January 2024 AI Recap

1. Amazon presents Diffuse to Choose, a diffusion-based image-conditioned inpainting model that allows users to virtually place any e-commerce item in any setting, ensuring detailed, semantically coherent blending with realistic lighting and shadows. Code and demo will be released soon [Details].
2. OpenAI announced two new embedding models, new GPT-4 Turbo and moderation models, new API usage management tools, and lower pricing on GPT-3.5 Turbo. The updated GPT-4 Turbo preview model reduces cases of “laziness” where the model doesn’t complete a task. The new embedding models include a smaller and highly efficient text-embedding-3-small model, and a larger and more powerful text-embedding-3-large model. [Details].
3. Hugging Face and Google partner to support developers building AI applications [Details].
4. Adept introduced Adept Fuyu-Heavy, a new multimodal model designed specifically for digital agents. Fuyu-Heavy scores higher on the MMMU benchmark than Gemini Pro [Details].
5. Fireworks.ai has open-sourced FireLLaVA, a LLaVA multi-modality model trained on OSS LLM generated instruction following data, with a commercially permissive license. Firewroks.ai is also providing both the completions API and chat completions API to devlopers [Details].
6. 01.AI released Yi Vision Language (Yi-VL) model, an open-source, multimodal version of the Yi Large Language Model (LLM) series, enabling content comprehension, recognition, and multi-round conversations about images. Yi-VL adopts the LLaVA architecture and is free for commercial use. Yi-VL-34B is the first open-source 34B vision language model worldwide [Details].
7. Tencent AI Lab introduced WebVoyager, an innovative Large Multimodal Model (LMM) powered web agent that can complete user instructions end-to-end by interacting with real-world websites [Paper].
8. Prophetic introduced MORPHEUS-1, a multi-modal generative ultrasonic transformer model designed to induce and stabilize lucid dreams from brain states. Instead of generating words, Morpheus-1 generates ultrasonic holograms for neurostimulation to bring one to a lucid state [Details].
9. Google Research presented Lumiere – a space-time video diffusion model for text-to-video, image-to-video, stylized generation, inpainting and cinemagraphs [Details].
10. TikTok released Depth Anything, an image-based depth estimation method trained on 1.5M labeled images and 62M+ unlabeled images jointly [Details].
11. Nightshade, the free tool that ‘poisons’ AI models, is now available for artists to use [Details].
12. Stability AI released Stable LM 2 1.6B, 1.6 billion parameter small language model trained on multilingual data in English, Spanish, German, Italian, French, Portuguese, and Dutch. Stable LM 2 1.6B can be used now both commercially and non-commercially with a Stability AI Membership [Details].
13. Etsy launched ‘Gift Mode,’ an AI-powered feature designed to match users with tailored gift ideas based on specific preferences [Details].
14. Google DeepMind presented AutoRT, a framework that uses foundation models to scale up the deployment of operational robots in completely unseen scenarios with minimal human supervision. In AutoRT, a VLM describes the scene, an LLM generates robot goals and filters for affordance and safety, then routes execution to policies [Details].
15. Google Chrome gains AI features, including a writing helper, theme creator, and tab organizer [Details].
16. Tencent AI Lab released VideoCrafter2 for high quality text-to-video generation, featuring major improvements in visual quality, motion and concept Composition compared to VideoCrafter1 [Details | Demo]
17. Google opens beta access to the conversational experience, a new chat-based feature in Google Ads, for English language advertisers in the U.S. & U.K. It will let advertisers create optimized Search campaigns from their website URL by generating relevant ad content, including creatives and keywords [Details].
18. If AI models violate copyright, US federal courts could order them to be destroyed
    Under copyright law, courts do have the power to issue destruction orders. Copyright law has never been used to destroy AI models specifically, but the law has been increasingly open to the idea of targeting AI. It's probably not going to happen to OpenAI but might possibly happen to other generative AI models in the future.
19. Microsoft, Amazon and Google face FTC inquiry over AI deals LINK

• The FTC is investigating investments by big tech companies like Microsoft, Amazon, and Alphabet into AI firms OpenAI and Anthropic to assess their impact on competition in generative AI.
• The FTC's inquiry focuses on how these investments influence the competitive dynamics, product releases, and oversight within the AI sector, requesting detailed information from the involved companies.
• Microsoft, Amazon, and Google have made significant investments in OpenAI and Anthropic, establishing partnerships that potentially affect market share, competition, and innovation in artificial intelligence.

20. OpenAI cures GPT-4 ‘laziness’ with new updates LINK

• OpenAI updated GPT-4 Turbo to more thoroughly complete tasks like code generation, aiming to reduce its 'laziness' in task completion.
• GPT-4 Turbo, distinct from the widely used GPT-4, benefits from data up to April 2023, while standard GPT-4 uses data until September 2021.
• Future updates for GPT-4 Turbo will include general availability with vision capabilities and the launch of more efficient AI models, such as embeddings to enhance content relationship understanding.