By far, the biggest crowd pleaser at last week’s TechCrunch50 was a demo by the Japanese startup Tonchidot for a mobile social tagging product it is developing called Sekai Camera. The Japanese CEO Takahito Iguchi overcame a very noticeable language barrier and deflected serious questions from the judges through sheer will of character. He had the audience roaring in laughter and rooting for him, as he answered lengthy questions about how his service would actually work with brief responses such as “Imagination!” and “We have a patent.” When judge Rafe Needleman suggested that Google would buy Tonchidot, he objected: “Never!”
The original video portion of the demo has been watched more than 108,000 times on YouTube. But I’ve embedded the entire demo above, including the follow-up Q&A. The entire video is 17 minutes long and takes about a minute to get started, but it is captures how Iguchi had the audience, and even the skeptical judges, eating out of his hand, despite a limited grasp of English. The Q&A starts about 8 minutes in. Iguchi was particularly adept at exasperating judge Tim O’Reilly. There is also a shorter edited video on YouTube of the judges’ Q&A with subtitles (embedded below). Anyone doing a demo can learn from this: keep your answers short, don’t drown in details, explain how you will change the world.
The demo starts with a video showing how Sekai Camera uses the iPhone’s camera viewfinder to overlay tags and information from a database onto objects in the real world. (I am not sure the iPhone SDK allows developers to access the camera in that way, but never mind). Pan the camera around, and different tags will pop up for stores, products, even voice or text notes left by your friends. “Look up, don’t look down,” Iguchi kept telling the audience. Sekai Camera includes an “Air Filter” that lets you see just the tags you are interested in. It is designed to be an interface between the real world and the Web.
We allowed a canned video demo instead of a live demo in this one instance because it was really the only way to show the full capabilities of the product (during rehearsals TC50 organizer Jason Calacanis did see a functioning prototype actually working on an iPhone). The judges, though, wanted to know who will be doing all of the tagging, and whether this thing really works. To which Iguchi answered, “Join us.”
Interesting clip!
Google doesn’t need Tonchidot. They’re actually in negotiation with another company called Enkin, a team of 2 german guys (ironically also based in Japan), who developped a WORKING PRODUCT for Android.
YEs, unlike Tonchidot, Enkin has a working demo for Android and it’s really impressive. Check it out here: http://www.enkin.net/
Enkin was not selected by Google as one of the Top-50 finalist (Android Application competition), because Google is actually trying to buy the company. And if I was Steve Jobs, I’d be also really interested by what these 2 geeks have developed. Seriously, think of the potential…
Google could and probably will develop a website (a-la AdWords) where every business in the planet will be able to geo-tag their shop, service etc… for free. But the trick is that you could tag not only your business name, but also place ads. And of course Google will charge for this service…
In Enkin’s video demonstration, we see an example with a hospital and a subway station geo-tagged and visible live through augmented reality. Now replace this with a restaurant or a shop. This shop pays Google to advertise so below their business name, the tag would also display for example a promotion for an article. Just imagine the potential!
This concept has the potential of becoming something similar to what adwords is to Google: a drastic profit machine!
On top of that, Google could license the service to TomTom or other GPS manufacturers, so in your car you will actually see in real-time the directions and tags through a live video, instead of the boring 3D-graphics we all know.
Chris,
do you work for Enkin, or are you an investor? I’ve seen your comments touting Enkin in a few places other than techcrunch….
Nevertheless, I do concur with your point of view. But please indicate your conflict of interest when commenting in the future.
and another way to build the database (should Tonchidot or Enkin decide to do this business alone without google), is eventually to partner with Yellow Pages in every country, because they already have sales teams taking care of business ads.
i really don’t understand why arrington schonfeld are skeptical about the ‘world changing’. this doesn’t matter. if your business moves to a different location in your city, you’ll simply update your geo-location on the Enkin (or Tonchidot) website. it’s that simple.
i think the confusion comes from the fact that the judges didn’t understand that geo-tags would come from businesses themselves. users tagging is just extra.
You, seem to throughly understand technology *and* the applications thereof. And, most importantly business sustainability.
Well said!
I was stupified listening to the (ridiculous) questions some of the panelists asked them. It was an incredible look into just how stupid so called experts (panelists) can be. I was surprised to learn they did not win TC50. This is here and now technology, complete with an excellent business model.
The Tags (watch the video first) will come from a myriad of sources… Geotagging, RFID, Bluetooth,etc. The tags will be user generated, or from a database, or sorced from spidering the net…
As a business strategist it was VERY clear to me what Tonchidot would need to do to be successful, and how to monetize it. What the heck where these panelists thinking…the audience (as measured by their applause got it right, the so called experts got it wrong).
http://www.twitter.com/A_F
I was stupefied reading the (ridiculous) comments one of the TC readers posted. It was an incredible look into just how stupid so called experts (TC readers) can be.
Perfectly valid question to ask how their system responds to the real world changing all the time. Yes, their demo was incredibly cool, but the panelists aren’t there to oooh and ahhh about the companies, they are there to ask tough questions. Considering none of their questions were really answered well, I can see why the didn’t win. But I do agree maybe the deserved to based on their tech.
real world is changing?… so what? this was a stupid question. Because the tags will be updated in real time through a website where business entities will geo-tag themselves.
Thank you.
I didn’t understand the questions either.
Put a compass in it and it could know which way its pointing… or save the pic with the tag so you could figure that out.. and then you could know if they changed the store.. or what you were looking at.. Didn’t seem that hard to imagine…
and Where is the data coming from? In my head I could easily figure out there would be layers, some search info, some user tags, some review site info, some phone directory info.. you could switch between them.
It was amazing! I was blown away.
I’m facing a similar problem at http://www.microspaces.com . I think microspaces is a profound and monumental innovation. But I am starting to build applications and show each use case, because until people see it in action, they aren’t imagining it. Most stop at ‘clever’ and ‘interesting’ and don’t seem to make the leap…
The questions made perfect sense since they wanted to understand how it works. Personally I think the demo they show in the end is how it actually works. Ie it downloads a list of all tags and stores in the area and display them on the screen. Kind of like a search on Google Maps, but without the map part being shown.
The augmented reality part was faked, and that’s what the panelists were asking about. Searching images for logos and such has been done by other projects before but it doesn’t work well and it doesn’t scale well.
You can’t use a compass ebcause the iPhone doesn’t have one. RFID and Bluetooth doesn’t give you any information on where in the image the devices are, so those are not useful either.
But using GPS to find stores and tags in the area is very possible. It has also been done by other programs before. The augmented reality bit is not possible with the tech they demonstrated, and the flippant replies made it quite clear that this was not a functional demo but more of a concept sketch.
Brilliant. Simply brilliant.
exactly one brillant idea….
Cool app….
Too bad the language barrier basically killed these guys on the spot. Their technology is interesting but its only geo-locational posting. There’s no QR code reading happening here, but that’s not to say that won’t come later.
I’m sure Google plans on rolling out something like this on android. But like anything we’ve just got a lot of cool tech to look at here, no real consideration on the actual use case for consumers. We’re assuming this all would rock and it would for the first 5 minutes or so for the average joe. Early adopters on the east and west coast will lead the way tagging the crap out of everything.
That one view of mass banner ad blindness is hilarious.
there is nothing on site, only contact email
This is cool stuff, straight from your favorite cyber punk novel (Tim even called that out on the panel), but I think its pretty stupid to just believe everything based on a video.
What are they ACTUALLY doing? The concept is great, but how have they executed it?
I gathered it was mostly GPS/location based and not image recognition based, so already its a bit of a hack. Their questions about the size of the database and how you would populate it are definitely relevant. Even Yelp and other “hot” review sites don’t have enough user generated content outside of some major cities.
I was disappointed that there was no translator, and I found it kind of insulting that TC would put this guy up there without being able to answer those questions. Great entertainment though!
IMAGINE
A lot of unanswered questions, which is why they didn’t win a prize. As for the translator, the company had an American executive on stage at the end of the video but decided for some reason not to put him back up during the Q&A. It was their call.
Yeah I think not putting that translator back into the cue was probably their downfall. I got the energy of the guy, and the IMAGINE, and the demo and technology was amazing, but… he couldn’t answer any questions around it.
And the thing is there are going to be a dozen players in this space all with own data “layer” like view finder apps. Google out of all of them, has the best chance to really win.. those bastards.. oh wait and then there’s Apple.
Maybe they didn’t want the prize money. They just wanted publicity to show off their technology to companies such as Google etc..
You guys could have done a better job editing the video. Like normalize the sound, maybe sub titles. All I heard so far about TC50 was “so, so” and this really seems to be one of the pieces sticking out. It would have been nice to “understand” more.
With no offense to the presenters, of course.
Will this incorporate facial recognition too.. so I can point it at people to get all their public INternet info?
People may scream privacy issues, but uh your in a public place and after all you placed all your info on the net anyway.
its too bad the audience audio wasn’t as loud as it was in person. People were literally falling out of their chairs (me included) with some of his non sequitur responses.
i can IMAGINE!
what an amazing device.
After seeing the demo (which was awesome), I thought about the feasibility of Tonchidot…here’s what I came up with:
http://martinma...idot-dissected/
I look forward to seeing what third-party reviewers have to say about this particular company’s implementation, as the demo looks implausibly good to me. My guess is that this will the top app of 2015, not 2009.
Cool service, but the concerns are valid. In order for the service to be useful, the information must come from a valid source and be updated to reflect changes to the physical world. The idea of UGC and community is great, but once someone looks at information tagged by someone out of their network, there is no way to know if it is sound. Perhaps a good model would be to allow business owners to tag their own stores and allow users to tag UGC on top. Users could check the UGC from people in their network, as well as, from others that could be vetted using a community based voting system like Digg. While business owners could update the tags to account for change, the UGC would still be outdated and there would need to be a system in place to delete old and irrelevant information. Either way, collecting relevant data requires a large scale initiative and properly accounting for change requires visual recognition software that doesn’t exist yet.
I would be a lot of money that this product is a fake.
Unlikely, its really just a better UI or graphical view of what Graffitio does on the iPhone.
Though they don’t really go into how accurate the tech is, thats probably a nice big messy grey area there.
be => bet
Who knew that the loophole for becoming a TechCrunch favorite is to severely limit your English fluency? This has all the characteristics of another “I cahn haz” or “all your base are belong to us” movement. Kids, start making your tee shirts now. Join us.
Idea that will certainly see the execution from many sources (including Google). However, the first “application” will be the spam flood of the system and it is unlikely that anything else would ever dominate it but spam. Predicting similar fate to it as the Secondlife…
Naw now its no second life. I mean its the first logical next step toward a type of location based augmented reality. I mean this is something that IS going to happen. How it looks, what form, and is it really useful to people that really the question.
sad is when some beauty pageant has interpreters standing right beside the bimbos, but a highly vouched for tech event goes on even after the very presenters are not getting the questions at all, and after this is evident, the entire thing is carried forward just for “entertainment”. SAD. those were attendees whistling?? at a tech event?? and the panelists, aaah, RIP ole’ people.
maybe we don’t hear of some tochnidot or whatever again[we will def hear of it some more times than some tamer yammer], but now in the above video we have seen how serious TC50 is with its “show-casing” business. good luck surviving now.
It demonstrates a really early, complex idea in a believable, easy to grasp way so it’s a great presentation of the product. Even with the language barrier, you can tell this guy is a smart guy. But, it’s early for mobile, even with the promising early adoption/use in other countries.
TC50 nabbed some really cool, clever ideas this year.
Think about how this thing can be so useful when you got a new city!
When you get to a new city*
Very cool _demo_, but … reminds me of those great autonomous navigation demos from the AI labs in the 80’s: “[whispering] assuming we have a completely accurate map of a world which never changes and we always have perfect location information, [shouting] look, we’ve solved the general navigation problem.”
Don’t get me wrong, I love it when people go after the really big problems, and I hope these guys find success, but they are battling the laws of physics here. That is, assuming you can get location and orientation data with the required accuracy (not a given, I think), how is _everything_ going to get tagged AND hyperlinked AND validated AND tracked when it moves/changes/becomes invalid? The DB and bandwidth implications alone are staggering, even if they limit their service to a restricted domain, like “stuff found in electronics stores.”
But, I’d love to be proven wrong. Like Tim O’Reilly alluded, we’re all waiting for our Neuromancer/Snow Crash moment — good luck, Tonchidot!
The panel experts were idiots.
well MrCashyCash, next year you should accept our invitation and show all those losers how to properly judge startups.
Accept Michael’s invitation MrCashyCash
I would pay and fly over from Australia just to see it !!
I think we can make an entire Seinfeld episode on this
Seriously though, great concept. It may or may not revolutionize the way we search. Of course, it all depends in the details and execution.
I am so impressed…just one brilliant idea after another. All I can say is when Sekai Camera and others like it make this the norm, what’s the next evolution?
The 21st Century rocks so far and I am getting excited at the brilliant ideas that are revolutionizing how we work, play and interact!!!
these guys should have won hands down y didn’t they??
even without the proper QNA
they would not have come all the way if they didn’t believe in their product in the first place plus
The energy of that guy and his passion for his offering was outstanding I guess with english translation their enthusiasm would have been lost and it would look like a boring college lecture and thats the reason why they kept the translator out of the picture for the QNA
With Japan being the early-adopters on mobile technology, I wouldn’t be surprised if even all of the forces of Microsoft and Google couldn’t catch up with this application if it builds an inimitable base in Japan and then spreads from there.
“Iguchi was particularly adept at exasperating judeg Tim O’Reilly.”
What’s a judeg?
Why was the full article corrected but not the post on the main page?
I think at the next TechCrunch50 I am going to demo a spellchecker. That clearly has not been invented yet. Amazing we’re talking about Web 2.0 and beyond and shit like this slips in ALL THE TIME.
so in the future we will all be walking around with apple’s new IPhone glued to our eye ?
the best q & a. ever.
join them, bitches!
Similar concept from a french company : http://disqvery.com
This is just the tip of the iceberg, good idea , but this can be taken even further
Most people on the panel are very narrow , since no one really addressed the business impact this application can have
Note : data is gathered from many sources ,obviously all information is really viral on the internet , you can draw on so many sources to populate a data base , just see how your own company builds data … DUHHHHHHHHHHHHHHHH
They won’t go with Google, because another company is far ahead of them and Google is most probably buying them: Enkin
It’s a working model built on the Android platform:
http://www.enkin.net/
Encompasses GPS, accelerometer and vision technology all together – that’s the key.
German guys who are living in Japan who also have a tad of trouble with English, but who can blame them, what with them knowing two other languages first.
LBS is obviously cool – but GPS – *indoors*?? Since GPS requires at LEAST 3 satellites in line-of-sight, I have to assume it “would” (if it were actually built) have to either use “dead reckoning” w/ the accelerometers, or combine that with the imager pattern recognition – which they aren’t claiming.
I’d love to see more, because what I’ve seen seems to clearly be a “cool idea”, not a technology…
oh those “crazy” japanese!
annoying, condescending laughter of the host made me want to sock him in the face
actually don’t need to worry about changes of the physical world, there are many solutions… shop owners definitely has the motivation to change that if they changed locations….
dont forget the power of “imagination”
Great application, some concerns about how accurate can it be when something changes. Definitely something from the future.
Great for wayfinding within offices, malls and tradeshows geared towards tech-savvy users (who happen to have an iphone or similar device) where the site owner provides the “credible” tags. The site would have to find a return on their time invested in adding tags; it would have to be subsidized by advertising. Can you image maintaining all of those tags?
Could be a problem trying to point at a location outside in a snowstorm or rain when you can’t get your phone wet.
Can you imagine if you could point the camera at a person, take their pic and attach their tags to their pic. I wonder if it would work with people?
So many possibilities. Let’s see, in 10 years in might catch on. I’d place my bets on Google streetview with tags embedded.
And here comes the Soundtrack for Tonchidot Madness:
http://files.gs...ve_a_Patent.mp3
N-joy!
OMG, what a retarded Q&A lol.
I should try that next time investors raise objections, see if they laugh and shrug it off and invest
.
I am still trying to load the initial video with the demo. The idea sounds great, but there are going to be a number of issues to over come depending on just how powerful/useful they want to make the service.
If the service simply shows data on buildings right in front of them at a street view level then they will need to have a sensor in the camera to see the angle of the camera and a compass to see which direction it is facing (not impossible), or they will need to have photo recognition software which will be cheaper in terms or hardware but will require some fantastic software.
But who is going to walk down the street looking through their phone and pointing it at buildings to find out what they are and what they offer.
If they take it to the next step and let people see down a street or through buildings for X meters (with the tags being smaller to represent how far away they are), then this will also require the above, but still has an issue to overcome: Where the tags should be placed in the image will depend upon how far above sea level the phone is as opposed to the base of the buildings. E.g. the phone use may be looking up a street that has a steep incline, and so the phone needs to know that there is an incline otherwise it will place underground :p.
I look forward to seeing how this develops.