Large Language Models
My latests articles
Fine tuning LLMs
Fine-tuning allows to adapt generalist models into specialists, but is it always the best approach?
LLM quantization
Quantization is a model compression technique that reduces the size and computational requirements of LLMs.
Understand LLM benchmarks
A practical guide to finally understanding most popular LLM benchmarks
Base and instruction-tuned models
What is the difference between base and instruction-tuned models?
Understand parameters in LLM
Parameter is a key concept in LLMs. This article explains the difference between total and activated parameters.
Web development and AI news
Web development
Bun's llms.txt File for LLM Integration
Bun includes a file (llms-full.txt) containing its documentation in a format suitable for training or querying Large Language Models (LLMs), facilitating AI-assisted development with Bun.
Discussion Around Potential Chrome Sale and OpenAI's Interest
Reports and commentary on the possibility of Google being required to sell its Chrome browser, including OpenAI's reported interest and opinions on the potential impact on the web.
Frimousse: Lightweight, Composable React Emoji Picker
A React component for selecting emojis, highlighted for its lightweight nature, lack of styling, composability, accessibility, and support for device-specific emojis.
React Compiler Release Candidate
The React team just published React Compiler RC, in preparation of the compiler’s stable release
RedwoodJS becomes RedwoodSDK
RedwoodJS evolved into RedwoodSDK - a new React framework built for Cloudflare that begins as a Vite plugin and gives you SSR, React Server Components, server functions, and realtime features.
Artificial intelligence
Baidu Launches Low-Cost ERNIE Turbo AI Models
Baidu introduced ERNIE 4.5 Turbo and ERNIE X1 Turbo models, offering upgraded multimodality, faster responses, and strong reasoning at significantly lower costs compared to competitors like GPT-4.5 and DeepSeek R1, aiming to challenge the economics of frontier AI.
xAI Updates Grok with Algorithm, Vision, and Memory Features
Elon Musk's xAI is enhancing its Grok chatbot with significant updates, including a much-improved algorithm, a new Vision feature allowing interaction via camera, and a memory recall capability, aiming to improve its performance and multimodal interaction.
Alibaba Launches Qwen3 Open-Source AI Models
Alibaba released Qwen3, a family of eight open-weight language models under Apache 2.0 license, with the flagship Qwen3-235B model rivaling top models like OpenAI's o1 and Grok-3 on benchmarks. The models support 119 languages and show strong performance in coding and tool use.
Firefly's Image Model 4
The latest release of Firefly unifies AI-powered tools for image, video, audio, and vector generation into a single, cohesive platform and introduces many new capabilities.
Google Launches Gemini 2.5 Flash and Gemma 3 QAT
Google introduced Gemini 2.5 Flash, a fast, cost-efficient model with controllable reasoning, and Gemma 3 QAT, which enables powerful 27B models to run on consumer GPUs, expanding accessibility for developers building AI applications.