@narendra I remember when you first pitched me on this idea! Glad to see it launch! What've you learned and innovated on from that first inception until now? What do people find interesting and compelling — or better, confusing — about what you're trying to do? When do they get the "ah ha!" moment?

@chrismessina I think the most important part of this process has been relentless cutting of features, ideas, and use cases to get to something clean and wide open with potential (i.e. like early twitter). I wanted the initial app to have some habitual use out of the gate and it is great for that if you enable your calendar and health (like a simple Fitbit). We spent a lot of time trying to make Peck something handy for a morning ritual and possibly an afternoon or travel checkin. We dialed back the haptic interface for peck intensity to just low and high. That was important. Embedded in the app are the seeds of a lot of things that have huge potential at scale: pecking back and forth with friends, endorsing peckables as a means of social discovery and filtering places/restaurants you might want to visit, and the foundation for the platform. We just finished a simple in-house design tool that allows for rapid creation and deployment of peckable micro-apps so yesterday we were able to roll out all NFL and MLB teams in less than an hour. Lots of stuff coming. I want to peck Uber and know how many vehicles are nearby and approximate wait time and then take the next step to request a car!

@narendra totally! Can you build it in an hour? developers.uber.com

@chrismessina I just glanced at the API and I think we can accept that challenge! I have to run out for a bit, but I'll start as soon as I get back. I'll post a screenshot.

@chrismessina done. search uber! screenshot here -- https://twitter.com/narendra/sta...

@rrhoover Thank you for trying it out. Our origin story is a bit longer but basically people open and close apps to get little pieces of information and that is wildly inefficient. You don't need an app for everything. The mobile web is pretty clunky. And conversational UI makes you type more not less. So Peck is a simple haptic interface to let you tap on people, places, and things. We started with the Green Bird as a way to get some habitual info and decided a "disposable" stream made the most sense. So when you tap on things responses flow into your stream. Some responses may have embedded actions if you tap on them (like Sharing, Viewing, etc.). And most things that you Peck will also come back with some more peckables so there is semantic/contextual discovery happening. At scale with our open platform, things get very exciting. 🐦

@rrhoover just peck away Ryan and the many secrets are revealed

@narendra I'm one of those incessant app switchers. I'll peck more!

@rrhoover @om I'm having the same experience :P I guess I don't know what to expect so every peck is a little bit confusing. I assume I need to connect more things (apps)?

@davidsfeng it becomes valuable when you create your own set of sources and a tight little network. Hopefully you will follow some of your friends and share your thoughts there after!

@lbrkr Thanks for pointing that out. Yes, we did decide to focus on the US initially, just from a practical point of view. Peck processes a fair amount of incoming data like weather forecasts and distances. Our current implementation does allow you to choose the output units, so we suspected non-American users would get frustrated at seeing Fahrenheit and miles. Also, all the initial content we loaded is in English. We have every intention to support the rest of the world in the not too distant future!

@tsunaze Yes, we have a setting for disclosing your city. It is NOT you exact location. Also, you know exactly who pecks you!

@dinakar_tumu A lot of AWS stuff including Lambda and also Neo4j.

@dinakar_tumu As @narendra mentioned, each Peckable is a bit of automation code written in Javascript (Typescript to be exact). That Peckable runs in response to a user's peck and the streams the data back to the app. Recommending Peckables and content to users is done using a graph database (Neo4j).

Thanks for the feedback, @knownhuman. We use Twitter in several places in the app. As you saw, we use it in the onboarding to grab your name and avatar. We also use Twitter as a source of data for a lot of Peckables. Since you yourself can be pecked, when you added your Twitter account as a source, that gets added as information others see about you. This does not use Twitter as an authentication method, just as a public data source. We will likely do a deeper Twitter integration down the road that will require you to authenticate again. As for the user feedback, yes, we've wrestled with this one. We've tried adding cards to the top, the bottom - but not the middle. We'll give that a try, too. As a side note, you can swipe left to remove unwanted cards or peck the Green Bird to reset everything.

@tyreus Interesting. From a UI standpoint, I didn't consider that a middle add. To me Card 1 is effectively a fixed of navigational card. For twitter integration - what would happen if I put in the name of someone famous on Twitter? Would I get pecks directed at them? Since I added later in the flow, and there was no oauth, am I missing out on pecks actually directed at me?

@knownhuman The Twitter source you set for yourself only dictates part of the content that others see when they peck you in the app. So if you set your Twitter source to be @mcuban, people will see content from Mark Cuban when then peck you. But the notifications go to you because your actual Peck account is linked to your phone number.

@ghobs91 Sorry if it isn't clear. Out of the gate, it is exactly what you said, a speedy way to get contextual data. As the platform builds out it will have very broad based utility and contextual smarts.