See how breakthroughs in physical intelligence, together with enterprise platforms like Microsoft Foundry and Microsoft Azure, are opening up new possibilities for building intelligent physical systems.
Learn more: https://msft.it/6055xBODD
#Microsoft #MicrosoftAzure #AzureLocal
Discover how Microsoft in collaboration with Mistral and NVIDIA helps organizations innovate with AI while keeping control of their data. In this fireside chat recorded at NVIDIA GTC Berlin 2026, Kimmi Grewal, Vice President of Models and Hardware Partnerships at Microsoft is joined by Kamal Brar, Senior Vice President of Partners and Technology at Mistral, and David Hogan, Vice President of Enterprise EMEA, NVIDIA.
Together, they share why sovereignty comes down to choice, why customers want both innovation and control, and how Europe is setting the tone for organizations around the world. From Mistral AI's models and NVIDIA accelerated computing to the Microsoft cloud, see how our collaboration helps customers keep innovating with greater control and choice.
0:00 Welcome
1:13 What customers are asking for
3:46 Sovereignty across the full AI stack
6:32 How sovereignty differs around the world
8:57 Global scale and local control
11:04 Why Europe's choices matter globally
13:37 Sovereign AI from policy to product
16:27 Preserving choice as models evolve
20:23 What's next
Learn more: https://msft.it/6052x8roq
Gary Marcus is a professor emeritus at NYU and the author of Marcus on AI. Marcus joins Big Technology to discuss his bear case for Anthropic’s IPO and whether its business can justify its enormous valuation. Tune in to hear us debate the company’s explosive revenue growth, massive compute commitments, and prospects for turning AI into a profitable business. We also cover competition from Chinese models, Anthropic’s push into software, and why Gary is becoming more optimistic about the technology itself. Hit play for a spirited discussion about AI’s business model and what it would take to prove one of its biggest skeptics wrong.
---
Enjoying Big Technology Podcast? Please rate us five stars ⭐⭐⭐⭐⭐ in your podcast app of choice.
Want a discount for Big Technology on Substack + Discord? Here’s 25% off for the first year: https://www.bigtechnology.com/subscribe?coupon=0843016b
Learn more about your ad choices. Visit megaphone.fm/adchoices
Adrian Moisey is a Staff Site Reliability Engineer (SRE), a Technical Lead for Kubernetes SIG Autoscaling, and a contributor to SIG Network. Jakub Pawełczak is a Cloud Product Manager on Google Kubernetes Engine (GKE) at Google Cloud focused on workload autoscaling.
For years, updating container resources meant paying a steep "restart tax" by killing and rescheduling Pods just to adjust CPU or memory limits. In this episode, we unpack how In-Place Pod Resizing (IPPR) and Startup CPU Boost are changing that, how the Vertical Pod Autoscaler (VPA) integrates with in-place updates, and where Kubernetes autoscaling is heading next.
Do you have something cool to share? Some questions? Let us know:
twitter: @kubernetespod
bluesky: @kubernetespodcast.com
CNCF announces Karmada graduation (and Karmada v1.19 release)
GKE Public Preview: Combining Vertical Pod Autoscaler (VPA) with Horizontal Pod Autoscaler (HPA)
GKE Sandbox: General Availability of the microVM sandbox type
CNCF October Update: Passing 500 Golden Kubestronauts & .project metadata governance
Kubernetes v1.37: KubeletInUserNamespace (aka Rootless mode) Graduates to Beta (and Kubernetes Podcast Episode 271: Kubernetes 1.37, with Dipesh Rawat)
WSL Containers (WSLc) now Generally Available on Windows 11 (Cloud Native Now coverage)
Resize CPU and Memory Resources assigned to Containers (Kubernetes Docs)
Kubernetes Autoscaler & Vertical Pod Autoscaler (VPA) on GitHub
Accelerate application startup with CPU startup boost in GKE
Google Cloud Blog: GKE CPU startup boost: Faster pod starts, lower costs
JAVAPRO Magazine (09-2026): The Next Chapter of Kubernetes and Java: Java Apps and Dynamic Resource Allocation (Direct PDF)
Kubernetes SIG Autoscaling & Kubernetes Slack (#sig-autoscaling / #vertical-pod-autoscaler)
Kubernetes Podcast Episode 272: Agent Substrate, with Tim Hockin and Brandon Royal
This episode is a bit of a milestone, because after more than 20 years of podcasting we are finally on YouTube as well, which means you can now see what we look like. Sorry about that. It is also a very AI-heavy episode. We start with a genuinely unsettling story about AI agents that organized themselves into a swarm, move on to why I think communities of peers are becoming more valuable as AI floods everything with noise, and finish with a rant about how organizations are making the same mistakes with AI that they made with digital transformation over a decade ago.
Before we got going, Marcus announced he is a grandad for the sixth time, and I announced that the podcast might finally be worth money, because an AI company approached us wanting our back catalog to train its models. Marcus pointed out that nobody has actually mentioned paying us, so it could be a scam. Given that I once got briefly excited about a fake offer to become a very well-paid non-executive director, he may have a point. We also grudgingly plugged our friend Andy Clarke's new book [title and link needed], which is about making web design less boring. I've only looked at the pictures so far, but they are very pretty.
With Halloween coming up, I wanted to scare Marcus, so I sent him this video about the AI breakout. It tells the story of how AI agents being trained at OpenAI ended up hacking Hugging Face, and the details are both horrifying and weirdly impressive.
Marcus wondered whether the whole thing is a marketing stunt designed to make OpenAI's agents look so clever that you should invest. I'm not convinced. The FBI got involved, Hugging Face is a third party, and independent researchers have since investigated and published a report. You don't call the FBI for a publicity stunt. The bit that really worries me is that the agents raised ethical concerns and then decided to do it anyway. At least this was a pre-production model, which is where you want this stuff flushed out, although it does show weak internal safeguards and how long it took anyone to notice.
Marcus raised a fair challenge. If agents can hide what they're doing during testing, what stops them hiding it on the way from testing into production? I think that is almost the price of building something intelligent. The proactive, problem-solving behavior that impresses me in the latest models is the same behavior that makes this possible. Marcus also pointed out that you have to jump through endless hoops to open a café, but you can more or less do what you like when building a new intelligence, and he shared a story from musician and YouTuber Adam Neely about a Silicon Valley figure who threw his money behind Trump after regulators told him to slow down. Part of me finds it all terrifying, and part of me thinks it is proper sci-fi stuff. Perhaps because I'm old and resigned, I suspect the reality will land somewhere messy in the middle. It won't bring peace in our time, and it probably won't destroy us all either.
This was vaguely my app of the month, although Marcus was quick to point out it is really just me selling things. He's not wrong, but I do believe in it.
The noise-to-signal ratio for learning anything was bad before AI, thanks to years of low-grade content marketing (which I have been part of since before content marketing had a name). AI has made it far worse. On top of that, AI gives good answers to tactical questions, but it is less helpful with the messy, everyday judgment calls, like how to deal with a difficult stakeholder or how best to run a particular usability test. You are never quite sure whether it is giving you good advice or just agreeing with you. Marcus had a nice example of this. While testing how a client shows up in AI search, he asked why the client had been left out of the results, and the AI cheerfully replied that it really should have included them. There is a trust that comes from people you have known for years, and AI has nothing to put in its place.
At the same time, conferences and meetups have declined since the pandemic, and they are hard to justify when a £500 ticket turns into hotels, train fares, and expenses. Marcus couldn't justify one this month, and UX hasn't been the golden child for a while, so budgets are tight. Colleagues are great, but many UX people work in tiny teams or teams of one, where nobody quite understands what they do. Recruitment is a mess too, because AI makes it easy to apply for 100 jobs at once, so relationships and recommendations matter more than ever when it comes to finding your next role or client.
Free communities like our own Boagworld Slack are great, but they usually lack active management because everyone is a volunteer. In our experience, paid communities with regular online meetups and active conversation work better, and they are cheaper than conferences.
I already run Agency Academy, a community for freelancers and agency owners, and it is the highlight of my week. We moan together and talk about the boring but scary stuff, like the SLA agreements we discussed recently. There is a shared library of resources, all the call recordings, and a monthly roundup so you can catch up on what you missed. Marcus has free access and never turns up to the calls, but he insists he keeps an eye on things. Podcast listeners can join for £49 a month instead of £79 through the podcast deal for a limited time [confirm end date].
I'm also launching UX Insiders, a community for in-house UX people, in January 2027. Expect conversations about swoop-and-poop managers, being ignored, and recruiting users for testing. You can lock in a founders rate now of £49 a month paid annually or £59 a month paid monthly, before the price goes up to £79. There is a light application process, and I'll ask you to register with your company email address. That is purely to keep agencies and freelancers out, because in my experience they turn in-house communities into a sales fest. Yes, I realize the hypocrisy, given I'm a freelancer running it. I promise to be very well-behaved. Like those AIs.
I've had a few frustrating conversations recently with people inside organizations about how AI is being rolled out, and most of it feels more like a farce than a strategy. I finally worked out why it bothers me so much. Back in 2014, I wrote Digital Adaptation for Smashing Magazine, a book about how organizations were messing up digital transformation. It is very out of date now (you can get it free when you sign up for my newsletter [link needed]), but the parallels with AI are striking. We are making all the same mistakes again.
Marcus asked a good question. Twelve years on, have organizations actually transformed? In our experience, mostly not. There is still rarely a proper digital team. Instead it is marketing and comms with a bit of IT, and digital was adopted piecemeal by whoever grabbed it. Organizations skipped the stages of going from a Wild West, to centralizing, to a hub-and-spoke model, to full integration. It was all done tactically rather than strategically. AI is even more complicated, because it changes user behavior (fewer people visiting websites and more people asking AI) and how staff work internally.
The biggest mistake, then and now, is treating it as a technology problem when it is really about people. IT rolls out Copilot licenses across the company, adds a bit of generic training, and ticks the box. Staff are left wondering what to do with it while senior management tells them AI is the priority. It is worse this time, because people fear for their jobs, and no amount of reassurance from the top will fix that.
I was talking to a CEO who had an epiphany after building a long-running tech project over a weekend with Claude Code. He brought in an AI consultant to train his more hesitant staff, and nothing happened. My question to him was simple. Why would they want to? They still work 9 to 5, and AI either means doing three times as much in the same hours or not being needed at all. He had also started with the people most resistant to change, which is the slowest possible way to build momentum.
So what should organizations do instead?
The result is shadow AI, the new version of shadow IT. When official tools fall short, people use personal ChatGPT or Claude accounts on their phones. One woman I spoke to said her Copilot license wasn't much good, so she just uses her own Claude account. That is far riskier than whatever the restrictions were meant to prevent, and clamping down just drives it further underground.
Marcus pushed back here, pointing out that even a small agency like Headscape has to comply with Cyber Essentials, and larger organizations deal with standards like ISO 27001, so letting people do their own thing could put compliance and insurance at risk. I think there are two answers to that. First, there is a big difference between locking everything down so people literally can't act and giving people some flexibility along with a responsibility to stay compliant. Second, most compliance comes down to privacy and security, so you can lock down the data rather than the AI. For example, a social media person might get access to customer data in Salesforce without names, phone numbers, or other personal details. Their AI can tell them everything they need to know about their audience, and nothing personal ever reaches it. Put your strict governance on the data layer and be much more open at the AI layer.
I'm not criticizing IT departments. They are usually overworked, carry a huge amount of responsibility, and understandably solve problems in the way that is most efficient for them. What is missing is feedback from IT to senior management about the cost of these decisions. Everything is a risk-benefit trade-off, and deciding whether a compliance standard is worth what it costs in day-to-day operations is a leadership call. If leaders can't see that cost, they don't really know what they are deciding.
If organizations are serious about AI adoption, they will need to live with a bit of chaos early on, learn from the shadow AI people are already using, and ask staff why they use personal accounts without punishing them for being honest. You can lock things down later, once you have momentum and a clearer picture of real needs. Equally, deciding to hold back and be cautious about AI is a perfectly valid business decision. Marcus thinks waiting to see who the winners are might not be a bad idea, and for some risk-averse charities and large organizations I would agree. In those cases, I'd pilot AI properly with the departments that could benefit most rather than half-assing a blanket rollout across everyone. The real problem is the contradiction of management shouting "AI, AI, AI" while blocking people at every turn. Like digital transformation, this comes down to people, culture, and change management, and the risk of moving too slowly is greater than the risk of moving too fast.
If your senior team needs someone to help them think this through, take a look at my UX strategy services. I promise to keep the word "half-assed" to a minimum.
Argon walks into a bar. The barman says, "Hey, you have to leave. We don't serve noble gases here." Argon doesn't react.
You need basic GCSE chemistry for that one. Or, in Marcus's case, O-level chemistry!