While companies readily use models that process images, audio, and text inputs, the practical application of generating multimodal outputs (like video or complex graphics) remains rare in business. The primary output is still text or structured data, with synthesized speech being the main exception.

The future of personalization may involve a two-step process. A centralized AI (like Criteo's) will provide strong recommendations. Then, a smaller, privacy-centric model running locally on the user's device (e.g., in their glasses) will perform the final, hyper-personalized adjustments, keeping the most sensitive data private.

Google's Embedding 2 model is a significant infrastructure upgrade because it is 'natively multimodal.' This allows AI to directly understand and retrieve images, diagrams, and text without first converting non-text data into lossy captions. This makes internal knowledge bases and co-pilots dramatically more effective and accurate for enterprises.

To move beyond keyword search in their media archive, Tim McLear's system generates two vector embeddings for each asset: one from the image thumbnail and another from its AI-generated text description. Fusing these enables a powerful semantic search that understands visual similarity and conceptual relationships, not just exact text matches.

The current focus on LLMs is a temporary phase. The true leap towards AGI will come from multi-sensory models that can process and integrate visual, auditory, and other data streams simultaneously, much like a human does. This moves AI from text generation to real-world understanding.

The future of creative AI is moving beyond simple text-to-X prompts. Labs are working to merge text, image, and video models into a single "mega-model" that can accept any combination of inputs (e.g., a video plus text) to generate a complex, edited output, unlocking new paradigms for design.

While today's focus is on text-based LLMs, the true, defensible AI battleground will be in complex modalities like video. Generating video requires multiple interacting models and unique architectures, creating far greater potential for differentiation and a wider competitive moat than text-based interfaces, which will become commoditized.

Standard Retrieval-Augmented Generation (RAG) systems often fail because they treat complex documents as pure text, missing crucial context within charts, tables, and layouts. The solution is to use vision language models for embedding and re-ranking, making visual and structural elements directly retrievable and improving accuracy.

The common belief that AI can't truly understand human wants is debunked by existing technology. Adam D'Angelo points out that recommender systems on platforms like Instagram and Quora are already far better than any individual human at predicting what a user will find engaging.