Good morning ☕ Looking to build your sleep stack? You may want to keep it low tech. A guy couldn’t sleep because his AI controlled mattress crashed. YouTuber Theo Browne found himself trapped awake when his Eight Sleep smart mattress system app went down, leaving him shivering and unable to adjust the temperature. They promised optimal sleep tech. All he got was a cold bed.
🤠 A Silicon Valley founder wants to build an unregulated tech city in Alameda. He just needs a presidential emergency order to do it. The pitch is to create Frontier Valley, a futuristic AI enclave on the Bay with glass towers, rocket pads, and zero rules. Think of it as Silicon Valley 2.0, free from pesky California laws and oversight. The catch? No one bothered to check with city officials, and the land is slated for development. It’s a lot easier to fast-track tech utopia when you ignore the norms.
🤮 ChatGPT internet pollution is breaking future AI. An explosion of AI chatbots has overrun the web with low quality AI slop. Now that junk is showing up in the datasets used to train newer AI models. Alarm bells are warning. When AI trains on its own output, the quality degrades, triggering model collapse. Clean data from before ChatGPT is now as rare and valuable as virgin land. Turns out when bots feed bots, everyone gets stupider.
🖼️ AI scraping bots are DDoSing museums. Bots harvesting data for training models are overwhelming libraries, archives, and galleries. A new report says the spike in traffic feels like a cyberattack. It’s knocking collections offline and forcing staff to scramble for solutions. At this rate, AI might erase history before it gets a chance to learn from it.
🤕 Robots can now feel pain. Scientists have developed a flexible synthetic skin that senses heat, pressure, and damage. Made from conductive gel, it turns the entire robot into a giant sensor. The fake skin can detect everything from gentle taps and hot surfaces to knife cuts and finger pokes. While not as sensitive as our own, it beats anything else on the market. Soon robots can know the agony of stepping on a Lego.
💸 xAI is burning $1B a month to stay in the AI game. The startup is scrambling to raise cash to keep its massive server farms running, despite making only a fraction of its rivals’ revenue. Now it has another expensive problem. The NAACP is suing them, alleging that xAI’s Memphis data center is illegally polluting Black neighborhoods. Looks like spending billions isn’t Musk’s only toxic habit.
🏖️ Finland just fired up the world’s largest battery, and it runs on…sand. Yes, plain old crushed rock. Piled inside a silo, the sand battery stores enough energy to heat an entire town for a week in winter. It’s slashing carbon emissions and reducing fuel use by 60%. Solving climate change might actually be dirt cheap.
Master AI search results
Adobe LLM Optimizer shows how often tools like ChatGPT and Claude mention your brand. It tracks where you show up in AI-generated answers and how you stack up against competitors.
But it’s more than just a tracker. You get one click recommendations to boost your presence in AI results.
Why it matters: Old school SEO is over. Generative AI is how customers find brands now. Adobe LLM Optimizer ties AI-driven discovery to real business results, not just vanity metrics.
Use it to stay front and center as AI rewrites search. Don’t just get found. Get featured.
⛵ Sorry Larry Ellison, Denmark beat you to autonomous sailing. Silicon Valley based Saildrone is helping the Danes patrol the high seas with solar powered robotic sailboats. Called “Voyagers,” these crewless vessels are packed with AI and surveillance tech to monitor threats amid growing maritime tensions. Wonder how long before sail bots go for the America’s Cup?
🚨 California refuses to wait for DC to regulate AI. A new state-backed report says tech companies aren’t equipped to police themselves and calls for mandatory third-party audits. Meanwhile Congress wants to block state-level AI rules for the next decade. California isn’t buying it. The Wild West days of AI might end right where they began.
⚡ By 2035, chips should be more efficient, right? A new report says no. It claims future AI chips could draw over 15k watts each, about 30x more than today’s high-end GPUs. The reason? New LLMs need way more memory bandwidth and compute. Data centers running these monsters will need serious cooling and massive grid upgrades – maybe even their own power plants. So much for AI solving all our problems.
fun stats
51%
of small businesses are still just testing AI
– 40% of workers used AI on the job in 2024
– 25% of small businesses currently use AI apps
– 24% say they won’t introduce AI anytime soon
[Inc]