Techcrunch40 Conference Coverage Session 1 : Search & Discovery

I’m at the TechCrunch40 conference right now and Michael Arrington & Jason Calacanis are kicking it off. I’ll be live blogging this event, starting with this long post - Techcrunch are also covering it. If you’re around, please say hi, and we’d love your poker chips (handed out to the best companies presenting in the DemoPit). Synthasite will be demo’ing tomorrow - and we’re very excited to be presenting!

Powerset

The guys from Powerset take the floor. They are a natural language search engine that’s received a lot of international acclaim. They’ve just announced the launch of PowerSet Labs - a way to :

• explore technology demos
• share ideas
• feed the learning engine
• improve your search karma

Sounds exciting - and definitely leaning toward utilizing the knowledge that’s out there in the search industry. Here’s a YouTube video on Powerset:

They demo some really great search technology, which is not linked to keyword based searches, but instead, semantic mapping of concepts and themes.

Cognitive Code

Next up are the guys from Cognitive Code, who are an artificial intelligence company.

• Practical, conversational artificial intelligence applications
• Development platforms for context-sensitive ai apps

This is very interesting - stuff like out of space age movies like Star Trek. They are talking to a computer and asking it questions, and it is understanding the questions and replying - very impressive, but I’m not sure how much of it is prepared and how much of it is scripted. They claim that there are no scripts running, which if this is the case, then it’s very impressive. They are trying to embed Silvia into mobile devices, toys, web servers and custom electronics over the next few years. This stuff looks far too early stage for prime time, but that’s just my opinion.

Cast TV

CastTV are up next. They’re a video search engine that indexes YouTube, Revver, and all the other video hosting sites. They are crawling and indexing videos from around the web. They aim to index every video on the web. They demo some comparative searches from Google & Yahoo, and compare with theirs, which has a far fresher index. Very interesting, again - not sure how much of this is pre-prepared.

Faroo

The next guy up is from Faroo, which is a P2P search engine. He starts out with a monologue tone and some fantastic claims of being able to index the entire Internet. Sounds very Dunning-Krugeresque to me. The concept is good, however, I’m not sure that these guys can pull it off, based on the presentation that he is giving. This is a really good example of how marketing is key - he speaks with a heavy swiss accent, and it’s difficult for me, and the rest of the audience to follow. I don’t think anyone understands him. Also, again, the idea is good, but I don’t think the implementation is done properly or with a large enough value proposition. He also claims to be able to save billions of dollars.

Viewdle

The guys from Viewdle start off with a nice presentation, which explains what they do. Basically, they index videos frame by fame, to capture people who are in the video, and keep and index of who is in a certain video. They allow you search for people and see which videos they appear in, by facial recognition. Very cool technology.

Q&A:

Chris Anderson, author of the long tail & Wired editor, ask the guy from Faroo on how the users benefit from allowing the web to be indexed from their PC’s. “I think it’s good for users because they’re using search together” - something got lost in translation here. I agree with Chris - it’s a bit wishy washy as a business model currently, either that, or I don’t understand what he is saying. Om Malik has a go at him - this is not going anywhere.

If you enjoyed this post, make sure you subscribe to my RSS feed!