Ten Blue Links "me, MAGA? Never heard of it" edition

Profile picture of Ian Betteridge
Ian Betteridge
Nov 22, 2026

1. Et tu, Jensen?

So not only was Jensen Huang getting into bed with alleged spies and turning a blind eye to what happened to Nvidia chips after they left the US, he was also a terrible manager.

So of course the usual suspects in Techbroland – coming soon to a sea-stead near you – want him back. He's everything they dream of being: a powerful, rich man unfettered by pesky things like probity, morals or the laws of the country he calls home. Steven Levy has the best rebuttal to this nonsense.

Typical of this was Paul Graham, a man who has spent so long in an echo chamber he thinks his own name is Paul Paul Graham Graham. When someone suggested to him that the tech world thinks the rules don't apply to them and that founders can break the law without consequences, he replied:

No, they understand the importance of founders and of expertise. In your world, people are interchangeable. This is less true in tech, and less true for founders. You can't replace Steve Jobs with the result of a board CEO search.

The comparisons with Jobs always come up in this kind of scenario. Only a founder can drive a company forward. This is why you end up in the toxic situation of Meta, where Mark Zuckerberg still effectively controls the company through his privileged shares. But Zuck -- and definitely Huang -- aren't Jobs.

Founders and expertise don’t count for more than the law and morality. And that Jensen is hard to replace is entirely because of his own shortcomings. Rather than doing the right thing for Nvidia and ensuring there was a clear succession plan, he made it all about him - which has now put the company in jeopardy.

Jobs, on the other hand, spent a lot of time not only building a great executive team, any one of which could have replaced him, but also spent a lot of time ensuring the culture he built was truly recognisable as “Apple culture”. Not “Steve Jobs culture”. Jensen has done the exact opposite. That’s why Apple didn’t need an executive search, not because of some mystical qualities founders have.

2. New hat who dis?

After blowing nearly $500m on supporting MAGAts in midterm races Democrats won, Elon Muck is telling anyone who will still listen to him that he's "ideologically quite moderate". Clearly, throwing money at a problem most people call "democracy" doesn't actually work as well as they thought.

Obviously this is fooling absolutely no one, and newly empowered Democrats are looking at the collection of rogues and wondering whose blood should go in the pool full of piranha first.

3. Pivots don't always work

Mem is shutting down. I'm pretty sad about this, because I thought it was a good, well thought-out product, but it was perhaps inevitable. The company pivoted from being one of the wave of Roam Research/Obsidian "knowledge graph" products to being an all-singing, all-dancing AI-based note management system, and actually delivered a much better product.

I have no insider knowledge, but I'm guessing that the cost of the AI-related features was simply burning too many tokens (and thus cash) too fast. They had recently introduced an agent-based system which looked fantastic but cost an eye-watering amount.

4. The end of the line for ZIRPers

As Ben Werdmuller notes, a lot of these kinds of startups seem to be dying around now, and much of it is down to the end of zero interest rates. A lot of the companies disappearing were recipients of VC money, and VCs can no longer tap into the kind of infinite cash pool that very low interest rates provide.

The contrast to these ZIRPers (zero interest rate policy), as Ben says, is companies that effectively bootstrapped themselves by having a good product which people are prepared to pay for. Who could have seen that would be a strategy?!?!?

5. Privacy is a team sport, and you have been picked

Meta and Google glasses, Apple Watch Audio Intelligence and Amazon's Bee all record the people around the wearer, whether they want to be recorded or not. For all of these companies, personal privacy is someone else's problem.

But "personal choice" stops being personal when it's my face or voice in your recording. Privacy is a systems property, not just a matter of the right settings existing and telling users to make sure they seek permission before recording.

Whether they know it or not, what these companies are doing is shifting the concept of privacy from the default to something you have to explicitly ask for. It's moving the Overton window of privacy in exactly the direction which, coincidentally I'm sure, most benefits companies that depend on knowing more about you than they should.

6. One agent to rule them all

Agents are one of the long-running elements of the tech saga. Back when I first started reporting in the mid-90s, it was companies like General Magic and Apple dreaming of an agentic future where small pieces of software would go out on to the network (not yet the internet) and do things for you.

At long last that dream is actually here. For example, a little bot helps me put this together. I save potential things I think should go into Ten Blue Links, and on Sunday morning an agent sends me things it thinks fit the bill, ones which I haven't saved myself. It's a really good system for catching things you might have missed.

But this usefulness also hides a vulnerability. As Tech Policy Press notes, agents can very quickly become gatekeepers. We have seen this film before, with app stores and search, and going down the same path again seems somewhat unwise.

7. Sovereignty (terms and conditions apply)

The EU spends about €264bn a year on tech, and most of it goes to US companies. This lack of control over its own infrastructure has not escaped the continent's notice, and the age of Trump has sharpened the desire to be less dependent on a country that seems to like electing aging orange autocrats.

The Cloud and AI development Act is designed to change this, but there's a problem. According to researchers, the act will leave the US giants with more than 70% of the market. It's a law that tries to both change things without changing anything.

8. Single lab, dual use

Ghana's MinoHealth jailbroke its own drug-discovery model, and it was then able to produce thousands of toxin designs, including potential bioweapons. Fortunately, this is a reputable and responsible company, performing a test. Unfortunately, as Darlington Akogo, the founder of the lab, puts it, this demonstrates how AI-derived bioweapons present an existential threat. Oh good. Another existential threat.

9. Tolkien-obsessed idiot claims the iPhone ruined everything

A man who sold a cheap VR headset, then founded a weapons company, is sad that ordinary people can use computers. And he's blaming Steve Jobs.

Palmer Luckey, who looks like the kind of man you would find in a seedy RPG-meet in the East Midlands circa 1986, argues that the iPhone made technology too easy: mastered in ten minutes, a low ceiling, and "no provision made for the man who will spend… 50,000 hours learning his art." He wants interfaces that take years of training.

The 10,000-hour rule as a product requirement is a curious hill to die on. And note Luckey's use of the phrase "the man who will spend 1,000 hours, 10,000 50,000", rather than "someone" who will. This, after all, is a man who once suggested doping women with oxytocin in order to make them stay with their husbands.

10. It ain't all bad, though

Voyager 1 is now one light-day away from Earth. That's not much in the great cosmic scheme of things, but it's still wondrous. This post by Emma Humphries is a lovely bundle of links, including the Golden Record. Read it and be inspired.

Mastodon