Phone GPS, in my experience, really doesn't. It is inaccurate, oftentimes 100ft off, and certainly drains battery faster than just about anything else.
What part of it works, on a phone, for applications such as these? Or were you just saying, in general, GPS technology does work?
I think GPS inaccuracy was the #1 reason Foursquare beat Gowalla. Gowalla used to enforce strict accuracy on check-ins and it rarely registered you in the place you were at.
GPS doesn't "work" in the same way that he is referring to. For the purpose of identifying a point on the globe, with line-of-sight to the sky - GPS works fine.
But, what is really being talked about here is much deeper spatial awareness based on a range of sensor and relational data modeling.
They are using GPS, direction, lighting, sound and finally - actual images to determine all the spatial attributes of where you are, what your surroundings are and whom else is in your vicinity -- even if you don't know them.
This is stunning.
They will be able to know which parties are groups - where they are in relation to one another etc.
If they are already using all those data-points. I am absolutely positive they are will have audio voice recognition - keyword recognition in the future as well.
There are incredible uses for this, such as "here were some highlights of the parties conversations"
What we NEED is for someone to sniff the backhaul from the client and lets look at exactly what data is leaving our devices.
In my experience, A-GPS works great. It’s hard to get a GPS signal in urban landscapes but there are tons of WiFi access points around which are very quick at determining your location and are reasonably accurate. I really don’t know what other sensors could bring to the table.
I think the worst thing they can do is come out and try to defend it. Instead, they should probably internalize the feedback and keep working on the product, so they can come back in a few months and say "See? THIS is why."
I don't think I ever heard any Twitter execs defending their product when everyone was saying it was just for posting about your egg sandwich from the toilet.
He actually offered no defense. I decided not to focus on the interface/user experience because everyone else did. He pretty much compared it to a venue that opened before it was ready. I decided to question them about the tech, because that seemed to be the part nobody was really discussing and I figured maybe all this money was because there was some interesting tech at its core. I think there is.
I'm clearly in the minority here, but damn if some of this doesn't seem very cool. Especially for the middle/high-school crowd, and college kids, who are friends, generally, with people geographically close to one another.
If they sort out the issues plaguing 1.0 (and I think is clear they will), this seems like something that would a least catch on with that younger crowd.
Plus, cool tech from what they said here. I'm glad to learn a little about it.
Sometimes when I see these investments I wonder if there's just a different opinion about what a lot of money is. To me, $3 million dollars is a lot of money and I'd have trouble spending it. But maybe to a VC, that's barely anything and an investment of $41M to people they've worked with before, trust, and who have a pretty good prototype/story is perfectly reasonable. That's why I haven't paid much attention to the griping other than to see the headlines.
I think the point he's making is that they are going to try to figure out your actual real-world social graph based on who you associate with and where. I'm sure that was the killer part of the pitch.
If you forget about the photo sharing app for now, and imagine they do manage to create such a graph - it seems to me that it would be a lot more valuable than what Facebook calls a social graph - which is really more like a social game.
Imagine being being able to identify the people who are 'hubs' and 'influencers' in real life as opposed to online. Now imagine you can predict where they are going to be hanging out so that you can promote things to them, and then track the fanout as they meet people in real life. That is what Sequoia think their $41M might get them.
I agree with this 100%. I'm really, really not sure I want anybody to have that kind of data about me - but I'm equally sure that most people are going to love it.
All the bang of Facebook, without any of the drag of data entry? And it's always freaking right, assuming all the people on your morning train into the city don't get added.
They are so right to pump money into this. Although I have to admit that the decision to spend $350K of it on an entirely unremarkable and unGoogleable domain name sort of leaves me wondering.
But they're in a situation where they need a compelling application first. You can't get the data without multiple people opening up the application. Color is really not compelling.
This is a classic problem I have in my own research at a university: in order to test the cool stuff, you have to convince people using stuff they already get. The convincing is way, way harder.
Maybe Color is just one of a suite of apps. Maybe the idea is to license the patents to Apple or Google so it all happens transparently (but the idea that my phone is silently collecting and transmitting data about me all the time makes me want to puke). However, what they've got is going nowhere fast.
Agree 100%. The trick for them is to build a compelling application.
I faced a similar problem with delineal.com. Like Color, my groundbreaking stuff is in the data choices, algorithms and data filtering and in my generalized approach to solving problems. I don't lead with detailed discussions of the technology because A) the vision and direction tell a more entertaining story B) The matching and filtering algorithms do not a product make. I need to build products that solve real-world problems to generate revenue.
I like the recognition they're getting for their technical work. It bodes well for me...and provides some nice hints about how to better frame my pitch... because I can say "I have technology AND useful apps".
If Color gains any initial traction at all in hopes of making a social graph, I wouldn't be surprised to see Facebook immediately going after it as well to cover their grounds. They have robust photo capabilities and strong mobile integration already, and it makes sense with their location check-ins to events and places.
assuming, of course, that the people you interact with all have smartphones with the color app running. as far as I know "passive" social networking or "let's interact with strangers" apps never gain traction, because most people don't find that idea compelling. outside of the handful of places with some density of early adopters and technohipsters (SF/NY basically), I just don't see anyone using this over the much more familiar model facebook provides.
Huh. Bill credits DJ Patil with creating this patent-pending technology that checks locale, but DJ just joined Color and LinkedIn, his previous company, definitely has an invention assignment agreement...
The company ... sees itself as "much more of a research company and a data mining company than a photo sharing site."
Wow - I bet users will gladly line up to upload photo and video to a data mining company. From the start they plan on mining their users rather than serving them. It really feels like all marketing BS when the core of the business is to track, record and mine their users.
It may be good tech, but that's a horrible user proposition and it certainly begs for a line-by-line analysis of their EULA since everything you do with the app is public.
"Lots of people are trying to create location-based services and using GPS," said Nguyen. "The problem with GPS is that it doesn't work."
This is exactly my pointbelow -- the tech to incorporate all the sensor data you can get from phones has real-world consumer uses -- but it has tons of uses in intel as well.
They will get their patents, they will build the next google of social intelligence and they will license all this to the NSA -- if they havent already.
Sounds cool, yay. But no thanks from me.
Where I said:
"Here is some tinfoil for you...
Assuming they have a back-door API - this could be a freaking BOON for intelligence gathering arms that are seeking to know the surrounds of anti establishment types.
Further - wait until you apply spatial modeling capabilities (what was that MS tech demo called?) where you can start building 3d navigable models of spaces that all the data collected in that 100' radius allows.
Dont think the military is doing CRAZY things with optical intel:
"The system can also be used for general night vision; it can follow bats five miles away in darkness."
Google analyzes tons of data because they have too. As the parent post said, the "NSA analogy" is wrong because it has a very wrong proposition. Users use it because it solves some real problems for them. AT&T doesn't work for NSA, they works for customers and has a fiber to NSA as by-product.
You cannot compare a for-profit organization to a research facility (e.g. MSR). Such thing as you mentioned is aresearch project, which has much limitations and some of them wouldn't possibly be commercialized. It is the same skeptical as I have to hunch.com, but hunch.com solves some real needs whereas Color from what I see, is pursuit a magic ball floating in the air. And should I remind you that selling backdoor API is not a business, it is a crime.
It's not like this doesn't already exist. I've specifically read about startups doing this from the carrier side several years ago. Whatever it's in service or not is another question. Unfortunately I can't remember the article but there are a lot of startup in this sector e.g. some of these: http://www.iqt.org/technology-portfolio/index-alphabetical.h...
Yep. If Color was launched earlier (which it didn't) and was a hit (which it's not), I imagine what would have happened in Egypt if the protesters were using it to share their movement. Could President "Richer Than MZ" Mubarak buy the data from Color? If they can actually do what they say they can, this is really a company that can change history. Not in the direction that I want, though.
What's fascinating about this is how much Nguyen feels he has to say. In a way, this is a serious blow to their plans; they have to reveal their world domination plans early so that bloggers won't keep dumping on them.
That said, their world domination plans are pretty awesome. I bet they come up with other ways to get you to 'soft checkin' on their platform. Just betcha. Wouldn't you like to just open an app that lets you, say, ask for contact info of people in a room with you? Like bump, but working, and low hassle? You could use it at parties, at networking events..
I'm guessing the application thoughts just keep coming for the color team.
Unfortunately, what they really can't afford is to dig into a hole based on bad PR right now; they need people out and using their app, and talking about it, hence the opening up a little bit.
I guess the VCs think that if Color can get their patents the worst case scenario will be that Color can become a patent troll.
The $41mil shouldn't even be a topic right now because the VCs that funded them are some very experienced betters in the game. If they made this bet they must have seen something. Honestly I don't see anything of value right now, but then again I'm not into sharing photo's with strangers thing.
I also think they are not a social company like Twitter/Flickr but are hoping to generate enough data using their mobile apps to be able to identify various locations without gps and sell that data, sort of like that one company which maps out locations using WiFi access point names.
As for this new bubble debate and Color's $41mil, no matter what the situation you will always have a few people saying the market is good and a few people saying that the market is too weak/strong. It's like the glass half/full debate that can never go away.
It has patent troll written all over it. Right now it seems like they don't have much more than an idea for a product and an idea for a business plan. Even if their photo-sharing thing doesn't work, it's inevitable that others will be looking to build social networks from cell phone data.
6 patents pending? Is this all a big IP bet for them & their investors? That could really set back under funded entrepreneurs working on similar solutions.
"The more you interact with someone, the more persistent that connection becomes."
Facebook already has been doing this for a couple years, but if Color sniped them on the patent by a few months then that would justify the valuation alone. Plus they have a solid team with at least a decent chance of building something themselves, which is basically just gravy if the investors have a guarantee of at least getting a 2x return by suing Facebook and a couple other companies.
Another way to think about it: Color is "Bump" without the need to bump phones! Using various sensors, it can tell who you are next to, and then you can exchange contact info, photos, and set up an ad hoc connection. Another use could be to figure out what store you are in and offer you relevant coupons...
Interesting that these guys go heavily against the idea that you should release something early and iterate on it. Do people agree that it's good to front load all this tech or should they have been getting the photo app right and say used GPS for a first version?
77 comments
[ 5.7 ms ] story [ 127 ms ] threadValuation eh- "Our tech knows where you are". Maybe, but very likely not as well as you claim, and even if so...eh.
What part of it works, on a phone, for applications such as these? Or were you just saying, in general, GPS technology does work?
But, what is really being talked about here is much deeper spatial awareness based on a range of sensor and relational data modeling.
They are using GPS, direction, lighting, sound and finally - actual images to determine all the spatial attributes of where you are, what your surroundings are and whom else is in your vicinity -- even if you don't know them.
This is stunning.
They will be able to know which parties are groups - where they are in relation to one another etc.
If they are already using all those data-points. I am absolutely positive they are will have audio voice recognition - keyword recognition in the future as well.
There are incredible uses for this, such as "here were some highlights of the parties conversations"
What we NEED is for someone to sniff the backhaul from the client and lets look at exactly what data is leaving our devices.
I am sure they are a lot further along than vapor at this point.
I don't think I ever heard any Twitter execs defending their product when everyone was saying it was just for posting about your egg sandwich from the toilet.
http://www.penny-arcade.com/comic/2008/4/23/
If they sort out the issues plaguing 1.0 (and I think is clear they will), this seems like something that would a least catch on with that younger crowd.
Plus, cool tech from what they said here. I'm glad to learn a little about it.
Just saying you'll have tens of millions of users in 3 months doesn't make it so.
If you forget about the photo sharing app for now, and imagine they do manage to create such a graph - it seems to me that it would be a lot more valuable than what Facebook calls a social graph - which is really more like a social game.
Imagine being being able to identify the people who are 'hubs' and 'influencers' in real life as opposed to online. Now imagine you can predict where they are going to be hanging out so that you can promote things to them, and then track the fanout as they meet people in real life. That is what Sequoia think their $41M might get them.
All the bang of Facebook, without any of the drag of data entry? And it's always freaking right, assuming all the people on your morning train into the city don't get added.
They are so right to pump money into this. Although I have to admit that the decision to spend $350K of it on an entirely unremarkable and unGoogleable domain name sort of leaves me wondering.
This is a classic problem I have in my own research at a university: in order to test the cool stuff, you have to convince people using stuff they already get. The convincing is way, way harder.
Maybe Color is just one of a suite of apps. Maybe the idea is to license the patents to Apple or Google so it all happens transparently (but the idea that my phone is silently collecting and transmitting data about me all the time makes me want to puke). However, what they've got is going nowhere fast.
I faced a similar problem with delineal.com. Like Color, my groundbreaking stuff is in the data choices, algorithms and data filtering and in my generalized approach to solving problems. I don't lead with detailed discussions of the technology because A) the vision and direction tell a more entertaining story B) The matching and filtering algorithms do not a product make. I need to build products that solve real-world problems to generate revenue.
I like the recognition they're getting for their technical work. It bodes well for me...and provides some nice hints about how to better frame my pitch... because I can say "I have technology AND useful apps".
Wow - I bet users will gladly line up to upload photo and video to a data mining company. From the start they plan on mining their users rather than serving them. It really feels like all marketing BS when the core of the business is to track, record and mine their users.
It may be good tech, but that's a horrible user proposition and it certainly begs for a line-by-line analysis of their EULA since everything you do with the app is public.
This is the telling quote:
"Lots of people are trying to create location-based services and using GPS," said Nguyen. "The problem with GPS is that it doesn't work."
This is exactly my pointbelow -- the tech to incorporate all the sensor data you can get from phones has real-world consumer uses -- but it has tons of uses in intel as well.
They will get their patents, they will build the next google of social intelligence and they will license all this to the NSA -- if they havent already.
Sounds cool, yay. But no thanks from me.
Where I said:
"Here is some tinfoil for you...
Assuming they have a back-door API - this could be a freaking BOON for intelligence gathering arms that are seeking to know the surrounds of anti establishment types.
Further - wait until you apply spatial modeling capabilities (what was that MS tech demo called?) where you can start building 3d navigable models of spaces that all the data collected in that 100' radius allows.
Dont think the military is doing CRAZY things with optical intel:
"The system can also be used for general night vision; it can follow bats five miles away in darkness."
http://www.wired.com/science/discoveries/news/2007/02/72632
the military now has autonomous robotic helicopters that can visually track bullets in flight.
http://www.wired.com/dangerroom/2009/02/gigapixel-flyin/
Now they just need the general public to be a part of the larger sensor intel fabric."
If the photos are dense enough, there's enough information to stitch together the photos into a scene you can jump around. It's really cool!
EDIT: Here's a scene in Sydney:
http://maps.google.com/maps?f=q&source=s_q&hl=en&...
Google analyzes tons of data because they have too. As the parent post said, the "NSA analogy" is wrong because it has a very wrong proposition. Users use it because it solves some real problems for them. AT&T doesn't work for NSA, they works for customers and has a fiber to NSA as by-product.
You cannot compare a for-profit organization to a research facility (e.g. MSR). Such thing as you mentioned is aresearch project, which has much limitations and some of them wouldn't possibly be commercialized. It is the same skeptical as I have to hunch.com, but hunch.com solves some real needs whereas Color from what I see, is pursuit a magic ball floating in the air. And should I remind you that selling backdoor API is not a business, it is a crime.
If you aren't paying for the product, you are the product.
And yet people seem to like Google...
"It really feels like all marketing BS when the core of the business is to track, record and mine their users."
I can't help but think that this also describes Google. :)
Edit: the person that downvoted me should at least have some courtesy and explain themselves.
That said, their world domination plans are pretty awesome. I bet they come up with other ways to get you to 'soft checkin' on their platform. Just betcha. Wouldn't you like to just open an app that lets you, say, ask for contact info of people in a room with you? Like bump, but working, and low hassle? You could use it at parties, at networking events..
I'm guessing the application thoughts just keep coming for the color team.
Unfortunately, what they really can't afford is to dig into a hole based on bad PR right now; they need people out and using their app, and talking about it, hence the opening up a little bit.
The $41mil shouldn't even be a topic right now because the VCs that funded them are some very experienced betters in the game. If they made this bet they must have seen something. Honestly I don't see anything of value right now, but then again I'm not into sharing photo's with strangers thing.
I also think they are not a social company like Twitter/Flickr but are hoping to generate enough data using their mobile apps to be able to identify various locations without gps and sell that data, sort of like that one company which maps out locations using WiFi access point names.
As for this new bubble debate and Color's $41mil, no matter what the situation you will always have a few people saying the market is good and a few people saying that the market is too weak/strong. It's like the glass half/full debate that can never go away.
The Onion as always is on top of things.
Wasn't this the premise of the ending of 'Batman: The Dark Knight?'
Scary, hilarious, and amazing at the same time.
"Your elastic social graph can and will be used against you in a court of law"
"The more you interact with someone, the more persistent that connection becomes."
Facebook already has been doing this for a couple years, but if Color sniped them on the patent by a few months then that would justify the valuation alone. Plus they have a solid team with at least a decent chance of building something themselves, which is basically just gravy if the investors have a guarantee of at least getting a 2x return by suing Facebook and a couple other companies.
What real capital costs changed between the $14 million they were getting before Sequoia and when they got $41 million total including Sequoia?
If they build a dark-knight-like smart phone data miner, they can always sell it as a service to (or get bought by) somebody who already has traction.
Why care about traction if someone else might pay you to use theirs more effectively?