This is the first (small) controlled study I have seen of GenAI on industrial quality control. Here, engineers commissioning new trains took part in an experiment using a GPT-3.5 powered troubleshooting system. Those who used the chatbot had significant increases in work quality
GENERATIVE AI
-
LLMs Learn to Write With Typos and Errors
By
–
Because LLMs could never, ever learn to rite with typos
-
RunwayML’s instantaneous inpainting
By
–
RunwayML’s instantaneous inpainting pic.twitter.com/0FQ6zSEoey
— AI Breakfast (@AiBreakfast) 25 juillet 2025RunwayML’s instantaneous inpainting
-
Upstage Document Parse: AI-Powered PDF Extraction for LLM Training
By
–
@upstageai document Parse extracts structured text from PDFs/DOCs with top performance. This is essential for tasks like LLM training or RAG. https://
console.upstage.ai/docs/capabilit
ies/document-digitization/document-parsing
… It accurately recognizes tables and charts, converting them into structured text. Try it—it’s excellent. -
Gemini Achieves Gold Medal in Math Olympiad, Launches 2.5 Flash-Lite
By
–
ICYMI here’s what shipped this week —Gemini achieved gold-medal standard in the International Mathematical Olympiad
—Gemini 2.5 Flash-Lite is stable and generally available for developers and enterprise customers
—You can now turn photos into videos in @GooglePhotos and -
Upstage Document Parse Extracts Structured Text from PDFs
By
–
@upstageai 's document Parse extracts structured text from PDFs/DOCs with top performance. This is essential for tasks like LLM training or RAG. https://
upstage.ai/products/docum
ent-parse
… It accurately recognizes tables and charts, converting them into structured text. Try it—it’s excellent. -
AI Draws Pelican With Visible Thinking Trace
By
–
Here's the pelican it drew me, and a snippet of the thinking trace behind it
-
Qwen’s New Thinking Model Takes 166 Seconds to Process
By
–
Qwen released their updated "thinking" model today. It thinks really hard! Took 166 seconds to think through the details of drawing me a pelican on a bicycle. The finished drawing wasn't great but the thoughts behind it were fun to see. https://
simonwillison.net/2025/Jul/25/qw
en3-235b-a22b-thinking-2507/
… -
Models Making Multiple Tool Calls Single Conversation
By
–
It's really the number of tool calls you can make a single conversation Many models can do one just fine, but I want to be able to do a dozen or more
-
GPT 4.1, o3, o4-mini and Claude 4 comparison feedback
By
–
I have been happy with GPT 4.1 and o3 and o4-mini for this in addition to Claude 4