Building something like Skype, Google Talk, or even Twitter is hard to do right at large scale. What real time service would you build if the technology was not a challenge?
In large cities in China they effectively have this through a million people being employed as couriers. You order something on taobao.com (kinda like amazon) and someone calls you up, asks you where you are, and half an hour later they show up on a motorcycle wherever you are and you hand over cash for your order.
It's basically UPS entirely staffed by bike messengers. It's realtime and it's awesome.
Was thinking that exact thing yesterday when I had to go to the mall for last minute presents - had major crowd/chaos anxiety, turned and bailed. If Amazon could do real time shipping, it would be a miracle. Happily pay a premium for it. They could have distribution hubs in all major cities...
Custom reality tv. Users pay to see a day/hour in the life of another user - possibly a celebrity, though other fun people would be popular - who wears a tasteful camera strapped to their head or neck.
Witness an afternoon as Michael Schumacher. What's life like as an Oxford student? My name is Joe and tomorrow I will go free running in the city - join me.
Madness! I guess my "idea" depends more on engineering advancements then. When I wrote the parent I'd just finished watching a series of the excellent Peep Show (filmed mostly via helmet cam), and I was thinking how I'd like to be able to jump into anyone's head and observe their life in that way.
An automatic, voice based translator: talk in one language on one end (to an iPhone or something) out comes another language and subtitles in another :-)
One thing that's interesting about this is that it's not too far off. We have speech recognition technology, computer voice technology, and translation technology. They need to get better, and faster, but they're there.
As of right now, I can find out a lot of information about a company by paying for the data from a data provider like Bloomberg or Reuters. This data, however, should be free and well-organized.
I would like to build a search engine that indexes and organizes regulatory filings, investment message board posts, investment blog posts, the portion of a company's website dealing with investment relations, and economic news.
You input the company's trading symbol to bring up all the relevant information about the company. Furthermore, there should be an API so that people can use the index in their own applications. I have no idea where revenue would come from.
Have you ever used Bloomberg or Reuters? Google/Yahoo Finance are far from useful for anything but an overview of a company. The service I am talking about would be aimed at more professional users.
Yes, I have used Bloomberg quite a lot, although I spent most of my time with its painful API. I just wanted to get your perspective on what Bloom feature you'd like to see in a web site/search engine.
Well, to start, if I could get correlation data from Yahoo/Google Finance that would be a great. In addition to that, if I could get my hands easily on data concerning pending lawsuits, better industry data, and better regression tools (though this may tie in with the correlation data requested above), that would also be wonderful. Those are just a few key things that would be nice for equity. Also, the data concerning anything outside of equities is severely lacking in Yahoo/Google Finance.
I agree with you completely. My question is what would you build if the technological barrier of building a massive, reliable real time service was not there. Everyone is talking about the Real Time Web (even organizing conferences) but Twitter is the only near-real-time service that comes to mind.
I've thought about making a real-time forum: basically the same thing as an ordinary forum software, but with presence capability and posts happening in real time without reloading. Or, if you look at it in another way, it would be a chat room with persistence, result is searchable, possibly threaded.
The hope is that conversations in such a medium would be something in between regular chat and regular forums, and they could be compelling enough.
The technology is not a big issue for building this, the Tornado framework has a chat example from which you can start. Text messages are short, they can be kept in memory and saved to disk once in a while either DIY or using something like Redis.
Yes, Wave has all the features, but seems focused on collaborative editing, rich media, send and synchronize individual keystrokes etc. - so I imagine a lot of the design decisions would be different.
I don't think I'm going to build it though - the type of forums I am interested in actually benefits from the perception that you have time to think through what you are saying.
I'm not uninterested in building something like Skype, Google Talk, or Twitter because of the technical issues. I'm uninterested because I have no good ideas (or passion) about solving a significant problem with them.
There are lots of things I'd do if you gave me technology that does not really exist. Stuff like speaker independent speech-to-text that's as accurate as an expert human could transcribe. Or language translation that was as good as an expert human. Or almost anything that let me have the power of a human brain for free (no, not MTurk).
Text to speech so that I don't have to consume realtime data like RSS, Facebook, Twitter, etc. while actively using a device. I'm just gonna wait for Google or someone else to solve that one.
