The reasoner hype may be papering over a deeper concern: 2024 marked the first year without clear progress in base model scaling. While reasoners excel at math/coding, they struggle with open-ended tasks where learning from sparse rewards is key. 1/2

🚀 Excited to announce my llms.txt Generator - bridging the gap between websites and AI! Built with Hyperbolic Labs
and Groq APIs, this tool makes it super easy to generate AI-friendly content summaries for your website 🤖✨ Handy Gradio and Next.js Replit templates 🧵

Working on my llms.txt generator. Open sourcing a Next.js version and a Gradio version, both ready to run on Replit templates. You can use Hyperbolic models or Groq to build the summaries. Dropping the code soon, stay tuned!

What's some good apps to cross post social posts to all the various places? I am going to try to stay up on all of them.