1. Vivaldi sets out its stall on AI
Vivaldi has taken a bold stance against the growing trend of integrating agentic AI into web browsers. CEO Jon von Tetzchner argues that embedding artificial assistants turns web exploration into passive consumption, threatening the diversity and vibrancy of the web. Vivaldi pledges to prioritize human curiosity, autonomy, and privacy over hype, focusing on personalization and resisting Big Tech’s push toward AI-driven browsers.
Unless AI enhances user autonomy without privacy risks or diminishing the open web, Vivaldi says it won’t adopt it, vowing instead to keep building a browser for those who value exploration and independence.
I get what von Tetzchner is saying, but I’m a little more positive about the idea of agentic browsers. Using something like Comet or Dia has actually increased my use of the web over apps, because having AI integrated into the browser is just useful.
Case in point: earlier today I checked out the list of upcoming rugby matches on TV and asked Comet to look through the page and put all the games into my calendar. It did it easily and quickly, which was a decent little time saving. That kind of automation seems to be useful. Summarising web pages, less so.
2. They put sal-mon in the fish tacos, Hank
If AI in the browser can be occasionally useful, how about AI in your fast food ordering system?
Taco Bell is re-evaluating its use of voice AI at drive-throughs after rolling out the technology in over 500 locations. While AI has processed millions of orders, customer reactions have been mixed – some dislike glitches or the impersonal experience, while others try to exploit or troll the system. Of course they do, and somewhere inside that system is FREE TACOS.
Part of me admires them for trying it, and another part just wonders what they hell they were thinking.
3. China’s vision of AI isn’t the same as the west’s
If most of the US-based AI companies are frothing at the mouth in an effort to out-talk each other on how close they are to general intelligence and true AI, China appears to be taking a somewhat different approach.
The key word is pragmatism: rather than chasing something that may be impossible, Xi Jinping is apparently pushing applications as being the most important thing. It’s going to be a fascinating few years as the different approaches play out.
4. The Trumpian tantrum on the DSA
If I didn’t know that he hadn’t read a book in his life, I would take Donald Trump for a student of history. Or at least that part of history which focuses on how empires can use economic power to get their own way politically.
The latest target is the Digital Services Act, Europe’s flagship piece of policy which regulates online platforms operating in the EU. The act aims to create a comprehensive framework to hold digital services accountable for content moderation, algorithmic transparency, and online safety.
Platforms must implement risk management systems, handle user complaints efficiently, and provide data access to authorities. With fines of up to 6% of global turnover for non-compliance, the DSA aims to create a safer digital space while protecting users' fundamental rights – a bold regulatory approach that has become a flashpoint in international tech policy.
But none of that matters to the US government, which, according to Wired, has been furiously writing to the tech giants demanding they ignore some of the DSA’s key provisions, particularly around encryption.
I’m not a fan of all the provisions of the DSA, but the EU has the right to determine its own laws, just as the US does.
5. The canonical example of enshittification
I’ve seen the word “enshittification” applied to many things which it doesn't really fit.
Enshittification, as defined by Cory Doctorow, describes the three-stage decline of digital platforms: First, they attract users with good deals and service. Second, they leverage this user base to attract businesses, gradually worsening the user experience. Finally, once users and businesses are sufficiently locked in, they degrade the experience for both, extracting maximum profit while providing minimum value. This downward spiral continues until the platform ultimately collapses under its own dysfunction.
The most perfect example of enshittification in action? I'd argue it's Booking.com.. Like most enshittified platforms, it started out as a fantastic service, but now the majority of Booking.com's revenue comes from hotel commissions, while consumer experience has deteriorated as hotels are pressured to pay for better placement and engage in various deceptive practices.
But Booking.com has a challenger in the race to be the perfect encapsulation of the concept. Peleton-wannbe Echelon pushed a firmware update to its exercise bikes that forced users to connect to Echelon’s servers for the equipment to work and locked useful features behind a costly subscription. This update broke compatibility with popular third-party apps and offline use, upsetting customers who’d bought for device openness and flexibility
Right to repair advocate Louis Rossmann, offered a $20,000 bounty to anyone who could jailbreak and unlock Echelon exercise bikes. And someone did just that.
But of course, it’s illegal to release that code. That’s right: you might think you own that flash exercise bike, but if you tamper with it in any way, you’re basically breaking the law. Which is, of course, why you’re way better off just getting a dumb bike in the first place.
6. All watched over by machines of loving grace
You’ll probably guess from what I’m writing about this week that I’m not a complete AI-refusenik. But there are areas where AI means creating the ultimate “computer says no” culture, making it possible for the people who deliver vital services to simply shrug their shoulders and blame algorithms.
Take, for example, the use of AI in delivering council services. Now there are applications where machine learning, in particular, can be helpful. Councils generate tonnes of data, and machine learning can offer insights which it might take humans years to spot, helping improve service delivery.
But summarising social worker case notes about children with special educational needs really isn’t one of those cases. Part of the point of reviewing notes is to fix more into the memory of social workers, and to help them see patterns and points which they might not have noticed at the time. It’s a human business – not one which we should leave to machines.
7. Sometimes, silence sends more of a message than words
One of the characteristics of authoritarian regimes is their ability to quickly control markets and companies. Donald Trump has very quickly mastered this, probably because threatening people was, and is, the way he does business.
The FT has published a long piece on how leading figures in American business and finance have largely remained silent in response to Trump’s series of aggressive interventions in the financial system and private companies.
There’s a deep layer of hypocrisy at work here. While corporate leaders were quick to criticise progressive mayoral candidate Zohran Mamdani’s economic proposals in New York, they have offered only muted or indirect objections to Trump’s actions, such as firing a Federal Reserve governor, taking stakes in companies like Intel, and pressuring currency and industrial policy.
This restraint is attributed primarily to fear of political retaliation and the greater power wielded by the president compared to a city mayor. Many in the business community believe that private persuasion is preferable to public criticism, although there is zero evidence that has actually paid off. Meanwhile, financial markets have shown little reaction to Trump’s moves, focusing more on rate signals than on the long-term risk to U.S. economic institutions.
Some industry voices and intellectuals have condemned Trump’s actions as “socialism” or dangerous interventionism, but most prominent business leaders and Republican lawmakers have avoided direct confrontation. Despite misgivings, self-interest and short-term market optimism have contributed to the broader “calculated silence” of the corporate elite as the rules of free enterprise in America are fundamentally abandoned.
8. AI plus nostalgia equals something dreadful
Have you ever hung out in a local Facebook group? if you have, you will have noticed there is a very strong theme of nostalgia. People are forever reminiscing about the good old days, when children played safely in the streets, people left their doors unlocked, and everyone knew their neighbours. This form of uncritical nostalgia – sometimes called "rosy retrospection" or "nostalgic bias" – ignores the negative aspects of the past while selectively emphasising the positive.
Now AI tools are joining in, amplifying this distorted view of history. Chuck Jordan highlights a recently-viral video which aims to capture the mood of the 80s, with veiled criticism of the present, while really doing nothing of the sort.
This AI-enhanced nostalgia is particularly dangerous because it comes with an artificial veneer of authenticity. It's feeding into a political narrative that suggests everything was better in the past – and could be again if only we returned to "traditional values." When you can recreate your rosy-tinted memories and share them, the truth of the past is going to be hidden.
9. RIP TypePad
A sad milestone in blogging history: TypePad, one of the pioneering blogging platforms that helped shape the early blogosphere, is shutting down. Launched in 2003 by Six Apart, TypePad offered a more user-friendly alternative to existing platforms with its emphasis on customisable designs and powerful publishing tools.
In its heyday, TypePad was the platform of choice for many serious bloggers, journalists, and publications. TypePad was my middle-ground between the Blogger era and installing my own Wordpress. In its day it was great.
Thankfully, the only TypePad blog that I still read is moving to Ghost. Phew.
10. What’s that about question marks?
I think Terence might be baiting me 🙂
Good article, though. I think the underlying question is not “should we attempt to keep vulnerable people safe from scams” but “who should be charged with protecting people from scams?” To my mind, that’s definitely not huge corporations that could – and will – use that power to generate profits that would have made Rockefeller blush.