Thoughts from Google I/O 2017 (Day 1)

I thought I’d do some reporting on my experience here at Google I/O. This is my first one (ticket courtesy of the Google Developer Experts Program), and so I’ll collect my thoughts/learnings/experiences here.

Getting in/Registration

Fairly easy. I’m splitting an AirBnB with a fellow GDE who is attending his 8th Google I/O, so I have a bit of an expert to give me advice. We went and got our badges on Tuesday evening before the conference started on Wednesday and had no issues. We did also benefit from reserved keynote seating because of GDE status, so there was no rush to get our badges early in the morning.

Getting in: to arrive at the conference, taking a shuttle or snagging a Google Bike was the best way. I was pretty stoked that yesterday when we’d decided to walk (it was alright, but not the most pleasant walk) we saw some bikes to nab en route, and then walked in the rest of the way. Taking a shuttle (from Mountain View Caltrain, since we are staying in downtown MV) gets you dropped off right next to the door.

Keynotes

Most of Day 1 is taken up with keynotes – I appreciate getting this taken care of, to be honest. Also that being in an outdoor amphitheater in the afternoon is really sunny, so morning/early afternoon was a way around this.

The general keynote was (no surprise) all over Google products. If there’s one theme this year, it’s “OMG MACHINE LEARNING.” It was also a nice peek of general announcements, which help for picking sessions the rest of the conference (ex. “I want to learn more about this”).

I enjoyed the developer keynote (after lunch before afternoon sessions) more. There was a bit more discussion about the big news for Android (official support of Kotlin). I also thought the announcement about device search was a big deal, but hey, I’m not an Android developer.

As to web, nothing very new in the continual push for PWA (progressive web apps) and AMP (accelerated mobile pages). I did think more myself about PWAs because as an app addict, I run out of storage space on my phone constantly, and if I could replace some apps with PWAs, that would be really great for me and my phone storage 😀

One web announcement I’m excited to look more into is that Lighthouse is in devtools. Rather than having to install Lighthouse as a Chrome Extension or in the CLI, it’ll now be right in devtools, under the Audits tab. You’ll be able to see this update if you have Chrome Canary available.

 

And of course, with my focus as late on serverless, I enjoyed hearing about Firebase functions, but I didn’t really get stoked about it until a later session.

Sessions

I only made it to two full sessions on Day 1 – after the keynotes I took some time to wander around the giant festival space to figure out where things are and to relax a bit, and also spent some time in the Firebase/Google Cloud/Machine Learning Sandbox, where Sandboxes are areas with demos and staff to answer any question you might have (neat).

The first session was on Firebase functions and was really a good use of my time. I’d really like to use Firebase for a side project. The video is here.

Two big things: one being that when you define firebase functions, you define the trigger in the code. If you’ve ever set up triggers on AWS Lambda, or even using Serverless Framework (yml), it is SO COOL to be able to set up the trigger in NodeJS.

The second factor being that in Firebase, you can use Express like you normally would, and send back server responses … including being able to set cache-control headers. That means dynamically serving content using FaaS functions, AND being able to cache that dynamic response in storage so you’re paying even less (theoretically) because you’re going to invoke your func much less often.

The other session I went to was interesting, but a bit of a miss. It was on the Mobile Vision API, when I thought it was going to be ex. more Cloud Vision (my bad). It’s very interesting for mobile developers, because you can do computer vision tasks for faces, text, and barcode without connecting to the internet. I do wonder if it’d be useful for hybrid apps, but since I won’t be doing mobile in the near future, it was more just interesting. The video is here.

Sandbox

In the one sandbox I visited on Day 1, the most facinating thing had to be seeing the TPU up close. You can look at a picture of it, but one thing I didn’t expect was how big it is (TBD if I’ll take a selfie with it for scale). It has quite big vents for pulling out heat, which then is pushed out with a fan in the data center.

And there was also a whack-a-mole game tied to Kubernetes where you killed containers to try and cause downtime faster than Kubernetes could spin up new ones 🙂

Social events

I debate about including this, but it’s such a big part of I/O that it’s hard to miss. Before the conference, there were a number of pre-parties available. I ended up at Women Techmakers and the Intel event. Intel had lots of demos which made it (one could argue) a bit more technical. Both had great food. Both were free. They’re parties, what can I say.

The official conference party on Day 1 was a “block party” which, besides music and comedy also did have a fair number of demos, including VR that (it is said) will be part of a new LCD Soundsystem music video, LCD Soundsystem being the band playing the big concert on Day 2. There was also a “Museum of Developer Art,” where there were large installations, one interactive with kinect, the other a large-scale data visualization.

Thoughts

Google I/O has some long days! Although it’s my first one, it does seem that they learned a lot last year about complaints about the sun/wind/etc. and there’s lots of water available, and there was sunscreen in the swag bags! I do have a big floppy hat that I think will prove to be a good decision for protecting my sun-damage-prone-skin. Thus far, Day 2 has started out a little more chill, but will report more in the next round-up 🙂

Tags: , , ,

No comments yet.

Leave a Reply