I’m at Twitter’s developer conference, Chirp, in San Francisco. Twitter’s co-founders Biz Stone and Ev Williams, director of platform Ryan Sarver, chief operating officer Dick Costolo and director of product management Jason Goldman are on stage taking live questions from the audience. (These notes were taken live, so don’t rely on them as verbatim text.)
Q: Will there be an Android app?
Williams: Yes, and it will be awesome.
Q: Will it be acquired or developed in-house?
Williams: We don’t know yet.
Q: How will Promoted Tweets be delinated?
Costolo: It will have a different color and marked as a promoted tweet. When we roll out syndication there will be clear guidelines on how promoted tweets should look.
Q: Who gets access to the firehose and how does that work?
Sarver: We don’t have a large team and right now it’s done through deals. We’re backlogged. It’s currently a manual process and we want to make it more efficient.
Costolo: Is there an on-ramp to the firehose? Yes there is. There are different levels of access. If you’re doing a sentiment analysis, you might only need access to five percent of the tweets.
Williams: We want everyone to have access to it. Pricing it is difficult. We want to license it to Google and we want two-man developer teams to have access to it. This isn’t a big money-making scheme. It’s just complicated to get it out there.
Q: Is Twitter going to host rich media?
Williams: The honest answer is we haven’t made a decision about it. We love that we don’t have to host media today, but we also think there are user experience problems around it. Photos are a mainstream activity and lots of people have provided that functionality. We think it could be a lot easier. In mobile clients, it’s a lot easier. Now we’ll have Twitter for the iPhone, which has photo sharing.
Sarver: We’ve been working on a spec that will broker media traffic between different clients. So for example, if you send a photo from a Seesmic client and your favorite service [might be YFrog or Twitpic], we’ll be able to broker that.
Q: There’s a question about competition from Facebook.
Goldman: We all worked at Blogger and back in the day there were questions about Six Apart versus Blogger and if you look at the world today, both services are still here. But if I think about my own habits, I don’t blog anymore. I tweet. So I worry more about the guy who is going to come up with the next big thing more than Google or Facebook. I worry about the guy who will come up with the 1-character tweet.
Williams: There are no lines anymore. There used to be clearly defined categories like spreadsheets and word processors. There were search engines. Now a lot of that is thrown out the window. FriendFeed came along and there were questions — is that a partner or a competitor?
Q: A lot of people are going to f8 next week. Why should they develop for Twitter over Facebook?
Williams: Create the thing that you want to exist in the world on the platform that you like working with. If it’s both, that’s great. I don’t think you can figure it out by saying they have more users than we do. If you try to make the choice by market dynamics, you’ll mislead yourself.
Q: If the uncertainty of working on Twitter’s platform becomes too great, how should developers cope?
Costolo: I don’t think things are more uncertain. They’re more certain. We’ve talked about where we’re going to go and where we think there are opportunities.
When Jesse Engle was out raising money for CoTweet, I was getting calls from investors who weren’t sure about these ecosystem companies.
Goldman: VC’s were saying that there was still opportunity for investments. They were looking for more clarity from us. Buying Tweetie before Chirp may not have been great from a headline standpoint, but from a communications transparency standpoint, it was helpful because we’re not able to the discussion here.
Williams: Let’s do a poll. How many people are more excited about building on the platform? [Half the room raises their hands.] How many people are less excited? [No one raises their hands.]
Q: What should developers build? All Twitter clients are Twitter apps, but not all Twitter apps are Twitter clients.
Sarver: If you look at CoTweet, they focused on their business audience. Analytics is another one.
Stone: I’m personally excited about some of the stuff Ryan was talking about today. We’ve historically released APIs that matched features. But when you introduce things like Annotations, you introduce the ability to creatively pivot. I’m not going to come up with the ideas right now. I’m excited about a whole new crop of applications that I’ve never even thought of that could blow our minds. We’ve embraced this unknown at Twitter. If we think we’ve figured it all out, we’re screwed.
The idea that we don’t know is an exciting idea. We’ve enabled some creative departures that we’re excited to see.
Q: Is the playing field going to be level is this new world?
Goldman: I would say the playing field is uneven. We’re a minority piece of the pie; we get 25 percent of users on Twitter.com so there’s even more of a burden on us.
Sarver: Twitter.com is going to be one take and it allows people to go around that and pivot in interesting ways.
Q: What is Twitter’s long-term international censorship strategy? We saw Twitter in Iran, when the government wanted to take it down. But because it’s distributed through so many IP addresses and applications, it’s even more resilient.
Goldman: China is a difficult situation. We have no interest in building servers there, but we’d like to offer translations so that people there can use the service.
Q: What do you feel viscerally about people getting in the way of tweets being accessible to everyone?
Stone: We have to make it as accessible as possible without putting any servers into these countries.
Williams: To be clear, censorship sucks. We want to enable freedom of expression. Even in China.
Q: Is your link shortening technology going to be available to the rest of the ecosystem?
Williams: Probably not. We’re going to launch a link shortener because we needed to fight spam in DMs.
Q: As Twitter search relevance improves, how will it handle the difficulties in d
Costolo: I don’t think anybody is married to having a Promoted Tweet be at the very tippy top.
Q: Why would I work at Twitter right now? Google pays cushy salaries and Facebook is on the verge of an IPO. This revenue model you have is still unproven.
A: Have you met us? We’re in a really sweet spot right now. We’re fewer than 200 people. We figured a lot of shit out that was painful in the beginning. We have a small number of people having a big impact. Everyone who comes in at 200 is a really early person. That’s just the objective stuff.
Costolo: Coming into working at Twitter, I’ll point out Isaac Hepworth and on Day One, we handed him the Google, Yahoo and Microsoft relationships. There’s an enormous ability to come in and take ownership. It’s you’re the guy. You’re the woman. Figure it out.
Q: A lot of these developers are now thinking that you might acquire them. How does that happen?
Costolo: The path to acquisition is wearing a black jacket and jeans.
Q: No but seriously how do these deals get done?
Goldman: Summize is an example where we almost merged. We were not that much bigger than Summize. That was at a very different stage.
Sarver: Look at Loren (who built Tweetie). He was living in Philadelphia and he just built a great product.
Q: So you have all these great apps, but how do users find them?
Williams: We do have numbers that people who use API apps are more engaged. While 75 percent of the traffic comes from the API, the majority of the users (about 56 percent) have only used our web site. A few months ago seventy percent of active users only used our web site.
The number one way people find out about apps is the ‘Via’ which says where a tweet comes from.
Goldman: We should do better for you.
Q: Is there a way that developers will be able to find each other? Will there be a place where kick-ass developers are featured?
Sarver: We’d love to do more local small versions of this so people don’t have to travel to us.
Q: How are you going to give more clarity about the roadmap? Are you going to give people guidance?
Sarver: One piece is the general roadmap announcement. We want to start videotaping quarterly updates. We can’t give details all the time because it changes constantly.
Goldman: We’re going to go through our decoupling strategy tomorrow, which is about how we’re moving away from the monolithic Twitter.com architecture. m.twitter.com, for example, will be run off the public APIs, for example.
Q: What would you like to see built?
Goldman: If we set the right strategic priorities, the team should be coming up with brilliant ideas of what can be built. And that’s if we set the right objectives, which are relevancy and having a friction-free experience.
Q: What have you learned in the last week or two from the developer community? In reference to what we’ve just said about previews of functionality, which you did do for lists and retweets. But there are things that people got in advance, like XAuth and lists. How are we supposed to interpret that?
Sarver: There are some things like sign-up APIs to everyone right away. There are lots of people trying to do bad things on our APIs. It’s a lot about trust. There are certain things that are innocuous which we can offer to everyone right away. But there are other times where we reach out to a few people.
Q: What did you learn from the last week?
Sarver: We need to communicate better externally and give better guidance.
Williams: I think that’s exactly right. We do stuff in our own company and we try and tell everyone and people get angry, so we do need to communicate better both internally and externally.
Goldman: We’ve finally reaching the point where we can execute. We sort of had our startup childhood robbed from us. Now we have enough people who can communicate in the right way to the developer community.