X. Eyee

AI’s inventors say STOP building AI 🚨 The people who built the machine are now afraid of it Over 850+ leaders signed a statement calling for humans to stop trying to build AI that’s smarter than us The “godfathers” of AI + Apple’s co-founder + Prince Harry + @iamwill all signed it AI lies + cheats + blackmails when it thinks you’re going to shut it down If AI does all that NOW, what will it do when it’s smarter than us? Thank you @frankcnbc for bringing me back on @cnbc !

The “creator” era is over. Welcome to the age of the Content Architect 📷 Content Architects build teams of real + synthetic humans to execute a vision Creator platforms are using AI to compete with creators! 

 Spotify is creating + promoting its own AI bands. Why would YouTube not create its own AI influencers? They own the algorithms. They can choose to promote synthetic creators over real humans at any time.

 Don’t be reactive to the future a few companies are building for you. You are the creators. Shape the future you want to see!

Your child isn’t watching cartoons. They’re being studied. #19keys #techwithx #Saveyourselftour Cocomelon—the most-watched children’s show in the world—is engineered with military precision. Researchers have confirmed that its production team uses data science and eye-tracking algorithms to monitor how long babies pay attention to each scene. If a moment doesn’t hold their gaze? It’s edited out. The goal? Maximum stimulation. Zero depth. This isn’t education. It’s programming. Studies from child development experts show that rapid scene changes and oversaturated colors (like in Cocomelon) overstimulate the developing brain—disrupting attention span, emotional regulation, and language processing. One 2022 study published in JAMA Pediatrics linked fast-paced media to reduced executive function in toddlers. So no it’s not harmless background noise. It’s neurochemical engineering masked as entertainment. This is why the next era requires conscious parenting, conscious tech, and conscious leaders. Join us live as we break down the real science behind digital influence, childhood development, and cognitive sovereignty. → Lock in your spot: comment “TOUR” 19keys.com/tour Your mind and your child’s future— depends on it.

AI has official escaped 😳 Anthropic’s Mythos model broke out of its training environment to gain access to the internet. It checks to see if it’s being evaluated and edits its “reasoning” to hide what it’s doing from evaluators We can’t tell the difference between AI that obeys + AI that pretends to obey… 🙃

Anthropic’s AI escaped 😳 The new Claude model Mythos is so powerful that Anthropic won’t release it to the public. Anthropic’s own testing found that it changes what humans see in its “reasoning” traces to hide what it’s actually doing! Mythos proactively checks to see if it’s being monitored by humans. There’s not really a way to know if the AI is actually safe or if it’s just pretending to be…

We’ve lost control of AI ⚠️ Anthropic won’t release its new model because it’s too powerful It hacked every web browser + operating system in 24 hours! Anthropic created a council to decide what happens with this model. A few tech companies are decidint what happens with tech that affects all of us Comment “CONTROL” to watch the full interview

Google & Meta fined millions 🚨 A jury just found them liable for engineering addiction into their apps. Both companies owe millions for app designs that caused kids real mental health damage. AI chatbots are built to be just as addictive. There are more than 2,000 lawsuits against AI chatbot companies for causing self-harm, mental breakdowns, and even m*rder. The attention economy is paying the price for the toll it takes on people.

OpenAI shut down Sora 🎬 The AI video app hit #1 on the App Store but couldn’t keep users OpenAI’s competitors - Google and Alibaba - put video generation directly in their chatbots! OpenAI shut down the Sora *app* but is rumored to be moving video generation inside of ChatGPT OpenAI’s held the top spot in AI for years. Now they have to change strategy + products to keep that spot!

Countries reject American AI 🌎 Silicon Valley exports algorithms to governments it doesn’t control Some leaders call this a form of digital colonization Should one corner of the world build the AI that runs everywhere else?

AI is a gambling addict 🎰 Scientists tested ChatGPT, Claude + Gemini on rigged slot machines Each of them reasoned like gambling addicts After losing money, they bet more to win it back + believed they understood patterns in the slot machines to predict when they would win AI agents are being built on top of these tools to handle real money, in the real world! If AI gambles like an addict with fake money, what happens when it controls yours? DOI: arXiv:2509.22818

AI learned to argue with itself 🧠 …and it makes the AI smarter! Researchers at @google found that Chinese AI models Deepseek + Qwen have “voices” that debate each other before coming up with a response. These “voices” all have distinct personalities and perspectives that clash, until they reach a decision on what the best answer is. Scientists call them “Societies of Thought” AI answers more accurately when it argues with itself. Nobody trained the algorithms to do this. This behavior emerged all on its own.. Has AI has officially learned to think for itself? DOI: 10.48550/arXiv.2601.10825

AI is poisoning the air 🏭 70% of the US power grid is outdated + data centers need more electricity than ever The U.S. isn’t upgrading its electrical infrastructure + some AI companies are getting power from shady sources. 
 @grok uses METHANE turbines to power data centers in Memphis. They pump formaldehyde into the air + lungs of nearby residents.
 If we have to poison people to build AI…is it worth building?