Hello again,

For our top story this week, we look at the very near future of internet search. Google's new search is here and its making people equal parts nervous and excited.

We also check in on open source darling Grafana Labs, workplace annoyance Microsoft Teams, and what happens when a billionaire buys a superyacht - but not for the reasons you think.

And for our blog spotlight, we look at Technofacism.

Let's dig in...

Top Story

🔍 Google search as you know it is over

A quick summary

Google has announced a complete rethink in the way it handles web searches. Soon, the list of blue links we've had since the early 2000s will be gone, in favour of "generative UI" - a system that will change its output based on whatever it was you searched for. Think interactive diagrams for scientific queries, or a digital shopfront for product searches.

Powered by AI (of course), the new search will accept everything from text and images, to lists of open tabs in Chrome. It's the biggest overhaul in online search since... well, since Google.

Why are they doing this?

In short, because of ChatGPT. For the first time in its history, Google has a genuine competitor in the web search game. More and more people, especially young professionals and academics, are using ChatGPT to do research and ask questions. With OpenAI reporting 900 million active users a week at the start of the year, Google needed to do something big. This is something big.

I think there's another reason too. Linked to needing to maintain market dominance, for most tech companies at the moment, the only real game in town is "Artificial General Intelligence" (AGI). Think Jarvis from Iron Man, TARS from Interstellar, or Sonny in I, Robot. AGI is a tool you can throw literally anything at, and it will perform that task. Whoever cracks AGI first gets to print all of the money. It won't matter who comes second. While ChatGPT and its ilk are undeniably impressive technological advancements, they are still a far cry from AGI.

The current bet is that in order to get there we need more data. Lots more. Keeping you locked into the Google ecosystem, sharing all that yummy data on who you are, how you think, what you believe, has never been more important for the company. Gone are the days where Google's aim was to point you to the most relevant website for your query. Now, they need to be the most relevant website for any query.

What does this mean for the web?

The most obvious threat is to publishers. Already we've seen dozens of online publishing companies collapse, because they weren't getting the eyeballs they needed to bring in ad revenue. Even big players like The New York Times, and HuffPost have seen dramatic drops in traffic. Business Insider saw a drop of over 50% in organic traffic between 2022-2025, leading them to cut 21% of its staff.

ChatGPT is getting the blame for this, but actually it seems that Google's AI Overviews have done the most damage. The precursor to what's to come, the little box at the top of Google searches have cut page views from Google to other sites by 34% in 2025. Concerning, when you consider that around 10% of those summaries are just wrong.

Even before Google's latest announcement, publishers like Reuters were gearing up to lose an average of 43% of their web traffic before the end of the decade. That number is going to seem like small potatoes now.

For a short window of time, AI was a tool that helped pull together data from different sources, and help point people in the direction of more information that would help with their query. It was like having someone help you Google stuff. Now, we're looking at a reality where all of that data comes to you. It might sound good in theory, but it seems unlikely that people will continue to produce high quality content if they aren't being properly rewarded for that - whether that reward is in views, subscriptions, ad revenue, or otherwise.

The longer term problem

"But if AI is trained on articles from publishers, and all the publishers go out of business, what will train the AI?" I hear you ask, wisely.

If you can't beat them, then some publications have opted to join them. The Guardian have jumped into bed with OpenAI, licensing their content to allow it to be used as training data. They aren't the only ones. News Corp., CNN, Fox News, and others have signed similar deals with Meta. It seems to be the path of least resistance for organisations not looking to rock the boat, while still maintaining a decent revenue stream.

Not everyone is satisfied with that plan though, and others have gone down the route of legal action against AI companies. The New York Times is one organisation that currently suing OpenAI and Microsoft for copyright violations (stealing articles to train AI). Separately, they are also suing Perplexity for similar reasons. They aren't the only ones. Press Gazette has a handy list of all the latest litigation action. The trouble is, the lawsuits themselves are expensive and drawn out, and the eventual settlements are not likely to be large enough to either a) recompense the publishers for all of the lost revenue, or b) dissuade the AI companies from doing it again.

As for how the AIs will be trained in a post-publisher world, don't worry! Plenty of AI is already being trained with "synthetic data". Synthetic data is data that has been artificially generated to be statistically similar to a base set. Yes, you understood that correctly. We are training AI using data that has been generated by AI. And it's going about as well as you'd expect.

What else is happening in the world of tech?

⛵ Billionaire Gabe Newell converts superyacht into a support and research vessel

Sick of hearing about rich tech bros doing nothing but ego stroking vanity exercises with their oodles of cash? Well dear reader, might I introduce you to Gabe Newell. For 90s and 00s gamers, the esteemed Mr Newell is well known. Co-founder of software development and distribution company Valve, Newell has presided over some of the most defining video games of the last 3 decades, including the Half-Life series, the Portal games, and the Left 4 Dead franchise, not to mention the surprisingly popular Steam deck console.

What does any billionaire do with all that money? Well, buy a disgustingly large super yacht, of course. The thing is Gabe didn't get the second half of that memo, where you turn it into a hedonistic den of rich frivolity. Or maybe he did, and his idea of hedonism is researching the shit out of marine life? Either way, his latest big spending is the overhauling of what was once the peak of seafaring luxury into a support vessel for his fleet of marine research ships. Newell is gearing up for some serious science-ing, and the flagship of his fleet is expected to launch in 2028, and play host to 70 scientists on a multi-month mission. All the data from his expeditions will be made available to the global research community.

Wonderful stuff. But Half Life 3 would also be kinda cool Gabe, if you get a chance...

💰 Grafana Labs hacked, refuses to pay ransom

Grafana Labs are one of the most popular open source analytics and visualisation tools on the web. This week, they reported that a hacker had gained access to some of its development code, and was trying to blackmail them for its release. Grafana Labs, apparently acting on the basis that a) none of the data stolen was customer information, and b) there probably wasn't a huge amount of code that wasn't already freely available, refused the pay the ransom.

This particular case seems to have been relatively easy for Grafana to shake off. But it seems open source software is coming more and more under fire. A hacker group calling themselves TeamPCP have recently claimed to have stolen access to over 4,000 Github repositories, in a so-called "supply chain attack". Rather than outright stealing programme code, these attacks take advantage of software deployment methods to inject bad code into open source applications. The eventual target is then any end-user that downloads and makes use of these bits of software.

Its causing concern among some in the open source community, while others point out that code reviews should be both part of a normal process, and would weed out most of these instances of bad code.

💻 Microsoft is retiring Teams' Together Mode

And finally, the Teams feature you never used is going away. Together Mode was a way for multiple people on a call to feel like they were all... watching the same uni lecture, I guess?

Turns out, this functionality was actually quite complicated to make work, so in a rare case of Microsoft axing something that nobody uses, we say cheerio to a feature that did - at best - absolutely nothing for the focus of a call.

Let me know, are you going to miss Together Mode? Or did you just find out about it right now?

Blog spotlight

👨‍💻Technofascism

For ages, I've been saying that I've got a blog post in my drafts folder titled "I'm not surprised that tech bros are suddenly fascists, and you shouldn't be either". I'll post it someday, probably.

But while I'm procrastinating about the subject, programmer, author, and educator, Greg Wilson has been attacking it head on. In part of an ongoing series, Technofascism looks at how our various technology overlords have been building towards "TESCREAL" for a while now.

Its well written and well cited. Go have a read, certainly of this post, but also of his series exploring how damaging big tech is to the fabric of our society.

And that's it for another LT3! Let me know what news caught your eye this week.

Thanks for reading, and I'll see you back here next week.

Will