Fred Wilson: “When will people start building apps/services that sit on top of multiple APIs?”
It’s a good question, and the answer is — we’re already doing it. The services are now so reliable and flexible that you almost forget how complex the systems are.
Let’s look at a Twittergram scenario:
1. Chris registers with the Twittergram site. That’s App #1.
2. She takes out her iPhone and opens the phone app. That’s App #2. (Yes, the mobile device is also a computer, it’s running an app, with an interface, it can dial a number and transmit audio. It’s old fashioned, but it works great.)
2. She calls BlogTalkRadio, 646-716-6000. That’s app #3. Records a 30-second message.
3. She hangs up. BlogTalkRadio calls the Twittergram app. App #1, again.
4. Twittergram calls Amazon S3 to store the MP3. App #4.
5. Twittergram calls TInyUrl to create a short URL for the MP3. That’s App #5.
5. Twittergram calls Twitter. That’s App #6.
So there’s a simple application that uses six different apps/services.
Another example, the Flickr-to-Twitter integration, also part of Twittergram.
1. Randy registers with the Twittergram site. That’s app #1.
2. He goes to the park, sees a beautiful flower. Takes a picture with his iPhone. That’s app #2.
3. He mails the picture to Flickr. That’s app #3.
4. Twittergram is monitoring Randy’s RSS 2.0 feed on Flickr. Some people might say this is another app, but let’s be conservative. App #3, again.
5. It notices a new picture, grabs the URL, calls TinyUrl. That’s App #4.
6. Grabs the title of the picture, appends the TInyUrl, sends it to Twitter. That’s App #5.
Another app that uses five different apps/services.
We’ve been doing this stuff for a long time, all the way back to 1998, when XML-RPC first came online. It’s always been about delivering functionality, quietly and reliably, to users.
It was intriguing for a day or two, but now it’s clear that the Leaderboard was the dumbest idea ever, because now more than ever, people are gaming Techmeme so they can climb the list.
Reminds me of something Ted Turner once said about how the Forbes list of richest people in the world was the worst thing ever for philanthropy. If you’re super-rich, now you don’t want to give it away because when you do, you move down (or off) the list.
Techmeme was already severely polluted by people saying stupid shit to rise to the top of the page. That was an ephemeral high. Now there’s a way to accumulate points toward more persistent rank, and everyone who isn’t on the list, wants to be on the list.
Or Scoble, who started on the list near the bottom — is rapidly rising. How’s he doing it? By saying extreme things that people will react to. That’s how you get points in the Techmeme universe. Scoble ain’t no idiot. If he wants to rise on the list, he rises.
A mind is a terrible thing to waste.
Well, this ain’t blogging, and we’re still getting ready to start a war with Iran, and the stock market is still acting weird, and there are still big ideas out there to pursue, and now Techmeme isn’t even worth reading when the top item on a weekday is guaranteed to be some idiot procliaming himself king of the hill. It’s worse than AM talk radio.
PS: A piece I wrote in May offering a vision for “Web 3.0.”
PPS: Mike Arrington weighs in. “Gabe sometimes edits stuff like this off of TechMeme to keep it stocked with real news.” Hmmm. I’d be surprised if that were true.
I was subscribed to too many podcasts, my old podcatcher was downloading far more than I’d listen to.
So when I started the project to write a new podcatcher, I decided to start with a clean slate, and go for a minimalist set, the ones I really listen to, and add slowly, and make it very easy to remove or suspend a feed.
And as I announced on Tuesday, my podcatcher has a Twitter account, if you’re interested in background programming while you twit the day away.
And yesterday I took another item off the to-do list, and came up with a public web page that shows my podcatcher’s discoveries in reverse-chronologic order.
This will develop of course, what you’re seeing is totally pre-alpha, not even 1.0.