Containment Is a Choice
Professors vanish into labs that won’t let them publish, PwC’s $1.5bn buys hallucinated reports, and the Vatican leaks 700,000 users.
Dear Friend,
This week, while compiling the list of links for this Briefing, it occurred to me that the average length of a blog post I consume has increased significantly over the last two years. Blog posts (at least from the sources I consult – which are quite a few) used to be pretty dense and to the point. Not anymore. Now they often are rambling multi-thousand-word monsters. And, of course, if you have seen your fair share of AI-generated/co-written content, you also see the telltales of AI use… Nothing against the use of AI – but when it comes to someone expressing their thoughts, observations, insights, and opinions, I’d side with Blaise Pascal (and no, it wasn’t Mark Twain who said this): “I would have written a shorter letter, but did not have the time.”
And now, this…
Headlines from the Future
Rethinking Legal Education in the AI Era. As educational institutions globally grapple with the question of how to prepare students for a future where AI will be a ubiquitous part of the workplace, the team at the Law School at the University of Chicago just published a thoughtful plan on the topic. Key element: strip devices from 1L courses to protect foundational skills while allowing AI use in later ones.
We will be piloting a coordinated approach to classroom and examination policies for the core 1L curriculum during the 2026–2027 academic year. Across all 1L sections, we will prohibit the use of electronic devices such as laptops, tablets, and phones in the classroom. There will be some limited exceptions to this policy.
All in all, definitely one of the better approaches to AI in higher education I have seen.
━━━━━
The OpenAI Hacking Incident Is That Gift That Keeps on Giving. By now, everyone with even a remote interest in AI has heard about OpenAI’s rogue AI agent hacking into the systems of AI company Hugging Face. Turns out, Hugging Face wasn’t the only victim:
Hugging Face was thought to be the only victim of the unprecedented hack - but OpenAI now admits its bot attacked several “publicly-available services”. The out-of-control AI found four logins online which allowed it to access four separate, unnamed services.
Personally, I am much less interested in the specifics of the attack (in case you are, Hugging Face has published a long and detailed analysis of the attack on its blog) and rather in the question of how, specifically, the agent broke out of its containment. And, much more interesting even, what this means for threat actors running their agents on systems and infrastructure which don’t even have such a thing as a harness to begin with.
Honestly, if your company has servers and data exposed to the internet (and who doesn’t), sleeping soundly at night, knowing that your cyberdefenses will hold up, might be a thing of the past.
What We Are Reading
Where Did All the Computer-Science Professors Go? The professors building the future of AI have left campus, and they’re not allowed to publish what they find. @Jane
Google Deepmind Dismantles Nobel-winning Alphafold Team in Strategy Shift Google DeepMind broke up its Nobel Prize-winning AlphaFold team, reassigning researchers to LLM development, enzyme design, and nuclear fusion, resulting in lead members leaving for Anthropic. @Mafe
AI Mania Is Eviscerating Global Decision-Making If you’re running (or even thinking of running) big AI projects in your org, you’d better be sure channels for dissenting opinions and sanity/reality checks remain open. @Jeffrey
The Newest Way to Go Analog iPods, digital cameras, and DVDs are returning – not because they’re truly analog, but because they do one thing instead of everything. The future may belong to products that deliberately do less. @Kacee
Elite Young Runners Are Becoming Freakishly Fast. Welcome to ‘Trackflation’ First, we broke the magic sub-2-hour marathon, then we ran the fastest mile; running is seeing records being broken which stood for years. There is a reason for that. @Pascal
Down the Rabbit Hole
👨💼 OpenAI’s latest report on how AI is reshaping work has an interesting twist: maybe one of the biggest changes is that AI is dissolving occupational boundaries. Workers use it to do tasks that once belonged to other jobs, reshaping who does what before titles change; 43.5% of occupation-specific messages (of 800k+) fall outside the user’s own occupation.
⛪ Even the Vatican isn’t save from cybersecurity foot-guns: Security flaw in Vatican’s ‘Click to Pray’ app leaves over 700,000 global users exposed – app has been leaking user data for over six months and still does.
🦹 Yikes! Hackers use DNS poisoning on hotel Wi‑Fi to steal Microsoft 365 accounts.
📦 Global oil supply chains are hyper-complex, an incredible feat of engineering and management, and fragile (as the whole Strait of Hormuz debacle shows). Here is a gorgeous visualization of how oil flows around the world.
👨💼 Replace your CEO with AI and save money – big time! “OverpAId is an Artificial Intelligence built from the ground up to do your CEO’s entire job – strategy, “vision,” motivational all-hands emails – better, faster, and without ever once asking the board for a bigger jet. Runs on a single desk-sized AI computer. Real hardware, real price, zero mystique.” LOL.
👔 You invest $1.5bn into your AI initiative and you will get AI slop: PwC published reports on AI marred by AI hallucinations.
👶 If you have kids, you really ought to think hard about the whole smartphone thing: “We found that there’s no age at which you can give a child a phone that’s better or worse for reading comprehension, so it’s just having a phone at all […]” – Kids who own cellphones have poorer reading comprehension scores than those who don’t, researchers find.
🗺️ What if Google Maps – but for ancient Rome? Here it is!
⛳ Love the game of golf? Here is a map of all the golf courses in the US.
↗ Dive into the deep end: Access our complete collection of 2,900+ radical links.
What We Do When We’re Not Writing
Hi! I’m Pascal from radical. When we’re not writing this newsletter, we help organizations turn volatility into advantage – without the “innovation theater.” We build your team’s capacity to handle disruption, so you stop reacting and start shaping. If that sounds useful, let’s talk.