Self-supervised learning creates training signals from the data itself instead of relying entirely on human labels. A model may predict masked words, reconstruct a missing image region, identify whether two views came from the same object, or learn the next element in a sequence. These objectives let systems pretrain on enormous collections of text, images, audio, or video and develop representations that can later support specific tasks. The approach reduces annotation needs but does not remove data-quality concerns; the model can absorb bias, private information, duplication, and harmful patterns from its source material. Downstream evaluation is necessary because success on the pretraining objective may not translate evenly across applications or groups.
Google Cloud and Figma have launched an expanded partnership to embed Google’s generative AI models – including Gemini 2.5 Flash and Imagen 4 – directly into Figma’s design tools, reducing latency and accelerating creative workflows.
Opera today introduced Neon, an AI-powered browser that can run tasks and navigate pages autonomously — aiming to shift web browsing from passive to “agentic.”
AI search startup Perplexity has reportedly raised $200 million at a $20 billion valuation, just two months after a $100 million round – signaling its rapid rise as a major rival to Google.