OpenAI presented CLIP (text–image contrast) and DALL·E (text-to-image generation), catalyzing the multimodal wave.
In early 2021, OpenAI revealed CLIP—contrastive training on web-scale image–text pairs—and DALL·E—a transformer generating images from captions.
CLIP became a ubiquitous encoder; DALL·E popularized prompt art. Competitors (Midjourney, Stable Diffusion) democratized access.
Multimodality moved from paper to daily product, with new questions on training-data copyright and deepfakes.
OpenAI launched GPT-4 with notable gains on professional exams, robustness, and later visual ability.
Vision in chatbots enabled reasoning over diagrams, screenshots, and scanned documents in one thread.
Frontier systems combined diagrams, PDF papers, and code in one tool-verified reasoning thread.