Practical AWS insight for engineers and the leaders who back them. Each issue: one honest take on what's actually changing in cloud and AI, plus a filtered list of launches worth your time.
Issue #01
Welcome to the first one. Here's the deal: I'm not going to relay AWS news you can already get from ten other places. Each issue you get one honest take from me, then a short, filtered list of what's actually worth your attention. Five minutes, tops. Let's go.
The Take
The real AI shift isn't happening in your IDEEveryone's still watching the robot type. That's the least interesting part. I've spent the last few months in rooms with engineering leaders, and the ones getting real value have quietly stopped obsessing over the coding assistant. An AI autocompleting inside the IDE is table stakes now, everyone has it, and it only ever touched the middle of the lifecycle. Picture your SDLC laid out left to right, the whole path from idea to running software:
Ideate→Define→Design→Build→Test→Secure→Deploy→Operate
where the leverage is now where everyone's been looking
For two solid years all the noise sat on one word, "build." But the real leverage is at the edges: the work left of build and the work right of test. That's where agents are moving now, and that's the shift most people are still missing. Here's the bit I keep coming back to. The winning move is not ripping out your stack for some shiny agent-native platform. It's pointing agents at the tools you already run. Wrap your existing test suite. Wrap your pipeline, your ticketing, your security scanners, and let an agent drive them. Claude and Kiro are genuinely interesting here (Kiro's spec-first discipline, Claude holding context across a long, messy task) but they're examples, not the point. The point is the orchestration layer around the tooling you already trust. And notice where the constraint actually moved. Generating code was never the real bottleneck. Governing it is - review, security validation, integration testing, policy. The teams pulling ahead are investing in the spine (evals, guardrails, audit trails), not just buying more agent seats. So my take: if your AI strategy still reads "we use an assistant in the IDE," you're a full lifecycle behind. The unit of work is shifting from the keystroke to the intent. Start at the edges. Wrap what you've already got. And govern it like you mean it. I've laid out the full argument, the whole lifecycle, left to right, over on the blog: Shift Left, Comrade. Where do you sit on this? Hit reply, strong disagreement is my favourite thing to open.
Worth Knowing
- a few AWS launches worth your time, and why
Claude Sonnet 5 is now on AWS
Top-tier coding and agent performance at Sonnet pricing, on Bedrock. Why it matters: if you're wiring agents into real workloads, the cost/quality line just moved in your favour, worth re-benchmarking anything you priced on the old model.
CloudFormation Express mode (up to 4x faster deploys)
Deployment confirmation in seconds, every commercial Region, no extra cost. Why it matters: the slow inner loop was always CFN's weak spot. This is aimed squarely at agents and humans who want to iterate fast, and it's free, so no reason not to try it.
EKS now supports Kubernetes version rollbacks
Reverse a cluster upgrade within seven days, no rebuild. Why it matters: version upgrades just went from a white-knuckle event to a reversible one. If fear of a bad upgrade kept you on an old version, that excuse is gone.
EC2 C9g / C9gd on Graviton5
Up to 25% better compute than Graviton4, 5x the cache, optional local NVMe. Why it matters: the easiest price/performance win in the cloud is still "move your compute-bound workload to the newest Graviton." This is the new default to reach for.
Signal to watch: Bedrock Agents → "Classic" (maintenance mode)
The original 2023 Bedrock Agents closes to new customers on July 30. Why it matters: exactly the shift from The Take, AWS moving the centre of gravity toward governed, orchestrated agents. If you built on it, start planning your migration.
On the Channel
AWS Certified Machine Learning Engineer - Associate (MLA-C01): Full Course The complete 205-minute run from zero to exam-ready on the MLA-C01. Block out an afternoon, and when you're done, go pressure-test yourself on knowza.ai. Watch on YouTube →AWS certification prep
Practice AWS exam questions from certified professionalsknowza.ai is my AWS exam-prep platform. Every practice question is written by AWS-certified professionals and built to reflect how the real exams test you. If you're working towards a certification this year, come and try a set. Try free questions on knowza.ai →That's it for the first one. Tell me if the format works for you, issue #1 is the one where your feedback shapes everything that follows.
You're getting this because you signed up on one of my channels. No hard feelings if it's not for you; the unsubscribe link is just below. |
Cloud Innovator by Day, Guitar Shredder by Night | AWS Golden Jacket
Practical AWS insight for engineers and the leaders who back them. Each issue: one honest take on what's actually changing in cloud and AI, plus a filtered list of launches worth your time.
Johnny Chivers: TECH IT FROM ME Issue #02 The Take The cheaper it gets, the more it costs AI just made building software cheap. Your bill is about to go up. There's a 160-year-old economics idea called Jevons paradox. When steam engines got more efficient, Britain didn't burn less coal, it burned far more, because efficiency made coal worth using everywhere. Cheaper per unit, higher total spend. That is exactly what's about to happen to software. AI has made generating code almost free, and...
