Image Generation
Jun 14, 2026

The Gist
Google DeepMind released DiffusionGemma, a new AI model that makes text generation four times faster when running on your own computer. Hollywood filmmakers are moving beyond basic AI video tools to create custom models for actual entertainment projects. Meanwhile, general-purpose AI chatbots are proving better at medical tasks than specialized medical AI systems.
Today's Stories
- 1
Google DeepMind releases DiffusionGemma to make local AI text generation four times faster
Google DeepMind launched DiffusionGemma, a new AI model that speeds up text generation by 400% when running on personal computers without internet connection. The technology uses diffusion methods (the same technique behind image generators like DALL-E) to create text instead of images. This represents a shift from the current approach where most AI writing happens on company servers.
People using AI writing tools on their laptops or phones could see much faster responses without needing constant internet connection.
- 2
Hollywood starts building custom AI models beyond basic video generators
Filmmakers at Tribeca 2026 showcased projects using custom-built versions of Google's Veo and Imagen models rather than standard AI video tools. The project "Dear Upstairs Neighbors" demonstrates how entertainment companies are training AI specifically for their creative needs instead of relying on generic video generators. This marks a shift from the current crop of short, inconsistent AI videos toward professional entertainment content.
Future movies and TV shows may increasingly use AI that's been specifically trained for storytelling rather than the basic AI video tools available to consumers today.
- 3
General AI chatbots outperform specialized medical AI in clinical tasks
A new study published in Nature Medicine found that large language models (AI that understands and generates text like ChatGPT) perform better at medical tasks than AI systems designed specifically for healthcare. The research challenges the assumption that specialized AI tools are always superior to general-purpose ones in professional fields.
Doctors and medical professionals might get better results using regular AI assistants like ChatGPT for certain tasks rather than expensive specialized medical software.
- 4
OVH Group enters negotiations to acquire voice AI company Gladia
European cloud computing company OVH Group announced exclusive negotiations to acquire Gladia, a French startup specializing in voice AI technology. Gladia builds tools that convert speech to text and analyze conversations for businesses. The acquisition would expand OVH's AI capabilities in the growing voice technology market.
European businesses may soon have access to more advanced voice AI tools integrated with their existing cloud services.
What to Watch
Keep an eye on how quickly DiffusionGemma gets integrated into popular writing apps and whether other tech companies follow Google's approach to faster local AI. The Hollywood AI projects from Tribeca may signal which entertainment companies will release AI-enhanced content first.
Sources
- General-purpose large language models outperform specialized clinical AI
- The Future of Work and AI
- Transcribing my old podcast locally with open-source AI
- The Day the US Government Shut Down the Most Powerful AI
- The future of Hollywood isn’t feeding prompts into vanilla gen AI models
- Validates AI
- OVH Grp enters into exclusive negotiations to acquire Gladia, expert in voice AI
- Debatable: Government Stakes in AI
- The Normalization of Deviance in AI
- Google DeepMind releases DiffusionGemma, a model that runs local AI 4x faster | Diffusion AI is most common in image generation, but it can make text outputs much faster
Share this with a friend
Send today's roundup to anyone who wants to keep up.
Get daily AI news free with AIToday
200+ AI sources, summarized in 1 minute. Email / LINE / Slack.
Sign up free