Stefan Köhler

I was born in 1961. Nobody had food allergies. There were no TRANS kids. I could ride public transportation without being set on fire. I could go to school without getting jumped. We played outside until the streetlights came on. Cartoons didnt try to turn us gay. Movies and TV shows didn't try to depress us about our lives. Families didn't break up over politics. You could say "Merry Christmas" without starting a war. A man was a man, a woman was a woman, and nobody pretended otherwise. Nobody lost their job or their home for merely stating biological facts. We respected cops, teachers, and the flag. Nobody was "offended" by EVERYTHING. And if anybody WAS offended by anything, it was THEIR PERSONAL PROBLEM, not a national crisis. Call me old-fashioned and outdated - but I really miss those days. I'd give anything to go back to BEFORE 24/7 cable news, social media doom-scrolling, and algorithm rage machines. Because tech turned families into enemies, politics into war, and every keyboard warrior into a lie-spreading narcissist. So Santa, if you're following me: All I want for Christmas is a global time machine. So we can get back to the glorious days of real community, actual news, personal responsibility - AND SANITY.

Yann LeCun's new interview - explains why LLMs are so limited in terms of real-world intelligence. Says the biggest LLM is trained on about 30 trillion words, which is roughly 10 to the power 14 bytes of text. That sounds huge, but a 4 year old who has been awake about 16,000 hours has also taken in about 10 to the power 14 bytes through the eyes alone. So a small child has already seen as much raw data as the largest LLM has read. But the child’s data is visual, continuous, noisy, and tied to actions: gravity, objects falling, hands grabbing, people moving, cause and effect. From this, the child builds an internal “world model” and intuitive physics, and can learn new tasks like loading a dishwasher from a handful of demonstrations. LLMs only see disconnected text and are trained just to predict the next token. So they get very good at symbol patterns, exams, and code, but they lack grounded physical understanding, real common sense, and efficient learning from a few messy real-world experiences. --- From 'Pioneer Works' YT channel (link in comment)