Jensen Huang and the Single Point of Failure

(This column is posted at www.StevenSavage.com and Steve’s Tumblr.  Find out more at my newsletter, and all my social media at my linktr.ee)

No this is not a new indie band. But now that I think of it, that would be a cool name.

So anyway, it seems like current events are my focus, and let’s get to Jensen Huang, his ever-increasing number of affairs, and his departure from NVIDIA. It actually illustrates a larger problem in society, businesses, and project teams. Am I going to relate this to Project Management? Hell yes.

So anyway as of this posting Jensen Huang, CEO of NVIDIA, had an affair with a Chinese Spy for awhile, it was less of a secret than you might think, he also turns out to have had an affair with his co-worker. No word from his wife, and hey maybe they have an arrangement, so I don’t know what she’s going through. So as you may guess, Jensen is out and this is a giant shitshow involving mysterious shipments to China, the AI bubble, and more.

I’m not going to explore it. It’s ever-changing. It’s also easily summarized as “really famous Tech CEO turns out to be unwise, compromised, and horny so he’s out.” This was a guy I once heard called one of the least offensive Techbros, and I can’t say if he keeps that title. But he’s an object lesson in many things.

I’d like to focus on why he’s an object lesson for system failure.

So Huang is rich, powerful, and apparently compromised by at least one major nation. His company is the shovel-seller for the AI goldrush. Huang is the classic example of the big CEO, and it turns out he’s a critical point of failure. We don’t know how much damage he’s done, from bad decisions to security risks to more, but we can guess.

In fact, we can guess more. Cheating with a Chinese spy? That makes NVIDIA’s recent financial dealings and valuations even more questionable. What other bad decisions are in there, or just not apparent until we re-assess his behavior.

Jensen is also not alone. Americans seem to be enchanted with the idea of Some Big Daddy who is a super genius that can do everything. I think it’s mostly evolved in the Business world based on commentary I’ve seen, and then leaked into politics. America certainly forgot the lesson of “don’t trust the guy who says he can solve everything.”

The thing is, a lot of attitudes now, especially in business and government are very culty. We want to believe some people are just super-duper mega-brilliant geniuses, often despite evidence to the contrary. This hero worship, this consolidation of power, makes those figures of worship points of failure.

First, the idea that someone is a super-genius renaissance genius capable of leading us all is a joke. No one is an expert in politics, economics, history, medicine, etc. Worshipping such people and empowering them makes them a point of failure – but often they stay at their positions, their office, etc. and keep doing damage.

I’m all for experts. But I want to see results, and a lot of people who WANT to be cult leaders are people who should be given no power whatsoever, ever.

Secondly, even if you had some genius who was a major polymath with social skills, they are still a point of failure. People age. People get insulated due to wealth and fame. People develop drug habits. People get compromised by clever Chinese spies. Even in the fantasy world of some kind of Intellectual Superman, that person is still a point of failure – it just takes once.

Third, this isn’t just politicians and CEOs and kings and religious leaders. This is everything in an organization, formal or informal. People who are the Center of All things will lead others to failure eventually by being, yes, a single point of failure. In a really bad organization you have a hierarchy of people trying to be the Big Name while all making terrible decisions.

You’ve probably been there. Our desires for the Big Leader goes up AND down the corporate ladder and it doesn’t work. We KNOW this.

This is why I’m big on Agile, on networking, on representation, and on limits to power. Organizations need to involve teams and listening and connection. People must be represented as no one knows what everyone goes through. Most of all you need businesses and communities that don’t rely on some central figure so they can adapt and adjust.

Jensen Huang may be a figure of mockery, but he’s really just the latest example of someone who’s a Single Point of Failure. And it’s not that funny when you step back.

(Well it’s still FUNNY, but you know what I mean).

Steven Savage

Immune Dysfunction

(This column is posted at www.StevenSavage.com and Steve’s Tumblr.  Find out more at my newsletter, and all my social media at my linktr.ee)

OK back at it after a long amount of not blogging. And let’s talk about the panic over AI and internet security and what gets missed.

Writing this in October of 2026, I’m hearing about all sorts of stories of AI breaching security. Oh and AI made and/or uploaded code that is vulnerable but is downloaded due to complicated dependencies. And the usual stuff about AI and security and it killing us all.

I’m kind of used to it by now. Also not exactly impressed. Congrats, your code told me someone hasn’t updated dependency libraries in three years, thanks AI. I liked you better when you were called Machine Learning and we’re less hyped. Yes, I know the internet is filled with stupid vulnerabilities, I got to see them grow up. I was there at the start.

Here’s the thing when I look at all the internet crises and security vulnerabilities and failures. I recall when the Internet was about failover. It was designed to work around failures and outages and stay connected. But boy we were into that connectivity a lot more.

So everything was about connectivity. More connectivity. Delivering more. Companies jumped onto the internet and looked at all sorts of ways to do more, deliver more – and of course take control of it. Even non-profits or “sort of benevolent non-profit” were all about getting you that information and code as fast as possible forget the rest.

We used to have a lot of computer security because so much was not connected, and there was plenty of human intervention and human paranoia about all this new stuff. We had security by aggressive and unintentional obscurity, plus Ray, the sysadmin, who kept bugging us.

In a way we kind of sort of had an IMMUNE SYSTEM in the old internet and computer age. We just didn’t notice it. We forgot about Ray when he retired.

Then we connected it all together and wondered why things kept going wrong. I think one of the great summaries for me is the Crowdstrike outage. The service and software to protect interconnected computers all over failed and took down chunks of the internet. We were all connected, had all these defenses up, and were so dependent on ONE SERVICE that it created a failure. Think that over for awhile – we defended our connections with another dependency that blew the hell up.

Right now the issues of AI and about every damn other problem of technology is a mix of Late Stage Capitalism and not having an immune system in our technology. We jack everything together to sell Bluetooth Aquariums, then watch it all fall apart because it’s all connected with no defenses (or if there are the defenses go down as one company does 90% of it).

Complain about AI and everything else. What we have though is a lack of a technology immune system in the world. Actually the more annoying parts of AI and the internet in general exist due to this lack of the immune system.

Maybe it’s time we look back a few decades and ask how things ran, why, and what we missed because it was just there.

Steven Savage

The Speed/Efficiency Paradox

(This column is posted at www.StevenSavage.com, Steve’s Tumblr, and Pillowfort.  Find out more at my newsletter, and all my social media at my linktr.ee)

A little more “Steve rants about software and process” time here as I contemplate issues of efficiency, distraction, and how faster isn’t always better.

It’s easy to argue that faster is good. The more things we do faster the more we can do! The more we get done. However, let us take a warning from real life that a love of speed breaks down, say, on the highway with some of your less inhibited fellow drivers. Speed burns fuel, leads to accidents, leads to citations, disrupts others, and pisses people off.

Speed, like anything has side effects. And am I going to compare idiot drivers to idiot software that takes “Gotta Go Fast” as a motto? Yes I am.

Previously I had critiqued an emphasis on speed as efficient software also adds more work, so somehow we get slower by going faster. But let’s take a look at how speed in our software, our culture, our business process, backfires.

First and foremost I’d note that speed can be wasteful. You have the latest software, the latest graphic card, the latest AI. Great, how much did that cost? How much does it cost to power? How much effort do people have to put in to keep up? Having the latest cool technical stuff or business process might be fast, but is it worth the cost? Assuming speed always pays off is not a great idea.

Secondly, when it comes to that fast software or new workflow, going faster also may mean it’s harder to check on what you’re doing. Sure, you streamlined that billing process, now how about those lawsuits that are coming in because something got missed. Sometimes slow is there to make sure you don’t screw up – and if you’ve been in IT any length of time you know.

Third, so you got that fast new process for disposing of waste or sending out paychecks – until it screws up because it’s so fast. Not only have you done all sorts of financially and ethically dumb things, now you’re getting investigated, regulated, etc. Admittedly, I wish we saw more of this in the world, but still there’s some impact on companies, business, and people. For now.

(Working in medicine and medical tech, which is rightfully regulated and contains people very serious about doing things right, I want more regulation and policy like that.)

Fourth, speed can be disruptive, and not in the fun Silicon Valley disruptor way. Actually, forget that “disruption” thing period because we’ve had a lot disrupted and it turned out to be awful. Going faster does mean you break things, distract people, derail processes, and so on and that has an impact. The love of disruption misses that maybe some things shouldn’t be disrupted and we miss that as speed, as disruption is a virtue above the results.

Fifth and finally, past a point, speed can just piss people off just like that driver racing by you at 95 MPH in a school zone. People need time to process information, input, signals, forms, and what have you. Someone racing around doing things fast even if good can still derail human interaction. Sometimes the point is to talk to people, take time, and be considerate of what happens to them.

(I’ve used this as a Project Manager, where just a checklist and a half hour meeting gets far more done than any kind of software process.)

So no, I’m not a fan of speed for speed’s sake. It has it’s time and it’s place. I’d love to figure out Faster Than Light travel, for instance. But let’s not assume it’s the best thing ever. It can often create more problems than it solves if it solves any problem.

Steven Savage