Yes — this also proves it's imagegen, i.e. it didn't just literally generate the whole thing via code. Fixing this was as easy as saying "Fix the numbers in the lower right." I did do this but decided it was more honest to leave the mistake in.
GENERATIVE AI
-
Creating helper images via tools then img2img on result
By
–
From the reasoning trace it seems to be creating "helper images" via tools and then does img2img on the result.
-
ChatGPT Images 2.0: Advanced Visual AI Generator Released
By
–
🚨ChatGPT Images 2.0 Drops: Next-Level Visual AI
— Futurepedia – Learn to Leverage AI (@futurepedia_io) 22 avril 2026
Meet the cutting-edge image generator tackling complex tasks with pinpoint precision. Expect razor-sharp edits, dynamic layouts, and AI smarts that think like a pro – delivering ready-to-use visuals instantly.
Video powered by… pic.twitter.com/HIcFGyLRb6ChatGPT Images 2.0 Drops: Next-Level Visual AI Meet the cutting-edge image generator tackling complex tasks with pinpoint precision. Expect razor-sharp edits, dynamic layouts, and AI smarts that think like a pro – delivering ready-to-use visuals instantly. powered by
-
Latent Reasoning Tuning: LLMs Reasoning Without Full Step Output
By
–
What if LLMs could reason without writing out every single step? Researchers from Harbin Institute of Technology propose a new method called Latent Reasoning Tuning (LRT). Instead of generating long, costly text chains, LRT teaches the model to use compact internal
-
Unauthorized Discord Users Access Anthropic’s Mythos AI Model
By
–
What? Although Mythos was "too powerful for public use" (Anthropic), several Discord users had access to the model from day one! A small group of "unauthorized discord-users" reportedly accessed Anthropic’s powerful Mythos AI model, exploiting a mix of insider access and online
-
Speculative Decoding Optimization for Mixture of Experts Models
By
–
Get more from speculative decoding in MoE models
-
Local AI Models Cannot Match Leading Labs Models
By
–
Nah, never found a single local model that can replace the best labs’ models.
-
Self-Improving AI Agent Live 24/7 Native Interface
By
–
Nice: One click and your self-improving AI agent is live 24/7 – main tagline
— Chubby♨️ (@kimmonismus) 22 avril 2026
Full Hermes power, zero terminal
43,200+ stars Hermes agent, now in a native interface
Your 24/7 coworker, live! :)) https://t.co/j1yJHMRzeaNice: One click and your self-improving AI agent is live 24/7 – main tagline
Full Hermes power, zero terminal 43,200+ stars Hermes agent, now in a native interface Your 24/7 coworker, live! :))