Hacker News new | comments | show | ask | jobs | submit login

The thing that bothers me about Qwiki is their "technology" is just a smoke and mirrors reproduction of things we see in the movies when a robot or A.I. talks to the characters. It's made to look and sound like A.I. except what they're doing doesn't involve machine learning or anything A.I-related. It's a simple mashup with a well-oiled marketing campaign (Techcrunch) behind them.

They raised a seed round of $1.5M to build a proof of concept and have now doubled down and raised $8M. What is bothering about that, isn't it ideal to "get something out the door asap" for this ambitious a goal (a visual search engine)?

It depends. Sometimes the goal of raising money is to give yourself enough time to tackle a really hard problem. QWiki's got huge upside so if they've got an innovative approach and a credible plan, $9.5M isn't an obscene amount to spend on the early stages.

I agree to some extent, but wikipedia and similar sites are what A.I. will use to "learn" in the future.

If this were 1990 or I didn't know about wikipedia, then I would be very impressed and think this was true artificial intelligence.

Nothing is being 'learned' here, the only thing being produced is a rote repetition of an article after it's been passed through a mediocre TTS engine.

If this were 1990 or you didn't know about Google, then a simple web directory connecting you to various sites would be very impressive.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | DMCA | Apply to YC | Contact