Devin: First AI Software Engineer 🤖 // Google Gemini AI on Elections 🚫 // WorkArena Benchmark 💼

Introducing Devin, the first AI software engineer that can plan and execute complex engineering tasks requiring thousands of decisions. Google's AI chatbot won't answer questions about upcoming elections to prevent inaccurate or misleading responses. WorkArena, a benchmark measuring the ability of large language model-based agents to perform tasks that align with the daily work of knowledge workers using enterprise software systems. Synth$^2$, a novel approach that leverages Large Language Models (LLMs) and image generation models to create synthetic image-text pairs for efficient and effective Visual-Language Model (VLM) training. Contact:  [email protected] Timestamps: 00:34 Introduction 01:59 Introducing Devin, the first AI software engineer 03:48 Google won’t let you use its Gemini AI to answer questions about an upcoming election in your country 05:35 AI Datacenter Energy Dilemma - Race for AI Datacenter Space 06:45 Fake sponsor 09:08 Branch-Train-MiX: Mixing Expert LLMs into a Mixture-of-Experts LLM 10:26 WorkArena: How Capable Are Web Agents at Solving Common Knowledge Work Tasks? 11:52 Synth$^2$: Boosting Visual-Language Models with Synthetic Captions and Image Embeddings 13:38 Outro