Good morning. What happens when kids raised on tablets meet paper books? Chaos. A UK study found nearly one-third of preschoolers can’t figure out how to use an actual book. They’re swiping, tapping, and pinching pages, baffled that nothing happens. Teachers blame screens and parents for creating a generation of “iPad babies.” Congrats, Silicon Valley, you’ve disrupted childhood. Can’t wait to see what happens when these kids grow up with brain-computer interfaces.
❄️ A shout out to all our readers weathering the massive winter storm. It’s a skating rink out there, but so far we still have power. Hope everyone is warm and safe!
Self correcting AI agents will never be autonomous. Researchers just tore apart the idea of AI agents fact checking their own work. Asking one AI to verify another’s work is about as useful as two drunk friends giving each other sobriety tests. Instead of catching mistakes, chained agents happily amplify each other’s nonsense. You simply can’t outsource sanity checks to equally clueless bots. So rather than paying twice for AI hallucinations, maybe the future of autonomous agents is just better data and tighter context. RAG for the win. (Wired)
Psst, hey you. Ever wonder if AI actually makes money? OpenAI just launched a new tool that calculates exactly how much your company could earn by implementing their AI. Think about that for a second. Businesses still debate if AI genuinely boosts productivity. So OpenAI is handing them a tool that predicts their exact ROI. Clearly, the fastest way to pry open hesitant wallets is by putting potential profits into a neat little spreadsheet. Executives love spreadsheets. (The Information)
As the pace of innovation accelerates, leaders need trustworthy, actionable insights to navigate AI with confidence.
This essential report by You.com’s founders will help you:
- Anticipate challenges
- Spot opportunities
- Make strategic decisions in 2026 (and beyond)
Be at the forefront of the next wave of innovation. Get the report.
AI leaders can’t agree how close AI is to human level intelligence. At Davos, Google DeepMind CEO Demis Hassabis said today’s AI isn’t even close to thinking like a person. Yann LeCun, a pioneer of modern AI, called large language models a dead end. Anthropic CEO Dario Amodei claims AI will replace software developers within a year, but is coding really the benchmark for human level AI? Hassabis says we have a 50% chance of achieving AGI in the next decade. Good to know we’re betting the entire economy on a coin flip. (Fortune)
China’s AI startups are valued like Nvidia without the profits. In just the past few weeks, Chinese AI model developers and chipmakers have raised over $4 billion in massive IPOs. Biren Technology is trading at an absurd 248x last year’s revenue, while Nvidia itself sits at a modest 24x. Startups like Zhipu and MiniMax make OpenAI’s valuation look timid, despite everyone involved still bleeding cash. Investors are betting big that China’s AI industry can close the technology gap, and with valuations like these… there’s never been a better time to shop for pre-IPOs. (The Information)
Clay hit $100 million ARR by outsourcing sales work to OpenAI. When Kareem Amin founded Clay in 2017, his original vision was ahead of what tech could actually deliver. Instead of giving up, he pivoted to automating the tedious tasks salespeople hate like finding leads, researching contacts, and drafting outreach messages. Clay went all in on OpenAI’s API, creating GPT powered agents that handled sales grunt work at scale. Soon they were processing trillions of tokens every month, becoming one of OpenAI’s largest API customers. Revenue exploded from $1 million to $100 million in just 2 years, proving yet again the fastest path to fortune is letting AI handle tasks nobody likes doing. (Inc.)
fun stats
🏄♂️ 1 million+. Number of “.ai” domains (including ours), pumping ~$70 million into the tiny Caribbean island of Anguilla. All thanks to the AI gold rush and ICANN’s random 2 letter gift.
🤨 50%. Companies Gartner predicts will adopt zero-trust data governance by 2028. Once AI generated slop passes for real data, betting your business on an LLM gets risky.
🔦 35,000+. Users reporting issues with TikTok since US takeover, with 22% experiencing total blackout.