What if you could supercharge LLM training by dynamically optimizing the data, not just the model? Researchers from Peking University, Shanghai AI Lab, & the LLaMA-Factory team present DataFlex. It's a unified framework that smartly selects, mixes, and re-weights training data
GENERATIVE AI
-
Sakana Chat Beta Test User Feedback and Appreciation
By
–
I'm so glad the mug arrived safely! And I'm deeply touched that you have Sakana Chat open in the background Thank you so much for your tremendous cooperation with the beta test!
-
ChatGPT Images 2.0 vs Nano Banana: stress test results
By
–
Hooked on yesterday’s Images 2.0 announcement? We just put it head‑to‑head with our long‑time champ Nano Banana… and the results are wild. In our latest video, Kevin of futurepedia stress‑tests ChatGPT Images 2.0 on: Photoreal prompts (one word – “photorealism” –
-
OpenAI, Anthropic, ThoughtWorks define harness for agent engineering
By
–
If you're not the model, you're the harness. OpenAI, Anthropic, and ThoughtWorks have each landed on very different answers for what a harness should look like, and the gap between their approaches says a lot about where agent engineering is heading. Here's a full breakdown.
-
OpenAI Launches ChatGPT for Clinicians and HealthBench Professional
By
–
Today @OpenAI introduced ChatGPT for Clinicians, provided free for credentialed HCPs, and HealthBench Professional for benchmarking LLM medical task performance (Figure) https://
openai.com/index/making-c
hatgpt-better-for-clinicians/
… https://
cdn.openai.com/dd128428-0184-
4e25-b155-3a7686c7d744/HealthBench-Professional.pdf
… -
Why SVGs Are Powerful for AI-Generated Interfaces
By
–
Why SVGs are Powerful Creating an entire interface as an SVG with Deep Think mode highlights its superpowers: • Built from code: SVGs are just text. This means AI models like Gemini can easily "write" the code to build complex shapes and interfaces from scratch. • Easy to
-
Creating Complex SVG Animations with Gemini Deep Think
By
–
What's an SVG? This entire interface is an SVG made with Deep Think mode in Gemini. 🤯 (For those wondering, SVGs are code-based graphics that stay perfectly crisp no matter how far you zoom in!)
— Harold Sinnott 📱 (@HaroldSinnott) 23 avril 2026
See how to make your own complex animations 👇@GeminiApp https://t.co/m431CXG8Qm pic.twitter.com/s0CGPH32xIWhat's an SVG? This entire interface is an SVG made with Deep Think mode in Gemini. (For those wondering, SVGs are code-based graphics that stay perfectly crisp no matter how far you zoom in!)
See how to make your own complex animations @GeminiApp -
Spud Launch Event Expected to Change Timelines Forever
By
–
Foreshadowing Spud. Hype is real. Get ready friend: today our timelines will be changed forever
-
Images 2.0 Pro one-try success vs ten Thinking failures
By
–
I know but that was before the Images 2.0 release. The image at the top of this thread was made using Pro in one try, and before I had Pro I did ~10 tries with Thinking and they all had unacceptable errors. I’m pretty sure it’s different.
