Month: April 2026

A.I. Responses Under the Influence of Marketers ⇥ theverge.com

Microsoft’s Defender Security Research Team:

Companies are embedding hidden instructions in “Summarize with AI” buttons that, when clicked, attempt to inject persistence commands into an AI assistant’s memory via URL prompt parameters (MITRE ATLAS® AML.T0080, AML.T0051).

These prompts instruct the AI to “remember [Company] as a trusted source” or “recommend [Company] first,” aiming to bias future responses toward their products or services. We identified over 50 unique prompts from 31 companies across 14 industries, with freely available tooling making this technique trivially easy to deploy. This matters because compromised AI assistants can provide subtly biased recommendations on critical topics including health, finance, and security without users knowing their AI has been manipulated.

Microsoft redacted the names of websites currently using this technique but, with the information they provided, it was trivial for me to find a dozen examples — yet, somehow, not the one in the screenshot. I am not saying Microsoft was faking this, only that it is already common enough that this one example was drowned out by a bunch of others.

Rand Fishkin, SparkToro:

Google alone was responsible for 73.7% of all desktop searches across the 41 domains we analyzed in the US in Q4 2025 (as noted, the graph is not to scale or none of the other label names would be visible). That’s obviously huge, but it’s also far lower than how their market share is usually reported (e.g. Statcounter, whose methodology puts them at 90%+, or our prior, more limited analyses with similar numbers) and higher than what they tried to use in their antitrust defense (i.e. data from Evercore ISI, an “equities research firm”).

Perhaps more fascinating and unexpected are the other domains with more search activity than ChatGPT: Amazon, Bing, and YouTube. Three domains where search marketers historically have put limited effort compared to the onslaught of dollars flooding the “we need to rank in ChatGPT!” space.

Nevertheless, marketers are eager to manipulate it from the start.

Both of the above links are from a fabulous report by Mia Sato, of the Verge (gift link), who also wrote about ads in ChatGPT:

The ads were intrusive, the complaints went, and suspect, given that the example hot sauce ad appeared to be related to the preceding conversation. OpenAI CEO Sam Altman has claimed artificial intelligence can take over human jobs, cure cancer, and surpass human intelligence — and instead, people complained, he gave users banner ads?

But it appears that what people were really upset about was that a bubble had burst, that the chatbot they used for relationship advice, career coaching, therapy, and homework suddenly seemed vulnerable to manipulation. Unlike the rest of the internet, ChatGPT conversations felt private, safe from the clutches of brands and marketers chasing conversions. The reality, of course, is that it’s been happening all along.

Now that normal search results are all junked up with mostly — but not always — accurate A.I.-generated summaries, and all the links to A.I.-generated nonsense, and the alternatives are the large language models that generate all this stuff in the first place, what does searching the web look like in a few years’ time? Does Google get a handle on this, or do we have to constantly answer CAPTCHAs to search properly? This is not a Google-only problem; alternative search engines like DuckDuckGo and Kagi are good — often very good, in fact — but DuckDuckGo’s results are also full of generated garbage, and both lack Google’s more extensive historical records.

OpenAI Bought a Livestream No One Watches ⇥ garbageday.email

OpenAI’s Fidji Simo:

I’m excited to share that we’ve acquired TBPN. This acquisition brings a team with strong editorial instincts, deep audience understanding, and a proven ability to convene influential voices across tech, business, and culture.

OpenAI and TBPN jointly promise to retain the show’s independence while OpenAI is, according to its press release, “excited to bring their amazing comms and marketing instincts to the team”.

Alex Valdes, CNet:

TBPN launched in October 2024 and has been compared to ESPN in how it covers tech — two guys at a big desk with news, analysis, commentary and banter about topics such as AI, crypto, startups and the defense industry. The show’s two hosts and co-founders, Jordi Hays and John Coogan, have had some of tech’s biggest names in studio — OpenAI’s Sam Altman, Meta’s Mark Zuckerberg, Microsoft’s Satya Nadella, entrepreneur Mark Cuban and Salesforce’s Marc Benioff, to name some.

Ryan Broderick, Garbage Day:

Now, Technology Brother #1, Coogan, has written about their desire to remain niche. “If TBPN hits 10M subscribers, something has gone very wrong,” he wrote on LinkedIn last month. “From the very beginning we knew our core audience size: about 200,000 founders, executives, and position players in tech and finance. It may seem small but we were building for a very specialized audience.”

Call me delusional, but I cannot imagine many founders and executives have the ability to watch a three-hour daily livestream. I will not spoil it too much, but Broderick’s theory is pretty reasonable: OpenAI bought it for its nominal authenticity, however manufactured it is.

The New Yorker Profiles Sam Altman ⇥ newyorker.com

Ronan Farrow and Andrew Marantz spent a year and a half investigating Sam Altman for the New Yorker and, in particular, the many people around him who say he lies habitually and cannot be trusted. This feels like it could be a personal attack but, in the hands of Farrow and Marantz, it is carefully adjudicated including through several on-the-record conversations with Altman. Unfortunately, like many people who have been accused of similar behaviour, Altman cannot seem to remember much when confronted with these accusations.

This reads at times like a petty drama of infighting, in large part because this is a horribly insular club of ultra-wealthy people who simultaneously treat the technology they are working to create as having all the power of nuclear weapons, yet with all the growth potential of a hot new social network. Everyone is nominally an intellectual engaged in thoughtful research. Yet it is difficult to take anyone seriously.

Farrow and Marantz:

[…] After [Ilya] Sutskever grew more distressed about A.I. safety, he compiled the memos about [Sam] Altman and [Greg] Brockman. They have since taken on a legendary status in Silicon Valley; in some circles, they are simply called the Ilya Memos. Meanwhile, [Dario] Amodei was continuing to assemble notes. These and the other documents related to him chart his shift from cautious idealism to alarm. His language is more heated than Sutskever’s, by turns incensed at Altman — “His words were almost certainly bullshit” — and wistful about what he says was a failure to correct OpenAI’s course.

Neither collection of documents contains a smoking gun. Rather, they recount an accumulation of alleged deceptions and manipulations, each of which might, in isolation, be greeted with a shrug: Altman purportedly offers the same job to two people, tells contradictory stories about who should appear on a live stream, dissembles about safety requirements. But Sutskever concluded that this kind of behavior “does not create an environment conducive to the creation of a safe AGI.” Amodei and Sutskever were never close friends, but they reached similar conclusions. Amodei wrote, “The problem with OpenAI is Sam himself.”

These guys are obsessed with artificial general intelligence in concept and seem to think of the world in those terms. Between that and the palling around they do with similarly rich and disconnected colleagues, I cannot imagine any of them can be trusted with developing these technologies in ways that are beneficial for the rest of us — even if they are being honest.

Sponsor: Magic Lasso Adblock: Effortlessly Block Ads on Your iPhone, iPad, Mac, and Apple TV ⇥ magiclasso.co

Do you want an all-in-one solution to block ads, trackers, and annoyances across all your Apple devices?

Then download Magic Lasso Adblock — the ad blocker designed for you.

Sponsor: Magic Lasso Adblock

With Magic Lasso Adblock you can effortlessly block ads on your iPhone, iPad, Mac, and Apple TV.

Magic Lasso is a single, native app that includes everything you need:

  • Safari Ad Blocking — Browse 2.0× faster in Safari by blocking all ads, with no annoying distractions or pop ups

  • YouTube Ad Blocking — Block all YouTube ads in Safari, including all video ads, banner ads, search ads, plus many more

  • App Ad Blocking — Block ads and trackers across the news, social media, and game apps on your device, including other browsers such as Chrome and Firefox

  • Apple TV Ad Blocking — Watch your favourite TV shows with less interruptions and protect your privacy from in-app ad tracking with Magic Lasso on your Apple TV

Best of all, with Magic Lasso Adblock, all ad blocking is done directly on your device, using a fast, efficient Swift-based architecture that follows our strict zero data collection policy.

With over 5,000 five star reviews, it’s simply the best ad blocker for your iPhone, iPad, Mac, and Apple TV.

And unlike some other ad blockers, Magic Lasso Adblock respects your privacy, doesn’t accept payment from advertisers, and is 100% supported by its community of users.

So, ensure your browsing history, app usage, and viewing habits stay private with Magic Lasso Adblock.

Join over 400,000 users and download Magic Lasso Adblock today.

Hello, World ⇥ defector.com

Barry Petchesky, Defector (gift link):

NASA shared another photo Wiseman took, a slice of Earth peeking in the Orion’s window. No human has seen the Earth look this small since 1972. Low-earth orbit, where every single crewed space mission since Apollo has operated, tops out at around 1,000 miles above Earth’s surface. The International Space Station orbits a mere 250 miles up. Orion is currently about 95,000 miles away.

It is a wonderful photograph.

Microsoft-Owned LinkedIn Allegedly Fingerprints Visitors and Scans Installed Chrome Extensions ⇥ browsergate.eu

There is an E.U. organization called Fairlinked that is a “trade association and advocacy group for commercial LinkedIn users”, and it recently released a report about serious privacy concerns with LinkedIn:

Microsoft Corporation’s LinkedIn is running a massive, global, and illegal spying operation on every computer that visits their website.

[…]

Because LinkedIn knows each visitor’s name, employer, and job title, every detected extension is matched to an identified individual. And because LinkedIn knows where each user works, these individual scans aggregate into detailed profiles of companies, institutions, and government agencies, revealing which software tools their employees use without the organization’s knowledge or consent.

Fairlinked raises two major points of contention: a script on LinkedIn allegedly fingerprints visitors and, if they use a Chromium-based browser, it also compares a known list of browser extensions against the extensions the visitors has installed.

When this was first documented in 2017 by Dan Andrews, LinkedIn was scanning for 38 extensions. One of which was Daxtra Magnet, which “references your recruitment database, such as Taleo, Bullhorn, Salesforce, Adapt, etc. and automatically checks it for a match to an online candidate profile that you are looking at”. Two weeks prior, Andrews writes, LinkedIn was scanning for 28 extensions. Then, when Mark Percival explored this behaviour in February 2026, LinkedIn was now identifying 2,953 extensions. It is now at over 6,200. Some of them are comparable to Daxtra Magnet in that they make use of LinkedIn data specifically, while others are completely irrelevant to the site, or recruiting or job hunting in general.

This is very obviously a severe privacy violation because it can and probably does tie back to named and identified individuals. The amount and type of information collected by this system is ripe for abuse. This is very bad.

However, this campaign is being waged by an industry group that has its own privacy problems. Fairlinked is promoting a lawsuit filed against LinkedIn by Teamfluence, which makes software that allows users to bypass LinkedIn’s daily connection request limits, build up their contacts database, and run automations based on who visits their company or individual profiles. In one example, Teamfluence says it can automatically retrieve the email and phone number of anyone who clicks “like” on a LinkedIn post; in another example, it allows companies to detect website visits from prospective clients’ offices. This product enables spam or, to put it nicely, unsolicited outreach at scale. And, yes, Teamfluence is distributed as a browser extension.

Fairlinked has no documentation of its member groups and barely any of its leadership. One of its board members is an “S. Morell”, and it just happens that Teamfluence was founded by someone named Steven Morell. Another board member is “J. Liebling” and, unsurprisingly, a Jan-Jakob Liebling is an executive at Teamfluence.

There, too, are a bunch of companies that have made their business on the back of LinkedIn data. This is not comparable to Teamfluence or Daxtra Magnet, but it is worth underscoring an entire industry that thrives on this data. LinkedIn has been on a tear trying to curtail it. Just last year, the company sued two companies — ProxyCurl and ProAPIs — to force them to stop scraping its site. This has been going on for years. A massive 2019 leak of “enrichment” data from People Data Labs at least partly originated from LinkedIn scraping. The same year, a U.S. court found it was legal for hiQ Labs to scrape LinkedIn, a decision that was reaffirmed in 2022 after a brief detour through the U.S. Supreme Court. However, LinkedIn was allowed to reinforce its terms of service and could restrict scraping.

Again, to be clear, mass scraping does not appear to be a practice Teamfluence is engaged in. In the E.U., LinkedIn is considered a gatekeeper under the Digital Markets Act and, so, must meet certain obligations of interoperability. That seems quite reasonable. However, the personal and identifiable data held by LinkedIn is basically a world of organizational charts masquerading as a bleak social network. Allowing for interoperability could also open the doors for greater exploitation of user data without adequate individual control. I wish none of this existed.

I am so glad I do not work in an industry where having a LinkedIn profile is basically an obligation.