Your radical Weekend Briefing is Back
And with that a new podcast episode, radical Connect Meetups, and sooo many links to explore…
Dear Friend –
We are back! I am happy to report that Jane and I achieved “Mission Accomplished” and made it to the summit of Mount Kilimanjaro (proof). Now we are back in the swing of it – with a big weekend issue of your radical Briefing (take a look at the “Rabbit Hole Recommendations” – we made some changes and provided some additional context… Useful?). And also a new episode of the Disrupt Disruption podcast – featuring Gero Hesse, CEO of EMBRACE, in a conversation about the future of HR, AI’s impact, authentic leadership, and a human-centric culture. Give it a listen!
Last but not least, we launched radical connect. The idea is simple: As I travel the globe, we will set up informal get-togethers for the community. No agenda, nothing formal – just us getting together, sharing ideas, and making connections over coffee. The first events will be in Monterrey, Mexico (Sep 12), Chicago (Sep 16), and San Diego (Sep 26). More info and registration here. I can’t wait to meet you all!
And now, this…
Headlines from the Future
Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence ↗
Stanford’s Digital Economy Lab is out with a new paper, analyzing the effect of AI on the early-career job market based on a huge dataset from the largest payroll provider in the US:
“We find that since the widespread adoption of generative AI, early-career workers (ages 22-25) in the most AI-exposed occupations have experienced a 13 percent relative decline in employment even after controlling for firm-level shocks.”
Not all is lost, though:
“In contrast, employment for workers in less exposed fields and more experienced workers in the same occupations has remained stable or continued to grow.”
—//—
Top AI models fail spectacularly when faced with slightly altered medical questions ↗
Shocking (not):
Artificial intelligence systems often perform impressively on standardized medical exams—but new research suggests these test scores may be misleading. A study published in JAMA Network Open indicates that large language models, or LLMs, might not actually “reason” through clinical questions. Instead, they seem to rely heavily on recognizing familiar answer patterns. When those patterns were slightly altered, the models’ performance dropped significantly—sometimes by more than half.
—//—
Coinbase CEO explains why he fired engineers who didn’t try AI immediately ↗
Here’s one way to deal with the developer talent shortage – just fire ‘em:
“I jumped on this call on Saturday and there were a couple people that had not done it. Some of them had a good reason, because they were just getting back from some trip or something, and some of them didn’t [have a good reason]. And they got fired.”
This is Coinbases’ CEO speaking – must be a lovely place to work at…
—//—
The AI Workforce Reckoning (How Will AI Affect the Global Workforce?) ↗
Goldman Sachs Research released a new analysis in August 2025 examining AI’s impact on global employment, finding that AI-related innovation will cause near-term job displacement while simultaneously creating new opportunities elsewhere. The research suggests economists expect generative AI to lift labor productivity by approximately 15% at full adoption while nudging unemployment up by about 0.5 percentage points.
Behind the news:
This latest Goldman research builds on their earlier 2023 analysis that predicted generative AI could raise global GDP by 7%. The updated findings align with broader industry research from McKinsey suggesting that by 2030, activities accounting for up to 30% of hours currently worked across the US economy could be automated. However, Goldman’s research takes a more nuanced view than some predictions of mass unemployment, emphasizing historical patterns where new opportunities in emerging sectors have ultimately offset jobs displaced by automation.
The Goldman findings suggest we’re entering a transition period rather than facing an employment apocalypse.
What We Are Reading
🤖 Can AIs Suffer? Big Tech and Users Grapple With One of Most Unsettling Questions of Our Times The question isn’t whether AIs are conscious; it’s whether we’re ready for them to lawyer up. @Jane
🏃 Nearly Half of Employees Say a Bad Onboarding Made Them Want To Quit First impressions really do matter: the quality of an employee’s onboarding can affect their long-term motivation and even how long they plan to stick around. @Mafe
📊 The Urgent Need for Revolutionizing Economic Statistics Continued digital transformation of commerce and culture is making it increasingly difficult to reliably measure economic activity. @Jeffrey
🧑🏼💼 AI Is Making the Workplace Empathy Crisis Worse AI may be great at efficiency, but it can’t carry the human side of leadership. If we’re not careful, it widens the empathy gap that is already plaguing the workplace. @Kacee
✈️ Is Air Travel Getting Worse? No, it is not just you. Well, partially. As someone who spends a lot of time on planes, it surely feels like air travel has gotten much worse in recent years – it turns out, it’s more complicated than that. @Pascal
Rabbit Hole Recommendations
Benchmarks are not everything! Despite the frontier model’s impressive scores, once confronted with imperfect information, they falter: “Top AI models fail spectacularly when faced with slightly altered medical questions”
You hear a lot about agentic AI – this is what happens when you try to actually use it: “We put agentic AI browsers to the test - they clicked, they paid, they failed”
For Agentic AI to really work, it needs to be able to complete long(er) tasks – something current systems can’t really do all that well: “Measuring AI ability to complete long tasks”
Vibe coding is bad for you – at least from a security perspective (not surprising): “We asked 100+ AI models to write code. Here’s how many failed security tests”
Another look at the AI Coding revolution – and why not everything that shines is gold: “Why AI isn’t ready to be a real coder. AI’s coding evolution hinges on collaboration and trust”
Read this with a grain of salt – the study has received pretty intense scrutiny: “MIT report: 95% of generative AI pilots at companies are failing”
A good reminder why “AI-first” as a strategy is not sufficient: “Most AI strategies fail for this reason.”
The common narrative is that AI is getting cheaper – specifically, tokens are getting cheaper. Well, not so quick: “Tokens are getting more expensive”
Think of AI what you want – the staggering investments into its infrastructure are propping up the US economy: “How AI conquered the US economy: a visual FAQ”
Venture Capital powerhouse a16z just released their updated “Top 100 Gen AI Consumer Apps” list
That Google study, which showed that AI’s water usage isn’t a big deal? Maybe some clever numbers engineering: “Google games numbers to make AI look less thirsty” and “3 problems with Google’s AI energy use data”
2nd and 3rd order implications of selling your digital likeness: “He sold his likeness. Now his avatar is shilling supplements on TikTok.”
It shouldn’t come as a surprise that GenAI is a boon for (organized) misinformation: “Foreign disinformation enters AI-powered era”
College grads are facing a very strange, very different job market: “Fear Of Super Intelligent AI Is Driving Harvard And MIT Students To Drop Out” and “Is AI why college grads can’t find jobs?” (h/t to Manos)
The onslaught continues: “Google has eliminated 35% of managers overseeing small teams in past year, exec says”
Is this the end of that brief moment in time where we actually cared about workers? “Howard Levitt: The myth of work–life balance is dead, and employers aren’t afraid to say it”
Short of market distortions through tariffs and other governmental actions, there is little doubt that China is winning the car race: “Ford CEO Jim Farley: China’s EV costs, tech, and quality “far superior” to the West”
Streaming was once hailed as the perfect antidote to piracy – the continuous enshittification of the services leads to this: “Can’t pay, won’t pay: impoverished streaming services are driving viewers back to piracy”
Happy Distractions
🛰️ Every satellite orbiting Earth and who owns them
🐟 Draw a Fish! (and make it swim)
🏸 Physics of badminton’s new killer spin serve
🧒🏼 The Nick Simulator: Welcome to the life of Nick, 30.
😱 Research shows the ‘compliment sandwich’ is no longer effective
🎮 Incredible collection of old PC game boxes
👷🏼 In case you’d like to geek out on a truly incredible work-from-home setup: Brooklyn Work-From-Home Study
💼 Not that this needs any repeating: Sunny Days Are Warm: Why LinkedIn Rewards Mediocrity
🚇 A music box of train station melodies from the JR Yamanote Line