The latest episode of @TheCTOAdvisor Podcast with @CalvinHP. Is @OpenAI's Codex runtime behaving like malware? We ask this question and discuss AI run-time security. Audio - thectoadvisor.com/podcasts/e… Video - youtu.be/_KEQyQp8sdM?si=TrEM…

Episode 700 of @thectoadvisor podcast dropped this morning. It's a reboot of the podcast. We are going back to the weekly format but with a co-host. @calvinhp, CTO of @sixfeetup, a Python development shop, is joining me weekly. It's a solid combination of an infrastructure CTO and a developer CTO. This week we discuss JEV and Qwen 3.8. thectoadvisor.com/podcasts/e…

Why did I launch knerd.ai? I'm very happy with @thectoadvisor. It's a nice lifestyle business. It could and will likely fund my retirement. I got or at least am getting mine. But the Bible teaches us that God have given each of us talents and expects a return. Simply taking care of my family is the bare minimum; returning to God what he gave me without a return. Moonshots like Common Knerd are my opportunity to create opportunity for others. Yes, I'm OK with taking risks. But that doesn't give me higher ground to go out and tell others to do the same. What I am undertaking is hard. And I'd be lying if I said I didn't also want to reap the benefits of that labor. But in the course, I have exceptional capabilities, and I need to use them for others, in addition to my family.

The latest 4 + 1 AI Infrastructure Layer 2C assessment is ready. @Snowflake layer2c.com/assessment/snowf…

I wanted one role in the platform: this analyst can run the assessment and cannot touch the evidence it reads. That role doesn't exist. Owner, editor, viewer, and that's the whole list. A viewer gets correctly refused four out of four ways in, including the route that skips the document pipeline and writes straight to the vector store. But a viewer can't start a conversation either, so a viewer can't run the instrument. The person operating the assessment is the same person who can rewrite what it reads. That isn't a product defect. It's ordinary access control, and no permission fixes it, because the capability you'd withhold is the capability you have to grant. So how do you control it? You don't prevent it, you catch it. Enumerate the corpus, diff it against the set you approved, and run that check before anything gets scored instead of as a report afterward. Same code, different place in the pipeline, and only one of those is a control. Inherit the boundary. Own the gate. Disclosure: @KamiwazaAI sponsored this lab. They didn't choose the vendors, see the instrument, or hold a veto.

This is my 10th anniversary with The CTO Advisor. It's also been 6 years since we had a major website refactoring. I'm all in on making my content machine readable for AI crawlers. thectoadvisor.com

SUSE Rancher makes our third Kubernetes-based Fourth Cloud Assessment to our research. cloud.layer2c.com/assessment…

We are now up to 25-vendors in our 4 + 1 AI Infrastructure Assessment. Welcome to the board, @AnthropicAI layer2c.com/assessment/anthr…

I published the 4+1 AI Infrastructure Assessment for @Supermicro. layer2c.com/assessment/super…

Ran into @StevenDickens3 right after the #HPEDiscover keynote, and it turned into one of those hallway conversations that reminded me why this event matters. People joke about @HPE being a legacy infrastructure company. But this is where enterprise architects hang out. The HPE story is broader than most give it credit for: networking, compute, storage, private cloud, AI ops, control planes, GreenLake, and a partner ecosystem that can be consumed through one commercial relationship. That matters because enterprise transformation is rarely blocked by the demo. It is blocked by procurement friction, transition cost, integration risk, and paying for the old platform while standing up the new one. That is why the VMware migration angle is interesting. Morpheus gives HPE a stronger control plane story. GreenLake gives them the commercial motion. The 90/9 program attacks the transition-cost problem directly. There is still a technology gap between VMware and what HPE can fully replace today. But HPE may be the largest vendor giving VMware customers a credible alternative with enough commercial leverage to make the conversation real. And Juniper changes the AI infrastructure story. AI is not just GPUs. Networking is now one of the battlegrounds. HPE is punching above its weight.

"Your agent just deleted prod. Now what?" AI agents are operating at scale — and when they go wrong, the blast radius isn't one file. It's entire buckets. Woon Ho Jung, CTO of @ClumioInc, breaks down why replication won't save you and what air-gapped backup across AWS and GCP actually looks like in practice. One policy. One portal. Two clouds. And a recovery point that no agent can touch. thectoadvisor.com/clumio-dat… Full conversation on the CTO Advisor podcast.

Ran into @bocanuts at the show. I asked him if he had any AMD Strix Halo under that hat.

Roadtrip work life.

We are at @Articul8_AI as part of our #CTOARoadtrip talking the impact of Claude Code and Openclaw on enterprise AI adoption and speed.

We are making road trip stops to Equinix and Articul8 this week. I’m looking forward to learning more about their AI solutions heading into GTC.

The flying cloud has seen some things. We checked off three more states on this #CTOARoadtrip -Kentucky -Kansas -Utah