Dear Friend,
This week I had another conversation with a friend who recently left the corporate world to branch out on his own. We traded notes on how we use AI – and agreed we both get massive value out of our subscriptions. Contrast this with our first headline from the future (and previous reports on the topic.) Maybe for the first time, the advantage of a new technology sits with the individual practitioner or small team, not with cash-rich corporations.
And now, this...
Headlines from the Future
The Question of AI Economics: Torsten Slok, investment firm Apollo’s Chief Economist (and a semi-regular in the radical Briefing), showed more evidence that the ROI on AI investment outside Big Tech isn’t there (yet). Revenue per worker and profit margins are climbing at the Mag 7 and flat across the other 493 companies in the S&P 500. Productivity growth shows no AI bump yet. I believe it’s down to the nature of the work (much of what these firms do doesn’t lend itself to AI-driven gains – at least not yet) and the difference between doing something as a one-off or pilot versus a reliable process running millions of times. As Slok puts it: “The bottom line is that the AI story will be decided in the S&P 493, not the Mag 7, and so far […]”
Plug Your Child Into the Matrix: Amazon’s latest tablet made for kids points toward a dystopian future of (even more) screen time for your toddler. While country after country establishes rules to minimize screen time and social media use for teens, Amazon pitches it as your three-year-old’s “first tablet” and touts its product’s ability to keep your kid pacified on the flight and the road trip (and everywhere in between). With AI rewriting education and our children’s future jobs, we should invest in our kids’ ability to create and think critically instead of outsourcing this to Amazon’s AI “Storybuilder” and “Art Studio” tools, which come bundled with the new tablet.
“It’s Offensively Cheap”: And now for a bit of good news: Solar energy is projected to overtake nuclear as a global power source this year – and the lead will only grow. Production prices keep dropping, battery storage gets better and cheaper, and the trend to electrify everything continues. Sunny days indeed – and much-needed news in a world which will consume ever more energy to keep the AI machinery running.
Behold the Artificial Intelligence Company (AIC): Delaware, the US state where most Fortune 500 companies are incorporated and long known for its pro-business legislative and tax stance, has drafted a bill for a new legal entity. Artificial Intelligence Companies would be legal entities run and managed by AIs – no human involved in running the day-to-day operations. The AIC “can sue and be sued in its own name, hold and dispose of property, and incur obligations” – but its owner isn’t liable for anything the AIC does. If this sounds to you like “an overly generous liability shield,” you’re in good company – Berkeley law professor Andrew Baker called it exactly that, and House Oversight chair James Comer just opened an inquiry.
The AI (and Your Boss) Is Watching You: In light of companies’ struggles with the ever-expanding spend on AI tokens (see the first post above), Anthropic now offers new reporting tools that show how and why your employees use AI – Claude reads the transcripts and flags “personal or off-topic” sessions as inefficiencies. Time to rein in those personal AI chats about your next vacation.
Trust AI, Think Less: A Microsoft survey of 319 knowledge workers finds that the more we trust AI output, the less we engage in critical thinking. The data is self-reported, so it shows perceptions of effort, not measured changes in thinking. The flip side is that workers confident in their own skills think harder. That matters for the human-in-the-loop discussion – a human who trusts the machine more than themselves stops checking.
What We Are Reading
The Myth of Safe AI Forget killer robots: this article argues that the hardest part of making AI safe isn’t technical or philosophical at all; it’s money. @Jane
The Best Strategy to Protect Against Getting Hacked Is Surprisingly Old-School AI-powered attacks call for the same security basics as always (limited data access, fast patching, backups, MFA), plus guardrails on your own AI agents and a recovery plan, since experts say a breach is a matter of when, not if. @Mafe
Face It: You’re a Crazy Person (or: Why Your Brain Needs a Boxcutter) The secret to finding your calling (or at least a role / career / business that could offer you long-term satisfaction) is committing to the work of truly unpacking the unglamorous details – of both the “opportunity” and what you as a uniquely strange individual bring to it. @Jeffrey
Frontier AI Labs Say Their Models Refuted the Exponential Time Hypothesis. What Does That Mean? AI may have just challenged a decades-old assumption about the limits of computing. The findings still need to be verified, but they’re already pushing scientists to rethink what’s possible and what questions to tackle next. @Kacee
The Safe Internet Is Shrinking Ever more powerful AI leads to ever more powerful cyber-attacks. With so much of our lives happening online now, we had better get used to an era of continuous discontinuity, as the tools, services, and websites we rely on will be taken offline with increasing frequency. @Pascal
...and now for some fun!
🍏 An incredible stash of photos from the early days of NeXT (Steve Jobs’ company after and before Apple).
🧱 All of Switzerland’s topography in 1x1 meter blocks, put into a game that makes this world walkable. It’s essentially Minecraft on steroids.
🧑🎤 Miss the 90s? Now you can create word art like it’s 1994.
🖼️ Here are some highly practical tips on how to best use AI image generators.
🍿 Disney’s lovely deep dive into what it takes to make an animated feature film.
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.
↗ Dive into the deep end: Access our complete collection of 3,200+ radical links.