Nine Nuclear Reactors Worth of Hype
Walmart's AI shopping experiment crashes, AGI benchmarks humble Silicon Valley, and the ads have officially reached the refrigerator
Dear Friend,
Pardon my French (in my defense, it’s not my headline), but Mario Zechner’s “Thoughts on slowing the f*** down” is a good reminder that all the wondrous things AI can and does do for us come at a cost – hence his reminder to: “[…] slowing the f*** down and suffering some friction is what allows you to learn and grow.” With that in mind – time to slow down, welcome the weekend, and dive one last time into our wild future before we call it a Friday.
P.S. A new episode of our podcast dropped: Jason Goldberg has spent 30 years watching companies survive – and get destroyed by – disruption in retail. His counterintuitive advice for the agentic commerce moment: stop trying to be first, and start asking what you’ll regret not doing when the future arrives.
And now, this…
Headlines from the Future
AIs Energy Demands Are Truly Bonkers. Japanese tech giant SoftBank is building a massive 10GW data center in Ohio to host AI models. Aside from the cool $30–40 billion price tag, it will require the build of a $33 billion natural gas power plant – with an insane output capacity (emphasis mine):
When completed, the new site could be one of the largest AI data centers ever built. Furthermore, it will be powered by one of the world’s largest fleets of gas turbines, equivalent to the energy supply of nine nuclear reactors.
It does leave you wondering where and how all this will end.
━━━━━
Maybe AI Isn’t Online Shopping’s Future After All. After the initial hype of online shopping results being incorporated into the answers LLMs give to the numerous product-related queries they receive, Walmart unveiled that the conversion they are seeing from those AI-referrals is just terrible.
After testing 200,000 items in ChatGPT, Walmart found sharply lower conversions and will use its own integrated shopping experience. Walmart said conversion rates for purchases made directly inside ChatGPT were three times lower than when users clicked through to its website.
Next: Agentic commerce. The jury’s out.
━━━━━
What 81,000 People Want From AI. Anthropic, the AI company which is not OpenAI, conducted what is, in their own words, likely the largest study on users’ desires, wishes, and fears when it comes to their use of AI. Anthropic being Anthropic, they didn’t survey people using a traditional questionnaire, but rather had their chatbot “talk” to people. The findings won’t surprise you – people want to use AI to better themselves: professional excellence and increased productivity, which translates into the very human desire to, ultimately, live better. And respondents live the Scott Fitzgerald quote we are so fond of quoting – they keep the light and the dark of AI in their heads simultaneously.
“AI should be cleaning windows and emptying the dishwasher so I can paint and write poetry. Right now it’s exactly the other way around.”
━━━━━
AGI? Not so Fast! AGI, or Artificial General Intelligence, is the thing Sam Altman and others love to talk about – and promise is just around the corner. To demo their respective companies’ progress, they roll out benchmark after benchmark showing how their AI beats humans on the sommelier exam. A new benchmark, however, shows that AGI is still a long, long way off. The ARC-AGI-3 benchmark pits leading AIs against humans in a series of computer games – and AIs don’t look all that great. To apply a lesson my statistics professor hammered into our heads: Never trust a statistic you haven’t faked yourself.
What We Are Reading
Ads Are Popping up on the Fridge and It Isn’t Going Over Well Ads are literally popping up everywhere (even on Google Maps starting this summer), but people are particularly irked by ads on expensive refrigerators with a big screen for recipes, weather updates, and, apparently, ads. @Mafe
The Insurance Catastrophe A deep dive into the history & future of insurance markets offers a fascinating lens for exploring how communities, societies, and economies deal with radical uncertainty and catastrophic risk. @Jeffrey
The 60-Year Degree: Why Universities Must Pivot from Recruitment to Perpetual PartnershipHigher ed has been at an inflection point for years; the degree is just the 1st casualty of a shift to lifelong contracts. @Kacee
AI Slop Is Infiltrating Online Children’s Content Surprised is, of course, no one. But it does leave you wondering what happens to the brains and cognitive development of children who are exposed to AI slop from an early age. @Pascal
Down the Rabbit Hole
🍿 Hilarious take on the world of Enshittification by the Norwegian Consumer Council (hat tip to Angel Grimalt for the link).
🥷🏼 AI agents going rogue: The more we rely on AI, the more we deploy AI agents, the more we see fun headlines like this: Meta is having trouble with rogue AI agents – now consider what this means for any company not the size of, or with the resources of, Meta!
🍺 Talking about cyberattacks and our ever-increasing reliance on Internet-connected technologies: Cyberattack on vehicle breathalyzer company leaves drivers stranded across the US.
👨🏼💻 Nerd alert! But super helpful: Here is a Claude Skill – Prompt Master – which helps you create better prompts, highly optimized for specific use cases, tools, and target LLMs.
🤦🏼 Yep, bro… Whatever. “Perplexity CEO says AI layoffs aren’t so bad because people hate their jobs anyways: ‘That sort of glorious future is what we should look forward to’”
⚓ The running and cycling app Strava has been used to track the location of military outposts before – now the French newspaper Le Monde has used it to track the location of France’s aircraft carrier. Note: Your public data is public data.
🧬 Fascinating read on the adaptability of the human body: Tribe in Kenya evolved genetic mutation that lets them survive with almost no water.
🥢 A Japanese glossary of chopsticks faux pas.
🧳 Lovely journey through 30 years of the web.
🎧 Peak 80s nostalgia: The Maxell Wireless Cassette Player.
↗ Dive into the deep end: Access our complete collection of 2,600+ 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.