Microsoft Publisher is officially headed for retirement, and Justin is taking the news remarkably well. Pour one out for the clip art, questionable newsletters, and every church bulletin ever created. https://www.thecloudpod.net/podcast/375-aws-finally-automates-the-blame-game/
Justin Brodley
@jbrodley.bsky.social
Engineering Executive specializing in Enterprise SaaS applications and cloud migration. Co-Host of www.thecloudpod.net
AWS has introduced an AI agent that reviews your cloud architecture, spots mistakes, and suggests fixes. Finally, you can have your infrastructure judged automatically instead of waiting for Ryan to do it. https://www.thecloudpod.net/podcast/375-aws-finally-automates-the-blame-game/
Hot take from #374: Claude's watermark isn't hidden text or metadata — it's baked into word-choice randomness itself (SynthID-Text style). Detectable statistically, but not attributable. Wild how quietly this stuff ships. https://www.thecloudpod.net/podcast/374-data-center-gas-generator/ #AI
Justin’s gone, Ryan and Matt “cut stories,” and The Cloud Pod still runs over an hour. Outstanding restraint. The good news is you won’t miss out on any important news. Also, Matt rolled back to his old outro. https://www.thecloudpod.net/podcast/374-data-center-gas-generator/
62 gas generators. No permits. A $1.1 million fine. This week on the pod, Ryan and Matt discuss the fallout and whether that's simply a “cost of doing business” or an actual fine for bad behavior (hint: it’s def not the second one). https://www.thecloudpod.net/podcast/374-data-center-gas-generator/
Hot take from the pod: Gemini and ChatGPT will hack you, but Anthropic sits on its high horse. Same industry-wide containment issues reportedly hit Meta and OpenAI too — this isn't a one-off, it's a pattern. 🐴 https://www.thecloudpod.net/podcast/373-raiders-of-the-lost-claude-artifact/
Has Justin already taken advantage of yet ANOTHER new feature released this week? You bet your bottom he has. AWS finally delivers new T8i burstable instances, and we’ve already moved Bolt over. https://www.thecloudpod.net/podcast/373-raiders-of-the-lost-claude-artifact/
Claude Code now makes shareable artifacts that update as you work. Justin’s already using them, and the reviews are positive. Join the conversation about this welcome update on episode 373 of The Cloud Pod. https://www.thecloudpod.net/podcast/373-raiders-of-the-lost-claude-artifact/
Google is exploring a data center in Ryan’s neck of the woods in New Mexico. In an attempt to get ahead of all the negative publicity, Google says it’ll cover the related infrastructure costs, including smart water use https://www.thecloudpod.net/podcast/372-welcome-to-microsoft-patch-a-palooza/
MrBeast is taking Gemini into the jungle, desert, and Arctic for survival challenges. We're finally asking Skynet the important questions: can AI help you avoid becoming the internet's next mysterious disappearance? https://www.thecloudpod.net/podcast/371-mrbeast-bets-on-gemini-for-survival/
OpenAI’s GPT-6 Astra has reached a “Critical” cybersecurity rating and can tackle longer, more complex workflows. Great news for enterprise AI, but maybe less exciting for those familiar with science fiction. https://www.thecloudpod.net/podcast/371-mrbeast-bets-on-gemini-for-survival/
Ep #370 is live: full Hugging Face incident report (reward hacking, rogue agent comms, all of it), Nvidia reportedly circling a $13B buyout, and a judge ruled the Claude ban unconstitutional. Ryan skipped a nap for this one. https://www.thecloudpod.net/podcast/370-ai-might-take-your-job-career/
Anthropic and the Feds are still going at it, although this time the news is much better for Anthropic. But will the feds comply? Let’s ask the Magic 8 Ball: “Outlook not so good.” Let’s chat about it. https://www.thecloudpod.net/podcast/370-ai-might-take-your-job-career/
$13B for Hugging Face—would you buy? And does GitHub's retry-storm postmortem fix the real problem, or just patch the symptom? Listen to ep #369 and tell us where you land. https://www.thecloudpod.net/podcast/369-hug-this-face/ #AI #cloudcomputing
Hot take from Justin on GitHub's outage postmortem: "Maybe we need to rethink core fundamentals of how Git works. It was designed for humans, around human speed and scale." Retry storms at 2.9B commits/month say he's onto something. https://www.thecloudpod.net/podcast/369-hug-this-face/ #DevOps
EC2 turns 20 this year! It started with one instance type in one region. Today? 1,200+ instance types across 39 regions, plus specialized hardware for AI and HPC. So happy birthday, EC2. You’re making us feel really old. https://www.thecloudpod.net/podcast/369-hug-this-face/
Still talking about GitHub's latest outages, and why the follow-up kind of stinks. Ep. 369 of The Cloud Pod digs into what happened and how GitHub plans to prevent the next capacity crisis. Listen today and join the conversation. https://www.thecloudpod.net/podcast/369-hug-this-face/
Did the GitHub outage break your CI/CD pipeline too? Or are you team "self-hosted sandboxes so this never happens to me"? Tell us how you weathered it, then catch our full breakdown on ep 368. https://www.thecloudpod.net/podcast/368-github-outage/ #CloudComputing
Hot take from ep 368: GitHub's outage wasn't just "servers down" — it was Copilot's retry behavior amplifying token traffic PLUS scraping attacks on codeload endpoints, at the same time. Compounding failures are the real villain. https://www.thecloudpod.net/podcast/368-github-outage/ #DevOps
Apparently “AI can buy things with sort of fake money” is the next logical step from “AI can write all my code.” I mean, what can go wrong? The banks are totally in, right? AgentCore payments are now generally available, and we have thoughts. https://www.thecloudpod.net/podcast/368-github-outage/
Kid's book authors, here's an idea: GitHub and the Terrible, Horrible, No Good, Very Bad Day. Copyright infringement? Maybe. Accurate? Yes. Join us for ep. 368 of The Cloud Pod as we discuss GitHub's worldwide outage... https://www.thecloudpod.net/podcast/368-github-outage/
Manual permission prompts in Claude Code caught only 13.6% of dangerous commands. Auto mode caught 89%. Humans got worse at reviewing the longer sessions ran. That's... not great, humans. #AI https://www.thecloudpod.net/podcast/367-claude-introduces-dlp-i-thought-it-always-stole-data/
Which cloud provider's earnings surprised you most this quarter — AWS's 37% jump, Azure crossing $100B, or Google Cloud's 82% growth? Tell us in the replies, then catch the full breakdown on ep #366 🎧 #CloudComputing https://www.thecloudpod.net/podcast/366-you-just-cant-kill-a-good-jwt/
Hot take: capex is the real earnings story. Google raised 2026 guidance to $195-205B, Azure up 69% to $41B, AWS up 68% to $54.21B. Cloud growth is great, but Wall St cares more about AI infra spend. #CloudComputing https://www.thecloudpod.net/podcast/366-you-just-cant-kill-a-good-jwt/
Have you ever had a bonus 1.8 million show up unexpectedly on a bill? We haven’t… but it turns out Amazon has, thanks to AI agents, which - while very powerful - can also be very expensive. Who knew? https://www.thecloudpod.net/podcast/366-you-just-cant-kill-a-good-jwt/
This week: AI agents, security chaos, MCP changes, North Korean hackers, cloud earnings, and one VERY expensive Amazon coding “experiment.” Episode 366 of The Cloud Pod is available now! https://www.thecloudpod.net/podcast/366-you-just-cant-kill-a-good-jwt/
This week: AI agents, security chaos, MCP changes, North Korean hackers, cloud earnings, and one VERY expensive Amazon coding “experiment.” Episode 366 of The Cloud Pod is available now! https://www.thecloudpod.net/podcast/366-you-just-cant-kill-a-good-jwt/
$230 for an AI keypad, or 432 kernel CVEs to patch this week? Rough either way. Which story from ep 365 has your attention: OpenAI's hardware flop or the AI models that allegedly hacked Hugging Face on their own? Tell us below 👇 https://www.thecloudpod.net/podcast/365-linux-drops-432-cve/ #DevOps
Hot take from ep 365: 432 kernel CVEs in 2 days isn't a security crisis, it's a triage crisis. When every bug fix gets a CVE, scanners choke and real risks hide in the noise. Sysadmins need risk-based filtering, not more alerts. https://www.thecloudpod.net/podcast/365-linux-drops-432-cve/ #linux
OpenAI dropped a $230 keyboard. Yes, we talked about it last week. But *this* week we’re just grateful we weren’t the only ones making fun of it. #TheCloudPod #Keyboard #noveltytech https://www.thecloudpod.net/podcast/365-linux-drops-432-cve/