Jeremy Howard leads Answer.AI, an AI R&D lab building practical applications from cutting-edge research. He focuses on AI, open-source tools, and educational platforms to help developers and researchers solve real-world problems efficiently.
18 articles from this blog
Introducing Cachy, an open-source Python package that caches LLM API calls to speed up development, testing, and clean up notebook diffs.
Introducing FastStripe, a cleaner, more efficient Python SDK for the Stripe API that simplifies payment integration with better IDE support and fewer lines of code.
Introduces fastmigrate, a simple Python tool for managing SQLite database migrations without requiring a specific ORM.
A technical exploration of flexicache, a Python caching decorator from fastcore, covering its LRU cache and time/file-based invalidation policies.
Introduces ReadBench, a benchmark for evaluating how well Vision-Language Models (VLMs) can read and extract information from images of text.
An AI researcher shares her journey into GPU programming and introduces WebGPU Puzzles, a browser-based tool for learning GPU fundamentals from scratch.
Explores using a masked language model's head for zero-shot tasks, achieving strong results without task-specific heads.
MonsterUI simplifies web UI development by providing pre-styled Python components for FastHTML, eliminating complex CSS and boilerplate.
A real-world evaluation of Devin, an AI software engineer, testing its capabilities on practical development tasks after its high-profile launch.
Introducing ModernBERT, a new family of state-of-the-art encoder models designed as a faster, more efficient replacement for the widely-used BERT.
Introducing nbsanity, a free service that instantly transforms public GitHub Jupyter notebooks into polished, customizable web pages using Quarto.
A guide on using ShellSage with iTerm2's tmux control mode for seamless AI-powered terminal assistance without learning tmux commands.
Introducing ShellSage, an AI-powered terminal assistant that provides context-aware help for sysadmin tasks directly in your shell.
Strategies for building an audience as a technical writer, focusing on authentic engagement and consistent content creation.
fast.ai announces a new 'AI-first' course, 'How To Solve It With Code,' and its integration into Answer.AI, focusing on Dialog Engineering for collaborative coding.
Explores the use of LLMs to generate synthetic data for training AI models, discussing challenges, an experiment with coding data, and a new library.
Explores methods for implementing background task processing and concurrency in Starlette-based web applications like FastAPI to avoid blocking on LLM API calls.
Introduces rerankers, a lightweight Python library providing a unified interface for various document re-ranking models used in information retrieval pipelines.
