I try my best to unpack one of these papers a week like this.
Things keep accelerating so keeping up is difficult, but that's half the reason I created this place, to create surface area to map out the new branches of cognition and technology we are building and exploring.
So dense yet easy to understand. Beautifully written as well, I can sense childlike wonder and admiration in this blog. It almost feels like you found beauty.
I’d love the opportunity to apprentice with you—no pay needed! I’m sharp, fast to learn, and ready to dive in headfirst. I worked at a U.S.-based digital marketing agency (remotely from India), saved up, and left when I realized AI would dominate the field. I’m too young to bet my future on it. Here is my LinkedIn Profile - https://www.linkedin.com/in/prashantmankani/
Today, i learnt even LLMs can learn like humans about practically anything. By developing logic and reasoning "by doing things" is the way to go, can't wait to see what the chads at DeepSeek build upon these models!
Thanks for distilling!
Thank you for reading!
I try my best to unpack one of these papers a week like this.
Things keep accelerating so keeping up is difficult, but that's half the reason I created this place, to create surface area to map out the new branches of cognition and technology we are building and exploring.
... And using to get to the next dimension.
What a time to be alive.
So dense yet easy to understand. Beautifully written as well, I can sense childlike wonder and admiration in this blog. It almost feels like you found beauty.
(I did) 😁
I’d love the opportunity to apprentice with you—no pay needed! I’m sharp, fast to learn, and ready to dive in headfirst. I worked at a U.S.-based digital marketing agency (remotely from India), saved up, and left when I realized AI would dominate the field. I’m too young to bet my future on it. Here is my LinkedIn Profile - https://www.linkedin.com/in/prashantmankani/
Today, i learnt even LLMs can learn like humans about practically anything. By developing logic and reasoning "by doing things" is the way to go, can't wait to see what the chads at DeepSeek build upon these models!
Thanks!