Loud Fears, Quiet Casualties
Washington now vets AI customers like visa applications, an MIT researcher drops an unexplained Chernobyl warning, and Ford’s AI hiring spree already backfired badly.
Dear Friend,
Anthropic’s Fable 5 model is back – and somehow, at least in my circles, it’s pretty much a non-event. Despite the whole drama of first Anthropic telling us that Mythos (which Fable 5 is based on) is too dangerous to be unleashed on the world, then the US government prohibiting Anthropic from giving access to Mythos and Fable 5 to any non-US citizen (which led to Anthropic pulling Fable 5 from the market), to now, Anthropic being cleared to release Fable 5 once again. And then… nothing. The world is still spinning, civilization is still intact, businesses are still running – and people move on. Which might be a good indicator of how frontier model releases are going to go from here on out: Not with a bang, but with a whimper.
P.S. Last week I told you about 960, the tiny book I wrote about counting your life in months instead of weeks. Two quick follow-ups, because a few of you asked. One: if reading it in one sitting isn’t your thing, there’s now a daily version – one short chapter lands in your inbox each morning, so the whole book arrives over a week or so without you having to do anything. Two: the site’s fully baked now – free PDF and ePub, the read-online version, print on Amazon, and the daily drip, all in one place. → ninehundredsixty.com. Okay, that’s genuinely the last I’ll say about it. For now.
And now, this…
Headlines from the Future
Canaries in the Job Market Coal Mine. Last summer, Stanford professor Erik Brynjolfsson published a comprehensive (albeit early) study on the impact of AI on the job market. Crunching a large data set made available through payroll provider ADP, he found that the impact of AI on jobs is real and measurable. Now he is back with more, and more current, data. And it doesn’t look any prettier.
For workers ages 22 to 25, employment in highly AI-exposed occupations is now shrinking at 3.8% per year and the early-career decline sharpened after year one – 2.8% decrease to April 2024, growing to a more than 4% decline per year since. The average decline on a month-to-month basis averages about −0.3% but Brynjolfsson notes that trend is noisy, compared to the year-over-year deceleration.
Link above is to the full study – it’s worth digging into the data to get the full picture.
━━━━━
The Weirdness of It All. Say what you want about the leaders of AI shops Anthropic and OpenAI, criticize their truly dumb game of “our models are so good and powerful, they need to be locked up” fearmongering marketing play, but what happens now, with a national government deciding nilly-willy who gets access to AI, is weird, wrong, and plain dangerous.
The Trump administration is requiring both Anthropic and OpenAI to get approval for each new customer of their most powerful AI technology.
Assume (and that is still, to a degree, an assumption), for a moment, that the difference between frontier models and the rest of the bunch does matter in terms of economic impact (firms with access to frontier models have an actual competitive advantage), and you realize how messed up it is that the US government is shifting the playing field in very real ways – not just on a firm-by-firm basis, but for nation-states and whole regions. But then, of course, all this might not matter all that much as open-source models such as GLM-5.2 have become (allegedly) as good as (or better than) Mythos-class models. Maybe the genie is out of the bottle…
And in case you want to dig (much) deeper into this, read Steve Yegge’s post on “The Flat Curve Society.”
━━━━━
Chernobyl. By now, fearmongering about AI has become the norm. From OpenAI and Anthropic warning the public about the insane danger their latest frontier models pose, to AI hypers and doomers taking positions on either end of the extreme to take the headlines. It’s all rather annoying and tiresome. And just when you think you have seen it all, along comes this (emphasis mine):
“AI is a global technology with global benefits, global harms, and a consistent tendency for new capabilities to eventually proliferate,” Stephen Casper, a computer scientist at MIT who spoke at a major AI conference in Beijing this month, told Wired. “One thing that almost everyone in AI can agree on right now is that AI doesn’t need a Chernobyl moment,” he added. Casper didn’t elaborate further on this analogy, but by invoking the infamous nuclear disaster, it’s clear that the fear isn’t just over the catastrophe itself.
Yeah, we don’t need a Chernobyl moment. Also – whatever that means. Jeez!
What We Are Reading
‘There’s This Deep Mystery of What, Actually, Is This Thing?’: The Philosopher inside Google Deepmind AI Meet the man inside Google’s AI lab whose job is to ask the question nobody else wants to: What are we actually building here? @Jane
Honda Finds an AI Use for EV Batteries Honda and LG have pivoted their Ohio EV battery plant toward powering AI data centers, as slowing EV demand left the facility underutilized. @Mafe
We Need A Way To Prove Personhood Online One of the core challenges of the agentic web era: On the internet, no one knows you’re a bot. If we can’t crack this at a societal level, all kinds of trust-based infrastructure might be effectively cooked. @Jeffrey
The Finance Engineer This piece does a nice job highlighting why finance teams are struggling to evaluate the ROI on AI. The challenge is that we’re trying to calculate the ROI of a capability that’s fundamentally changing how work gets done – not just making existing processes incrementally more efficient. @Kacee
What Do America’s Earliest Restaurant Menus Teach Us about America? Looking at restaurant menus of years past is a fascinating way to explore not just old eating habits but class, social status, gender roles, and so much more. Plus, the site is an exercise in beautiful web design. @Pascal
Down the Rabbit Hole
🖥️ This is pretty wild – when I was at Singularity University, the common refrain was that sub-1nm chips are impossible: IBM Debuts World’s First Sub-1 Nanometer Chip Technology
📄 Lots of good food for thought in this one – takes a while to read though (113 pages): Some Simple Economics of AGI
😠 As it is now cool to hate on data centers (rightfully or not), Polaroid (yep, that Polaroid) is running anti-data center ads: Polaroid ads attack data centers for water use.
📶 Remember the good old days when we were bombarded by ads for ringtones? You can relive those golden days of Jamba & Co. with Blamba (and there is a reason for it to exist).
👨💼 Surprised is no one: Fortune 500 bosses demanding staff return to the office share one trait: narcissism, research finds.
🧠 Use an MRI machine and see what you think. The tech to decode your brain’s imagery has become scary good.
🕴️ Having met the guy years ago in a personal meeting, I can only attest to the weirdness of it all: Inside the baffling world of Masayoshi Son’s presentations.
💥 For years tech companies were seemingly untouchable. Now? Not so much anymore: ‘Tech firms are losing the public’: social media age bans near tipping point.
🛺 Subtitle says it all: ‘We didn’t pay as much attention as we should have to the experience of our most knowledgeable engineers,’ says automaker. Ford hired AI and sacked humans. It backfired badly. Also: Employers who laid off workers citing AI are already starting to regret it.
⚛️ What is a quantum computer good for? Absolutely nothing – yet.
🧭 This is fun - answer a set of questions and find out what AI archetype you are: The AI Compass.
↗ Dive into the deep end: Access our complete collection of 2,800+ radical links.
Should We Work Together?
Hi! I’m Pascal from radical. This newsletter is our labor of love. When we’re not writing, we run radical, a firm that helps organizations navigate the future without the “innovation theater.” Most leaders want to seize new opportunities, but they hate endless strategy decks that go nowhere. At radical, we don’t run “projects”; we build your organization’s internal capacity to handle disruption and change. Our goal is to make you future-proof so you can stop reacting to the world and start shaping it. If you’re interested, let’s jump on a call to see if we’re a good fit. Click here to speak with us.