@patio11: Sometime in the last ~two model releases from the big labs they went from “this would be acceptable output from the med…
Summary
Patrick McKenzie observes that recent AI model releases from major labs have seen a dramatic improvement in output quality, from average coworker level to impressively good.
View Cached Full Text
Cached at: 09/29/26, 11:53 AM
Sometime in the last ~two model releases from the big labs they went from “this would be acceptable output from the median low-seniority coworker” to “this is frighteningly good.”
I think people who are not daily users of the models are unlikely to grok that, so, saying it.
Similar Articles
Video models are getting good
The article discusses the increasing capabilities and improvements in AI video models, highlighting their growing effectiveness in generating or processing video content.
I stopped comparing models months ago. My output improved .
The author shares that they stopped comparing AI models and focused on workflow design, leading to improved output. They argue that workflow has more leverage than model choice for most practical use cases.
@rohanpaul_ai: Arcads' Actors+ is a good example of why AI video quality is becoming a perceptual problem now. You can keep improving …
Arcads' Actors+ is an AI actor library that makes AI-generated videos nearly indistinguishable from real user content, highlighting the perceptual challenge in advancing AI video quality.
AI models got way more reliable but the AI labs forgot to make them smarter (SciCode)
The article argues that AI models have become more reliable but intelligence gains have plateaued, suggesting older models like GPT-4 turbo could rival newer ones with proper enhancements.
@mattpocockuk: Uncle Bob gets it
Robert C. Martin (Uncle Bob) shares his experience that using AI agents significantly boosts productivity, though 30-40% of time is spent tuning and hardening the output.