Choose how you want to continue.
An LLM that's 7500× stupider
Summary
The post humorously critiques a newly released large language model (Kimi K2.5) with 1 trillion parameters, contrasting it with an older model (smollm2:135m) that is described as '7500× stupider'. The author details amusing interactions with the smaller model, showcasing its incorrect yet comical responses, and reflects on the overall utility and entertainment value of LLMs in software development.
MORE POSTS LIKE THIS
Meta to release new AI model with advanced coding capabilities ‘soon’
Siliconangle ·
Jul 3, 2026
The “back and forth” I'm having with an LLM over a non-issue it thinks is a bug
Boston Conman ·
May 26, 2026
How Small Language Models Are Key to Scalable Agentic AI
NVIDIA Corporation ·
Aug 29, 2025
Introducing Claude 4 in Amazon Bedrock, the most powerful models for coding from Anthropic
Amazon Web Services ·
May 22, 2025
AUTHOR
Sponsored
Organized team chat for people who take work seriously. Topic-based
threading keeps conversations focused.
Try Zulip