed intervals are better at triggering memories than posed photographs.
“To recharge the camera’s batteries, you connect the camera to your computer; at the same time the photos are automatically uploaded to Memoto’s servers”
What? I can’t just keep the photos without putting them on Memoto’s servers ??
The cloud model would seem problematic for this
much data, given the bandwidth constraints of many users.
For my DSL connection (3Mbps/0.8Mbps) I would need to
saturate my outgoing bandwidth half of each day
to upload 4GB. And I think that these numbers aren't
far off from the global average.
If this is going to be a viable product, I don't
understand at why the developers wouldn't provide
a local storage model. Without that, I for one
would have zero interest in the service as described,
particularly given all the other concerns of handing
over this much intensely personal data.
2000 geotagged pictures a day, every day of my life, automatically uploaded to their servers, and no word on how would one set up their own private server? Let me think about it... #not
I'm not sure why they wouldn't let you have the photos yourself instead of forcing you to upload to their server. They want to sell the hardware, not lock you in (I think?).
I wish i was more qualified to identify the myriads of cliches in their video. Hipstery shaky 70s-ish looking nostalgic video, rounded edges, yellow plastic, heavy SLR.
But really, this seems like a cool idea for a phone app, why the extra hardware?
I think this is a really cool idea, but it's still about 2 or 3 times too big for me to wear it. It's big enough that people will notice it and ask about it, which is enough to discourage me from wanting one.
Is this real client side encryption, or is it Dropbox/Hushmail style encryption where they can do a little server side trickery in order to obtain your encryption key if the US government compels them to?
I'm just thinking how useful it would be to law enforcement if they could just specify a location, date and time and go through pictures taken there for evidence.
I hate this whole idea. It's inevitable that it will happen, but I hate it nonetheless. You wont be able to go outdoors without being constantly recorded by strangers. My only hope is that the popular services manage the technology such that we don't end up with databases that governments, or organisations can read private data from.