This data shows small AI teams are building massive empires. Startup revenue per head is hitting all-time highs. And it's cutting across multiple types (powerpoint decks, code writing, voice generation, image generation), though not all types are performant (ex: AI detection).
MULTIMODAL AI
-
ChatGPT Projects Gets Deep Research, Voice Mode, Mobile Updates
By
–
Projects Update We’re adding more capabilities to projects in ChatGPT to help you do more focused work. Deep research support Voice mode support Improved memory to reference past chats in projects Upload files and access the model selector on mobile
-
Instagram’s AI 3D Photo Integration Beta Impresses with Seamless Stereoscopic Conversion
By
–
The beta 3D photo integration with Instagram is very well done! Every static photo becomes an AI generated stereoscopic 3D photo, and there is a “3D” button that lets you toggle the feature on and off for comparison. Every photo I looked at “just worked”, with no glaring
-
Google Announces New Gemini Model Releases
By
–
BREAKING : Google is preparing several new model releases for next week. – Gemini 2.5 Flash Light Preview 06 17
– Gemini 2.5 Pro 06 19 Plus, "prowlridge" model on LM Arena h/t @arrakis_ai -
AVGenL Workshop 2nd Edition at ICCV2025 Announced
By
–
We are excited to announce the 2nd Workshop on AVGenL: Audio-Visual Generation & Learning at #ICCV2025! This year, in addition to an outstanding lineup of speakers, we’ll be featuring industrial demo sessions. Stay tuned—more details are coming soon! https://
goo.su/p8cj -
Imagen 4 Fast Model Performance Costs Comparison
By
–
Same prompt with each Imagen 4 model (4:3) 1. Fast: ~3s, $0.02
2. Regular: ~4s, $0.04
3. Ultra: ~9s, $0.06 Fast still has great prompt adherence and renders text well, it seems very good. I like the raw look here too. -
New Imagen 4 Models Released with Pricing Options
By
–
We've got some new Imagen 4 models for y'all: Imagen 4 Ultra: $0.06/img, the highest quality outputs https://
replicate.com/google/imagen-
4-ultra
… Imagen 4 Fast: $0.02/img, for speed https://
replicate.com/google/imagen-
4-fast
… And we’ve reduced the price of regular Imagen 4 to $0.04/img -
Using Flux Context AI to Colorize and Restore Damaged Vintage Photos
By
–
2. J'obtiens ce prompt que j’envoie à Flux Context avec la photo Colorize this severely damaged vintage photograph of four young girls in traditional or ceremonial clothing. Carefully reconstruct the heavily creased, torn, and missing parts of the image, especially across the
-
Using ChatGPT to Generate Optimized Photo Restoration Prompts
By
–
1. Vous envoyez d'abord l'image à chatgpt ou autre, et lui demandez ça : J’ai une photo ancienne que je souhaite restaurer avec un générateur IA. Peux-tu me rédiger un prompt en anglais, optimisé pour obtenir la meilleure restauration possible ? Sois stratégique et précis.
-
ByteDance Video Generation Advances: Quality Speed Control Improvements
By
–
Tenéis más ejemplos en la web de Bytedance para comparar. Lo que está claro es que los avances en generación de vídeo no tiene pinta de estar parando, mejorando calidad, velocidad, control! https://
seed.bytedance.com/en/seedance