• Created by developers who have had access to most mainstream LLMs for various projects, ArtificialAnalysis is a site where you can compare different AI models and hosting providers to get a better idea of what you need for your project. - Artificial Analysis
• The National Retail Federation’s latest conference was (almost) all about generative AI, with 81% of retail decision-makers wanting to urgently adopt it – retailers like Walmart, Target, or Carrefour are integrating generative AI into customer service, merchandising, marketing, and internal operations. - Modern Retail
• BCG reports that 85% of business leaders say that they will increase their spending on AI and GenAI in 2024, 89% of executives rank it as a top-three tech priority, and 54% expect AI to deliver cost savings this year primarily through productivity gains in operations, customer service, and IT. - BCG
• If you want to convince your boss that AI has value, show them this video of Copilot for Office creating a PowerPoint presentation from an AI-generated Word doc in 47 seconds – apparently this is the most impressive and scary AI capability to managers, according to Ethan Mollick. - X (Twitter)
• Only 23% of wages paid for tasks involving vision are economically viable for AI automation despite significant cost modeling advancements in computer vision, suggesting that AI won’t take your job soon – it will take even longer if computing requirements grow, data becomes scarce, workers are hard to find and costs are too high. But, if AI implementation costs drop rapidly, AI could take over more jobs, faster. - MIT CSAIL
• The modern AI enterprise tech stack has four layers: Compute and foundation models, Data, Deployment, and Observability. We’ve transitioned from model-centric to product-centric AI development, and the biggest current trends include retrieval-augmented generation (RAG), smaller & task-specific models, new tools for observability and evaluation, and a move toward serverless architectures. - Menlo Ventures
