This week has seen several notable breakthroughs and announcements in the field of AI:
OpenAI’s Developer Conference and New Products: OpenAI introduced an enhanced version of GPT-4, new text-to-speech models, and an API for DALL-E 3. They also launched the GPT Store, a marketplace for developers to publish and potentially monetize their AI systems based on OpenAI’s models.
Samsung’s Generative AI Family – Samsung Gauss: Samsung unveiled its generative AI family, Samsung Gauss, which includes a large language model similar to ChatGPT, a code-generating model, and an image generation and editing model. Currently used internally, it will soon be available to public users.
Microsoft’s AI Compute for Startups: Microsoft announced a no-cost Azure AI infrastructure option for startups, including “high-end” Nvidia-based GPU virtual machine clusters, to train and run generative models. This update is part of the Microsoft for Startups Founders Hub program.
YouTube’s Generative AI Features: YouTube plans to experiment with generative AI features, including a conversational tool for content recommendation and summarization of video comments.
Kai-Fu Lee’s AI Startup – 01.AI: Kai-Fu Lee’s startup 01.AI, valued at $1 billion, released its first model, the open-source Yi-34B, just seven months after its founding.
GitHub’s Customizable Copilot Plan: GitHub announced an enterprise subscription tier allowing companies to tailor Copilot based on their internal codebase.
Hugging Face’s Two-Person Model Team: The AI startup Hugging Face has been making significant strides with tools developed by its two-person team, H4.
Mozilla’s AI Chatbot – Fakespot Chat: Mozilla launched Fakespot Chat, an AI agent for online shopping assistance, following its acquisition of Fakespot.
Machine Learning in Short-Term Predictions: Researchers are using machine learning for short-term predictions in complex data structures, such as extending warning periods for earthquakes and predicting weather patterns with models like MetNet-3.
AI-Generated Text Detector for Chemistry Articles: University of Kansas researchers developed a detector for AI-generated text, specifically for chemistry journal articles.
AI in College Admission Essays Analysis: Researchers at Colorado University and UPenn demonstrated that AI could identify qualities in student essays, potentially aiding admissions officers.
AI’s Limitations in Accessibility: A study at the University of Washington highlighted mixed experiences with AI tools’ accessibility, particularly for people with disabilities.
AI Assistance for People on the Autism Spectrum: A person on the autism spectrum found that using a language model to generate messages on Slack helped in overcoming communication challenges.
Governing Principles for AI in Seattle: Seattle released a set of governing principles and toolkits to regulate AI use in official duties.
Interactive VR Experiences with Machine Learning: Machine learning models are being used to create innovative VR interactions, enabling more intuitive and efficient virtual object handling.
Meta’s Generative AI Research Milestones – Emu Video and Emu Edit: Meta announced Emu Video for high-quality text-to-video generation and Emu Edit for precise image editing, both leveraging generative AI technologies.
Microsoft Ignite 2023 Announcements: Microsoft’s Copilot for Microsoft 365 has shown substantial productivity gains. The company also announced Azure Maia, an AI Accelerator chip, Azure Cobalt, a cloud-native chip, and Azure Boost for enhanced storage and networking.
These advancements demonstrate the dynamic and multifaceted growth of AI technology across various sectors.
