Imagen 4 is available at https://
poe.com/Imagen-4 and across all platforms. (2/2)
MULTIMODAL AI
-
Imagen 4 Now Available on Poe and All Platforms
By
–
-
Google DeepMind Launches Imagen 4 Text-to-Image Model
By
–
Now on Poe: Imagen 4! Imagen 4 is the newest text-to-image model from Google DeepMind. It produces realistic images with sharpened clarity and impressive speed. (1/2)
-
AI Image Generation Handles Upside Down Perspectives Unexpectedly Well
By
–
I didn't expect this one to just work.
— fofr (@fofrAI) 23 mai 2025
> an upside down woman is saying "hello, I seem to be upside down"
Someone should help her down. pic.twitter.com/ntjSYhVEuxI didn't expect this one to just work. > an upside down woman is saying "hello, I seem to be upside down" Someone should help her down.
-
Gen-4 AI References for Building and Interior Design
By
–
Use Gen-4 References to design buildings. And their interiors.
-
Scale Any Shoot with AI: Text-to-Image and Gen-4 Production
By
–
Turn any shoot into a large scale production by combining Text to Image, References, Gen-4 and some simple tracking and color grading. pic.twitter.com/bSXXD4r0li
— Runway (@runwayml) 23 mai 2025Turn any shoot into a large scale production by combining Text to Image, References, Gen-4 and some simple tracking and color grading.
-
Flow Still Using Veo2 Video Generation Model
By
–
I could be wrong but I think Flow is still using Veo2.
-
MIT Study Reveals Vision-Language Models Fail with Negation in Medical Imaging
By
–
MIT study reveals that vision-language models, which are often used to analyze medical images to streamline medical diagnosis, can’t handle queries w/negation words like "no" and "not": https://
bit.ly/4dDo9Hj -
ChatGPT gains molecule analysis capabilities via RDKit library
By
–
ChatGPT now can analyze, manipulate, and visualize molecules and chemical information via the RDKit library. Useful for scientific work across health, biology, and chemistry.
-
AI Models Today: Specialized Tools for Every Task
By
–
AI is getting wild—models today can already handle almost anything: o3: expert planner
Sonnet: tool use & code
Gemini Flash: fast & cheap
Qwen-235b: open-source beast
GPT-4.1: daily tasks
Grok-3: real-time info
Kling 2.0: video
ChatGPT image-gen: visuals You can try all of them -
Diffusion Models: Generation, Guidance, and Flow Matching Techniques
By
–
I’ve been away this week, but next week we re-start the diffusion lectures! We’ll be doing generation, classifier (free) guidance, flow matching with ODEs, and some pretty cool math What an amazing week for diffusion: text diffusion, jaw dropping video models, high fidelity