I’ll be honest with you. Our @googlecloud Gemini Code Assist was useful, but with too many rough edges. My DevRel team, along with product, engineering, and even @ThomasOrTK had enough. So we’ve spent months removing friction from it.

It’s not perfect, but it’s better! A 🧵: Reminder, Gemini Code Assist is our AI assistant that extends VS Code and Jetbrains IDEs. It’s free for individual use (and doesn’t require a @googlecloud account), and available for teams via Cloud subscription. Gemini Code Assist uses Gemini 2.5.
codeassist.google

It's possible that @Google AI Studio is the best AI tool you haven't used yet. @DynamicWebPaige got me hyped about it this morning.

It's free to use, model tuning is free, experiment with multi-modal scenarios, get API keys, and more. Here are 7 killer features. A 🧵 ... First, we've had JSON mode for a few months now. Get structured output, and even do easy schema editing.

Here I asked for US states and capitals, and just had to set the "JSON mode" slider.

ai.google.dev/gemini-api/doc…

We announced Duet AI in @googlecloud at Google IO and are working hard on a product that makes the cloud experience better for you. We’ve also been kinda quiet about our progress. Let’s change that.

I got permission (at least I think I did) to show off my favorite things. A 🧵: As a reminder, Duet AI in @googlecloud makes it easier for you to use the cloud and build cloud-ready applications.

We’ve shared intro videos, but it’s time to show you the real products in Preview that will GA later this year.

It's been fun to watch @googlecloud Anthos find it's fit both for customers primarily in Google Cloud, and also those earlier in their journey.

So what's new with this Kubernetes + management plane product that makes it a good choice for anyone running containers? Quick 🧵 ... First, I like seeing a new dashboard experience! This the view is at the "fleet" level. Regardless of where your clusters are—Google Cloud, other clouds, on-prem, edge—you get an all-up view of health and policy compliance. Awesome.

Ok, the #googlecloudnext announcements are coming out, and let's explore one in particular.

Cloud Workstations are in public preview! This isn't your parent's virtual desktop. It's a fast, flexible dev environment with security in mind. Let's explore together in a 🧵 ... Browser-based IDEs are a thing now. It's how Googlers do a lot of their engineering work. It's ready for primetime.

To start with @googlecloud Cloud Workstations, you create a cluster. It manages lifecycle and networking things. I can put these around the world, close to devs.

You're not going to watch every talk at #GoogleCloudNext. I know it. You know it. Not a big deal, I won't either.

But there are a handful you might want to bookmark to learn what's new, or what the other clouds will be talking about in a year or two. A 🧵 of your best bets ... – “What’s new” talks –
Opening Keynote cloud.withgoogle.com/next/catalog?s…
Next for app dev cloud.withgoogle.com/next/catalog?s…
Next for data scientists cloud.withgoogle.com/next/catalog?s…
Architects and infra cloud.withgoogle.com/next/catalog?s…
SysAdmins/DevOps cloud.withgoogle.com/next/catalog?s…
Security pros cloud.withgoogle.com/next/catalog?s…

One of the fastest ways to deploy an app? Push to @googlecloud Run by running this command against your source code: "gcloud run deploy ."

But do that for prod deployments? Unlikely. We just added Cloud Run support to Cloud Deploy, and I wanted to try it out. A 🧵 … For reference, Cloud Deploy is our continuous deployment service that supports staged rollouts to GKE, and now Cloud Run (in preview). It relies on declarative configurations to define deployment pipelines and targets.

cloud.google.com/deploy

Let’s make all this “secure software supply chain” stuff feel a little more real, shall we?

How do I ensure that *only* container images built by a particular build engine get deployed to production? We just made this super easy to do. Demo 🧵 time! As a dev, I barely want to think about this stuff. Just make it easy to do the right thing! I’m going to show you how we made @googlecloud Cloud Build SLSA level 2 compliant (slsa.dev/spec/v0.1/requ…) by adding attestations automatically. You’ll see in a moment why that matters.

#GoogleAlloyDB. Let’s talk about why it’s a BIG deal for developers and DBAs, and why it might be the best way to do PostgreSQL in the cloud.

And how about a quick look at what the provisioning experience looks like? Quick 🧵 with a bunch of links at the end …

#GoogleIO First, it's just PostgreSQL, but operationalized in a way that @googlecloud does so well. It's 100% compatible PostgreSQL 14.

If you ONLY care about using the simplest Kubernetes in a given cloud, use the native managed option (GKE, EKS, AKS, etc).

If you're expanding outward from your anchor cloud, you care about more.

We just shipped a new multicloud Anthos. Here's a 🧵 of how it works. Buckle up. As a refresher, Anthos is a platform for container-based apps. You get GKE, config mgmt, service mesh, fleet mgmt and more, everywhere. It's GA on @googlecloud, vSphere, bare metal (bring your own OS), AWS, and Azure.

We just made a big improvement to how multicloud works.

I don't think about which data center I'm using when I upload a pic to Google Photos. Or when I perform a search. Or use Gmail.

Why should the public cloud be so different? Here's a 🧵with 10 @googlecloud services that are unique because of their global backplane … First, VPC. Most VPC products in the public cloud take a regional approach. If you want to interconnect a bunch of regional VPCs later on, it's tricky.

Not with @googlecloud. A single VPC is global with automatic communication across regions.

cloud.google.com/vpc/docs/vpc

I just spun up an Amazon EKS cluster because I like to live dangerously. I hoped to access it via AWS CloudShell, but surprisingly there's no kubectl there.

I could complain, or I could just use the dev-friendly, k8s-ready @googlecloud to make my AWS experience better. A 🧵 ... So I did the steps to get an EKS cluster, and then separately provision the worker nodes. The dashboard experience is a bit light for management but we can fix that.

From my local machine, I logged into the EKS cluster, and ran a single command to register it with GKE.

When you've got Kubernetes clusters flung all over the place—on-prem, edge, clouds—how do you access them?

It's easier when they're connected to a cloud. In Anthos 1.7, we shipped the Connect gateway which lets me access any cluster, anywhere.

A @GoogleCloudTech demo 🧵 ... Right now, I have two GKE clusters, one AKS cluster, and one kind cluster (on my desktop!) all registered with @googlecloud.

Once registered, I can view workloads, add service endpoints, etc. It's cool.

I'm still getting used to working at a company that everyone has an opinion about.

To understand more about where our @Google culture came from, I read "How Google Works" (2014) by @ericschmidt, @jjrosenberg, and @aeaglejr.

Here's a 🧵 with passages that stood out to me. On why they wrote the book …

Virtually every software system has a workflow engine. The only question is whether you build your own, or drop one in. Today, @GCPcloud shipped Cloud Workflows.

Declarative definitions, a rich syntax, and no operational effort? Let's take a look in this 🧵thread 🧵 You use Cloud Workflows to execute a series of steps, typically callouts to HTTP endpoints. A workflow may be long-running or short-running. Here, I just call out to a public API endpoint (Chuck Norris facts!), parse the result, and return it.

I’ve been thinking about the cool ways @GCPcloud reinvented public cloud. We took many first-generation cloud concepts, and made them better for you.

It's 🧵 time! Modern apps need modern data stores. We re:built the relational database for the cloud era with Cloud Spanner. The classic relational database is terrific, and you may want those capabilities AND effortless scale.

You're going to ship more software, more often, to more places—different cloud regions, or different clouds.

Can GitOps and the Kubernetes Resource Model (KRM) make it easier? I used Anthos to deploy apps AND managed services to @GCPcloud and @Azure.

A 🧵 with link at the end. GitOps offers an interesting model of declaring your infrastructure intent, storing it in a git repo, and applying it to your environment. Combine that with the KRM, and you have a way to manage a wide range of things (besides containers).

New @CloudNativeFdn survey is out (cncf.io/blog/2020/11/1…). After four years of doing this survey, they have good year over year trend data. 1300+ filled out this year's edition.

A few things stood out to me. A 🧵... Good large-company representation this year. It's not just startups. 2/3 come from companies with more than 100 employees, and almost a third came from companies bigger than 5000.

Shoveling a VM into a container? Bad. ☹️

Vacuuming an app out of a VM and leaving behind its bloated carcass? Good. 😀

I just tried out the latest @GCPcloud Migrate for Anthos, and thought I'd share the experience, and when you'd use it.

A 🧵 ... First, I needed a VM. Wanted one that had some stuff on it, so chose a Wordpress image from the GCP marketplace.

Ok, let's take a quick look at the @jetbrains "state of developer ecosystem 2020" report and what was interesting to me.

First, almost 20,000 developers took part. Many from midsize companies, so keep that in mind. People still read books! That makes me happy. Devs get most info from community sites and social media. A few of you love the radio, apparently.