IBM has released Granite Embedding Multilingual R2, an Apache 2.0 licensed multilingual embedding model with a 32K context window. This model achieves state-of-the-art retrieval quality among models under 100 million parameters. It is designed for efficient and accurate information retrieval across multiple languages.
→ This sub-100M multilingual embedding model with a huge context window is a strong contender for RAG applications, especially for those looking for open-source, performant alternatives to larger models.
Simon Willison released inaturalist-clumper 0.1, a tool he uses to publish his iNaturalist sightings to his blog. This release follows several weeks of production use and iterations on its functionality. An example of the output is available as a JSON file.
→ This is a low-signal item, but it's a good example of how developers are building custom tools for data integration and publishing, which could inspire similar RAG-related projects.
Simon Willison released inaturalist-clumper 0.1, a tool he uses to publish his iNaturalist sightings to his blog. This release follows several weeks of production use and iterations on its functionality. An example of the output is available as a JSON file.
→ This is a low-signal item, but it's a good example of how developers are building custom tools for data integration and publishing, which could inspire similar RAG-related projects.
Simon Willison released datasette-llm-limits 0.1a0, a new plugin for Datasette. This plugin integrates with datasette-llm and datasette-llm-accountant to enable per-user or global spending limits for LLM usage within Datasette. It allows configuration of limits such as a rolling 24-hour window with a specified USD amount.
→ This is a practical tool for managing LLM costs in Datasette, especially useful for controlling spending in multi-user environments or when experimenting with various local LLMs.
Simon Willison released datasette-ip-rate-limit 0.1a0, a new plugin for Datasette. This plugin was developed with the help of Codex (GPT-5.5 xhigh) to address issues with poorly-behaved crawlers. It allows for configurable rate limiting based on IP addresses, with options for exempt paths and specific rules for different site areas.
→ This is a great example of using an LLM (even a non-local one) for practical, targeted development to solve real-world infrastructure problems, which is highly relevant for tooling and RAG improvements.
IBM has released Granite Embedding Multilingual R2, an Apache 2.0 licensed multilingual embedding model with a 32K context window. This model achieves state-of-the-art retrieval quality among models under 100 million parameters. It is designed for efficient and accurate information retrieval across multiple languages.
→ This sub-100M multilingual embedding model with a huge context window is a strong contender for RAG applications, especially for those looking for open-source, performant alternatives to larger models.
IBM has released Granite Embedding Multilingual R2, an Apache 2.0 licensed multilingual embedding model with a 32K context window. This model achieves state-of-the-art retrieval quality among models under 100 million parameters. It is designed for efficient and accurate information retrieval across multiple languages.
→ This sub-100M multilingual embedding model with a huge context window is a strong contender for RAG applications, especially for those looking for open-source, performant alternatives to larger models.
Simon Willison released inaturalist-clumper 0.1, a tool he uses to publish his iNaturalist sightings to his blog. This release follows several weeks of production use and iterations on its functionality. An example of the output is available as a JSON file.
→ This is a low-signal item, but it's a good example of how developers are building custom tools for data integration and publishing, which could inspire similar RAG-related projects.
Simon Willison released datasette-llm-limits 0.1a0, a new plugin for Datasette. This plugin integrates with datasette-llm and datasette-llm-accountant to enable per-user or global spending limits for LLM usage within Datasette. It allows configuration of limits such as a rolling 24-hour window with a specified USD amount.
→ This is a practical tool for managing LLM costs in Datasette, especially useful for controlling spending in multi-user environments or when experimenting with various local LLMs.
Simon Willison released datasette-ip-rate-limit 0.1a0, a new plugin for Datasette. This plugin was developed with the help of Codex (GPT-5.5 xhigh) to address issues with poorly-behaved crawlers. It allows for configurable rate limiting based on IP addresses, with options for exempt paths and specific rules for different site areas.
→ This is a great example of using an LLM (even a non-local one) for practical, targeted development to solve real-world infrastructure problems, which is highly relevant for tooling and RAG improvements.
Learn how new ChatGPT safety updates improve context awareness in sensitive conversations, helping detect risk over time and respond more safely.
Tool: QR code generator Claude helped me build this tool for creating QR codes, for both text/URLs and for connecting to WiFi networks. Tags: vibe-coding , tools , generative-ai , ai , llms
This Mitchell Hashimoto quote about Bun migrating from Zig to Rust reminded me of a similar conversation I had at a conference last week. I was talking to someone who worked for a medium sized technology company with a pair of legacy/ legendary iPhone and Android apps. They told
OpenAI and Malta partner to expand AI access, offering ChatGPT Plus and training to help citizens build practical AI skills and use AI responsibly.
Use Codex anywhere with the ChatGPT mobile app. Monitor, steer, and approve coding tasks in real time across devices and remote environments.
We normally focus on technical stories, but occasional large fundraisings are noteworthy in themselves, and the Cerebras IPO (after one pulled S-1 and a fantastic 750MW partnership and $10-$20B stake/deal with OpenAI) this week, certainly qualifies as a growing theme supporting t
Special discounts up for AIE Melbourne ( LS discount ) and AIE World’s Fair (group discounts up to 25% - CFPs still open for Autoresearch and Vertical AI ) Cya there! Abridge did not start as an “GPT wrapper”. It was founded in 2018, years before the Cambrian explosion of AI appl
This might be the first AI demo in a LONG time that felt like there was actually a big step up in intelligence. Thinking Machines Labs just showed off their new “Interaction Models” and some of these demos genuinely feel like the next evolution of AI. They showed off real-time tr
[...] On the interesting side is how fungible programming languages are nowadays. Programming languages used to be LOCK IN, and they're increasingly not so. You think the Bun rewrite in Rust is good for Rust? Bun has shown they can be in probably any language they want in roughly
Preview a new personal finance experience in ChatGPT for Pro users in the U.S. Securely connect your financial accounts and get AI-powered insights and guidance grounded in your financial context, goals, and priorities.
See how data science teams can use Codex to build root-cause briefs, impact readouts, KPI memos, scoped analyses, and dashboard specs from real work inputs.
Databricks uses GPT-5.5 for enterprise agent workflows after the model set a new state of the art on the OfficeQA Pro benchmark.
Sea Limited's CPO explains why the company is deploying Codex across engineering teams to accelerate AI-native software development in Asia.
Here's the AI News you probably missed this week. Stop choosing between performance and budget and start building today with Crusoe Managed Inference at www.crusoe.ai/cloud/managed-inference?utm_source=mattwolfe&utm_medium=influencer&utm_campaign=bring_your_own_model_launch Disco
See how sales teams can use Codex to create pipeline briefs, meeting prep packets, forecast reviews, account plans, and stalled-deal diagnoses from real work inputs.
I'm always being asked to pick my favorite AI model, but the truth is my favorite changes all the time based on what's out there. At the moment, GPT-5.5 is probably my go-to for almost everything like coding, brainstorming, and general questions. But a month ago I probably would’
tp3_memories_localgemma3:4b)