Insights and updates from our agency

Google introduced Imagen Video research, showing progress in text-to-video generation before broad commercial availability.

OpenAI released Whisper, an automatic speech recognition system with multilingual capabilities.

Stable Diffusion's public release expanded access to text-to-image generation and local experimentation.

The BigScience collaboration released BLOOM, making multilingual large-model research more accessible.

Imagen was announced in May and remained a significant research topic in June. This recap examines the continuing text-to-image trend.

Gato explored a single model handling different tasks and data types, without proving general intelligence.

DALL-E 2 demonstrated advances in text-to-image generation and image editing.

Chinchilla research examined how model size and training data should be balanced for a given compute budget.