Good Morning!
AWS Upgrades its CEO — Adam Selipsky has been deprecated like a Google product and replaced with Matt Garman. Let me tell you a few things about Mr. Garman from personal experience! He embodies Customer Obsession better than anyone I’ve ever met at Amazon. He’s historically demonstrated the uncanny ability to push back on my nonsense both firmly, but also without being a jerk about it (let me stress just how rare this is). And he absolutely has a personality of his own. In other words, this is a typical Amazonian solution: instead of fixing the first service, instead they replace it with a second version of the service that’s both better and more expensive.
I will say, though: I’m more enthused about re:Invent this year than I’ve been since before the pandemic.
From the Community
The Worst Website In The Entire World talks about trying to get VMware Fusion for free from Broadcom. You can guess how it goes; this is an entertaining read.
"Look, it’s not really a good time to talk, I–" Shut up and listen to me; another Steve Yegge post came out, drop everything and read it immediately.
My friend Ana Visneski’s book F*ck it, Watch This: Saying the Quiet Parts Out Loud has dropped. Check it out and buy it immediately. I’m mentioned!
A deep dive into AWS’s microsecond-accurate time.
I periodically get cited in the media, but it’s not every day that one of my tweets becomes the story. This one’s about Slack training its AI on user content.
Did IBM make a $6.4 billion blunder by buying HashiCorp? asks the Register. I think it’s way too soon to say, but I have faith that they’ll manage to IBM it up somehow.
Podcasts
Last Week In AWS: New Cognito Pricing Dimensions, More GenAI Boosterism
Screaming in the Cloud: Challenges of AI in Cloud Computing with Justin Brodley
Screaming in the Cloud: Google’s Graveyard and Lack of Hunger with Cody Ogden
Choice Cuts
Announcing Timestream Compute Unit (TCU) for Amazon Timestream for LiveAnalytics – I missed this gem; they’re replacing the old Timestream billing method with something that for many use cases is many times more expensive. I will remember this the next time AWS touts a price cut for a service nobody uses in a region no one knew existed as evidence of "their dedication to helping customers reduce their costs."
Amazon S3 will no longer charge for several HTTP error codes – While this is being touted as a win for customers, let’s be very clear here: this was first identified by Colin Percival as an issue, which he reported to AWS in 2006–18 years ago. My take on it? This isn’t "suddenly AWS discovered the problem," it’s that someone wrote a blog post that got enough traction that they couldn’t ignore it. Doubtless someone ran an analysis before committing to fix this that validated the money on the table was a relative pittance–had it not been, I’d bet they’d have continued to ignore the sporadic customer reports that this was patently unfair.
Amazon VPC Lattice now supports TLS Passthrough – I’d almost forgotten Lattice existed! I thought it was crappy when it first launched because somehow it neglected to mention in its rollout that it COMPLETELY REPLACED DATA TRANSFER CHARGES which made it not just a technical win, but a cloud economics win as well. Seriously–look into it. They massively buried the lede here.
Application Load Balancer launches IPv6 only support for internet clients – Ah, at last, and only about 9 years after the rest of the world ran out of IPV4 addresses. Because why hurry to support a protocol that’s merely a core part of the modern internet? Around the AWS office, they call this "innovation at our own damn pace".
Safely remove Kafka brokers from Amazon MSK provisioned clusters – This is quite the spin. "Good news, you can scale in your MSK clusters WITHOUT HAVING TO COMPLETELY DESTROY IT IN THE PROCESS." That’s how it should read. And yes, that’s as hilariously bad as it sounds.
… and that’s what happened Last Week in AWS.
