Some months ago, a Japanese startup called Tonchidot showed a demo of their “Sekai Camera” (Sekai means World in Japanese btw) to the crowd at Techcrunch50. It is a pretty amazing video and got a lot of people excited, even if they really didn’t understand what they were seeing, or how this sort of technology works (or could work). Augmented Reality is very new and is still beind defined and redefined as we explore and experiment. I wasn’t impressed with their video then, and I’m not impressed with their new demo now. Read more for why…
This is the first video (which is mostly nothing more than pre-recorded video with composited effects)
Yesterday, Tonchidot organized an event at some exhibition in Tokyo and invited some press folks for a real demo. Clink the link and read the article.
My initial thoughts about Tonchidot were more than skeptical, and I felt like their Techcrunch50 presentation was misleading…you can’t show a conceptual video and present it as a technology demo, and I’m surprised they were picked to show anyway, without any clear business model or explanation of how their tech could or should work. Ideas are a dime a dozen…show me something that works (even if it is barely a proof of concept).
Anyway, it seems that Tonchidot, as a company, is real. Their second demo shows that there is something here and it is worth keeping them on the radar. On the other hand, there are a lot of holes that need filling. They claim that only GPS is needed, which is fine if you are just linking a location to media tags that people have uploaded in that specific spot, combined with a interactive interface to select the tag you want to see up close (the second video shows this). They claim that no image recognition technology or wifi-triangulation is needed, and I think in this case that is probably true. But then, what are they really doing here? How is this different from a dozen other projects out there that tie some type of media to a location? Isn’t this pretty much similar to putting a GPS coordinate on a picture you have taken and uploading it to Flickr? (letting people in that location see what photos are tied to that spot). It seems that the bulk of the value here is the interface (shown in the second video).
Definitely not enough to get impressed about or build a company around. What happens if you have 500 people all putting tags in the same spot? What is the value in scraping tags from various web services and websites? Do you really want to navigate through floating spam, random tags, and 3D google adsense?
The new Tonchidot demo is a far cry from their original video and almost a completely different thing. Yes, it does look cool, but ultimately this is going to fall into the novelty category unless Tonchidot kicks things into a different gear and gets back to their original concept (which seems, to me, that they have no real idea how to build).
That’s what I think.