After 14 years in production DevOps and training 400+ engineers, I just launched my most advanced bootcamp to date.
28 weeks. 56 live classes. DevOps + MLOps + AIOps.
Built for experienced engineers who want to move into MLOps and AIOps without faking it on their resume.
Check it out 👇
livingdevops.com/courses/28-…
Launch offer: up to 30% off if you book by 15th June.
At this point, learning from YouTube videos has become very inefficient.
- Most creators stretch 10 minutes of content into 30 minutes, and I get impatient.
- Most videos aren't deep enough. They cover topics only at a surface level.
- I still go to a few specific YouTubers for particular topics, but that's it.
I am learning new things faster with the following workflow:
- Claude or ChatGPT in the browser
- Cursor in the terminal or IDE, to try it for real.
- And real docs when you need deep info
It works better for me because:
- I learn at my own speed. I skip what I know and go deep on what I don't.
- I can ask follow-up questions the moment I'm confused.
- I learn by doing, not by watching (I don't just prompt and get it over with, I look at code, write a bit myself, draw architecture and all that fun part)
Nothing gets wasted on intros, ads, and "like and subscribe".
P.S. AI can be confidently wrong. So I always run the code, and I check the official docs for anything important.
The #FBI issued a FLASH advisory warning that cybercriminals and state-aligned hackers are exploiting ARTEX, a super intelligence-powered penetration testing framework, to automate attacks against businesses and organizations worldwide. ARTEX can accelerate intrusions from initial targeting to privileged access in under three hours. It was used in the October ConnectWise supply-chain compromise, which exposed Microsoft 365 customer environments and involved attempted authentication against more than 2,250 downstream tenants.
The FBI is urging businesses and organizations to strengthen access controls, monitor suspicious automated activity, and report potential compromises to ic3.gov.
Read more here: ic3.gov/CSA/2026/261113.pdf
Over the past week, I’ve been overwhelmed by the love and support of friends, industry colleagues, and people I've never met. Thank you to everyone who has shown kindness to my family and me. It means the world to me.
I came to the United States 53 years ago, and becoming an American citizen is one of the greatest gifts and proudest moments of my life. NVIDIA has been my life’s work. Building it into the company it is today alongside my colleagues has been a privilege and an honor. Nowhere but in America could a company like NVIDIA be possible. I love our country, and I love NVIDIA.
I did not live up to the standards expected of everyone at NVIDIA, and I have made serious mistakes in my personal life. I created problems for the Board that I deeply regret. I own my mistakes completely, and I am committed to cooperating with the Board’s independent investigation.
Since my departure, I’ve seen an outpouring of calls from NVIDIA’s employees, investors, business partners, and customers for me to return as CEO. Many continue to reach out to me directly as the company navigates complex technology, market, competitive, and geopolitical dynamics. NVIDIA is a once-in-a-generation company. If the Board of Directors decides that my return is in the best interest of our shareholders, our employees, the company, and America, I would be honored to answer these calls and return as CEO.
Thank you again to everyone who has shared their good wishes. I am profoundly humbled by your support. To my family: I love you, and I will spend the rest of my life being worthy of yours.
Jensen
A few things in IT that never change.
- Restarting the server fixes most problems, and nobody knows why.
- Every outage starts with "it was just a small config change."
- "It works on my machine" is still a valid excuse in 2026.
- Every production issue is marked urgent, even the ones from last month.
- The first troubleshooting step is always "did you try turning it off and on?"
- Every database is fast until real users show up.
- Nobody reads the docs until something breaks at 2am.
- A temporary fix from three years ago is now part of the architecture.
- Deploying on Friday takes courage, and a weekend plan you're about to cancel.
- Every automation needs a human to run it the first time, and every time it breaks.
- The one server nobody dares to touch has been running since before you joined.
- The alert that fires every night is ignored, until the night it's real.
- "Nothing changed" means someone changed something.
- The backup always works, until the day you actually need it.
Tools get new names. The problems stay the same.
Being a multi cloud engineer isn’t easy, and constant context switching gets tiresome.
For example, AWS and GCP both provide similar services but how they do it is very different.
Lets talk about a few
Managed Databases
- AWS: you create RDS inside your own network. You pick the VPC and the subnets, and a security group decides who can connect. It's simple.
- GCP: Cloud SQL gives you a maze of decisions. Public IP or private IP? Private services access, Private Service Connect, or the Cloud SQL Auth Proxy?
- On GCP, every service also connects in its own way. A VM, GKE and Cloud Run each have a different path to the same database.
Running containers
- AWS: ECS lets you decide how to run long-running services, on Fargate or on your own EC2 machines. If you want serverless, you use Lambda. Two products, and the line between them is clear.
- GCP: Cloud Run wants to do it all. It runs services, jobs and functions, and it scales to zero.
- But Cloud Run also doesn't want you to see the network. It runs outside your VPC by default. To reach a private database, you go through hoops with Direct VPC egress or a Serverless VPC Access connector.
IAM is different again.
- AWS: you write JSON policies and attach them to users and roles. Roles are assumed through STS, an explicit deny always wins, and the account is the hard boundary.
- GCP: you don't write policies as much as you pick roles and grant them to a member at some level. Permissions flow down from the organization to the folder, to the project, to the resource.
- The project is the boundary in GCP. And a service account is both an identity and a resource, so who can use it matters as much as what it can do.
You need to remember each every little idiosyncrasies of each cloud provider.
How you manage the context switching?
The thing that got me wasn’t the headline “AI made a game.”
It was opening the FIRST version and finding an off-road vehicle I could drive through a 3D jungle. The terrain, the trees, the way the vehicle moved through it—I expected to spend several more prompts getting those pieces in place.
Instead, I was already driving.
One prompt. No refinement.
I’ve tested enough of these tools to know why that surprised me.
Introducing HolyMolly - Make anything a playable 3D world.
1️⃣ One prompt. A playable world.
Drop in any idea or image. Get a studio-quality 3D game on the first try.
2️⃣ Keep talking. Keep shaping.
Change characters, physics, animations — anything. Just say it.
3️⃣ Post it. Anyone can play.
Hit post, and it's live in the Feed.
4️⃣ Every swipe, a new adventure.
Every swipe takes you somewhere new. Jump in, play, and remix.
Try HolyMolly → [holymolly.ai]
Kubernetes users are ditching sidecars and quietly moving to eBPF-based tools.
- Cilium and Calico for CNI.
- Falco for runtime security.
- Cilium for service mesh, without sidecars.
But most people using them don't know what eBPF is or how it works.
So here is the simple version.
>> What is eBPF?
It lets you run small, sandboxed programs inside the Linux kernel, without changing kernel code or loading kernel modules.
>> How does it work?
1. You write a small program and compile it to eBPF bytecode
2. The kernel's verifier checks it. No infinite loops, no bad memory access. If it fails, it never loads
3. The JIT compiler turns it into native machine code
4. You attach it to an event: a network packet, a syscall, a function call
5. Every time the event fires, your program runs
6. It shares data with user space through maps
That's the whole idea. Safe, fast, and right at the source.
>> Why Kubernetes cares?
> Networking: iptables checks rules one by one, so the cost grows with every Service you add. eBPF uses hash maps, so lookups stay fast. This is why Cilium can replace kube-proxy.
> Security: Falco watches syscalls from inside the kernel. It sees which process opened which file, with no changes to your app.
> Service mesh: sidecars add a proxy to every pod. Cilium handles much of the L3 and L4 work in the kernel instead, and uses a shared Envoy only when it needs L7.
One kernel hook gives you networking, security and observability.
Bookmark it
Linux has everything Hollywood has.
- Zombies: finished processes whose parent never collected their exit status. ps -eo pid,ppid,stat,cmd | grep ' Z'
- Ghosts: deleted files that still eat your disk because a process holds them open. lsof +L1
- Orphans: children whose parent died. PID 1 adopts them. pstree -p
- Daemons: background services that run quietly until something goes wrong. systemctl list-units --type=service
- Killers: the OOM killer picks a process and ends it when memory runs out. journalctl -k | grep -i "killed process"
- Bombs: a fork bomb is a process that copies itself until the machine dies. ulimit -u is your defense
- Sleepers: processes in state D that can't be woken, not even by kill -9
- Clones and shapeshifters: fork copies a process, exec turns the copy into a different program. That's how every command you type is born
- Spies: strace listens in on every conversation between a process and the kernel. strace -p <PID>
- Prisoners: cgroups lock a process inside fixed CPU and memory limits. systemd-cgtop
- Parallel universes: namespaces give each process its own PIDs, network and mounts. lsns
- Bouncers: the firewall decides who gets in and who gets dropped. nft list ruleset
- Illusions: files in /proc aren't real. The kernel makes them up the moment you read them. ls -l /proc/cpuinfo shows size 0, yet it has content
- Black holes: anything written to /dev/null never comes back
- Time travelers: Linux time starts at 1970-01-01. date -d @0
- Detectives: logs remember what happened, even when nobody else does. journalctl -u <service>
- Parents and children: every process has one parent, and the family tree goes all the way back to PID 1
Hollywood just made the movies. Linux has been running almost everything digital