AI Experiments is live.
There’s a bunch of amazing experiments on the site; but this one below is the one I spend the most time with during its early development phase.
Honestly; I never felt more out of my depth on a project than at the beginning of this one. Sat in the kickoffs with Alex, Kyle and Yotam who were deep in the weeds talking about t-SNE, dimensionality reduction, hi-dimensional space, convolutional neural networks, and supervised vs un-supervised learning. Was a full-on nose-bleed, crash course, in ML. But so worth it. Do not fear this stuff. It’s a different world to start; but after a few weeks it starts to take. So please enjoy….
The Infinite Drum Machine
Sounds are complex and vary widely. This experiment uses machine learning to organize thousands of everyday sounds. The computer wasn’t given any descriptions or tags – only the audio. Using a technique called t-SNE, the computer placed similar sounds closer together. You can use the map to explore neighborhoods of similar sounds and even make beats using the drum sequencer.
Here’s the explainer video:
For an extra sneak peak into the development process; here’s a video showing an earlier prototype. This one has around ~40k short samples from Freesound! For the final version we licensed ~17k.
This is one of the last projects I started working on in New York, so it’s great to see it out in the real world. Mad props to Alex, Catherine, Manny, Yotam, Eric, Jonas, Kyle, Gene and bunch of other very smart people.
And yea…. what Kyle said.
AI Experiments website:
Most of the stuff I work on is confidential so I don’t get to share it publicly—but my last project could hardly be more visible. I’m not going to write lots about it as there’s already plenty of coverage out there. All I wanted to say is that it’s the meatiest (and most rewarding) thing I’ve ever tackled. We kicked the project off in January 2015 with ten people in a room sketching ideas. By the end of August we had over 200+ engineers, designers, writers, product managers, and marketeers preparing to flip the switches on over 30+ product updates. As well as the product updates and a ton of guidelines and toolkits – we also made this Google, Evolved video, a Google Doodle for the occasion, and shared the thinking on the Official Google Blog.
Everything went live on September 1st 2015.
Bonus: we also broke down the process + thinking in much more detail over on the Google Design Blog post If you’re into how things get made you should definitely take time to read it. You’ll get a better understanding of how the process worked, why the system & framework were designed to hold together, and what we wanted to reflect in the brand by making Google more accessible and useful to our users—wherever they may encounter it.
Here’s a little teaser.
Early this year, designers from all across the company, including Creative Lab and the Material Design team, convened in New York for an intense, week-long design sprint. We drafted a brief that identified four challenges we wanted to address:
- A scalable mark that could convey the feeling of the full logotype in constrained spaces.
- The incorporation of dynamic, intelligent motion that responded to users at all stages of an interaction.
- A systematic approach to branding in our products to provide consistency in people’s daily encounters with Google.
- A refinement of what makes us Googley, combining the best of the brand our users know and love with thoughtful consideration for how their needs are changing.
It was a huge team effort. Hope you like the work!
I’m excited to be able to share something I’ve been working on with team at Google’s Advanced Technology and Projects (ATAP) group over the past few months.
Project Soli is a new technology that uses radar to enable new types of touch-less interactions. My team worked on the overall project design + branding, early interaction ideas and use-cases, demo & prototype ideas, narrative storytelling and connected ATAP with the talented Jack Schulze and Timo Arnall.
Project Soli was announced at Google I/O in May 2015 to rave reviews, and the project team are now building out the DevKit. Developers and interested parties can now sign-up for updates on the official site.
Huge props to Ivan Poupyrev, Carsten Schwesig, and the entire team at ATAP. Excited to see where this will go.
Tweetfuel from Stinkdigital on Vimeo.
TweetFuel is an Arduino-controlled experiment that uses the Nike+ FuelBand to measure the health of our Twitter account.
This is something I meant to write about a long time ago. I also wanted to write a detailed and smart post about how and why the project came about, what it’s meant to say about Stinkdigital, how we had to get Nike onboard because we do a lot of work for them, maybe share some details about how and why we simplified the site design to make it feel less like a ‘product’ launch, how we used HTML5/JS, Arduino, Python, and the coverage we got in FWA, FastCo Create, FastCo Design, and lots of other magazines and blogs.
But time has passed, it’s was launched almost 3 months ago now…. so all you really need to know is that we got it from concept to working prototype in 5 days.
Whenever someone follows, RTs or mentions @stinkdigital, our custom made motorized kit spins a mounted Nike+ FuelBand.
So far the video has had 24,100 plays — and I’m still super happy with it as a piece of work.
Retweets count. (So do mentions.)
The annual social clean-up is in progress. I’ve already shut down my Posterous and Tumblr accounts to focus on my blog. Now it’s time to deal with Twitter.
I love Twitter. It covers such an incredible spectrum of information; but recently it’s begun to feel out of control. The signal-to-noise was wrong and stuff from people I cared about was being drowned out.
I started by unfollowing a few heavy/active users. But I enjoy a lot of the stuff they share so it didn’t feel massively practical. Then @sermad suggested I made more of lists. I created a few and moved a bunch of people over.
Great! But my timeline was still too noisy.
I tried unfollowing a few more people but couldn’t create a clear mental criteria in my head.
Take my friends and ex-colleagues at glue as an example. I don’t want 180+ people on my timeline, but once you start to add a few, you’re into a strange mental world of… If I add X, I should add Y, because we did X together. In the end I was trying to do something ridiculous like:
People I worked with for more than 4 years + that I regulary went out for beers with + that use Twitter regularly + but not so much that they drown everyone else out = Following.
It wasn’t working. I got frustrated.
So I went BIG and unfollowed everybody.
It might’ve been a bit impulsive but it’s democratic and I can slowly add people back over time.
I have a nagging doubt though. It’s counter intuitive. Lists guarantee I see peoples stuff; but they think I’ve unfollowed them. Not great really. But I needed to wrestle control back of the timeline so I’ve taken the plunge.
I’m sure some people will unfollow me.
Or add me to Lists…
You may have heard classic tunes recreated on barely functional pieces of electronics before, but this is a really impressive effort.
All made using:
a. HP Scanjet 3P, Adaptec SCSI card and a computer powered by Ubuntu v9.10 OS as the Vocals. (hey, the scanner is old)
b. Atari 800XL with an EiCO Oscilloscope as the Organ
c. Texas instrument Ti-99/4A with a Tektronix Oscilloscope as the Guitar
d. Hard-drive powered by a PiC16F84A microcontroller as the bass drum and cymbal
I quite like this as well.
Daft Punk Derezzed (from the Tron Soundtrack) performed on 5 Hard Drives.