Encoderfile’s New Format: Why a “Dull” Design Wins
Encoder models don’t chat, and they don’t get much attention for it. But they’re behind most of the NLP that actually runs in production, powering…
Read MoreEncoderfile’s New Format: Why a “Dull” Design Wins
The Hardest Part of Running a Small Business in the Trades
The AI revolution has triggered a massive shift in daily life for knowledge workers. Developers, writers, analysts, and designers have seen their output transform dramatically…
Hardening Your LLM Dependency Supply Chain
On March 24, 2026, LiteLLM, a Python package, with over 95 million monthly downloads, was compromised. Versions 1.82.7 and 1.82.8 on PyPI contained a credential-stealing…
cq: Stack Overflow for Agents
Side A: Turtles all the way down / Side B: Mo’ tokens mo’ problems If you’ve been around long enough in anything you start to…
llamafile Reloaded: What’s New in v0.10.0
We are happy to announce the release of llamafile 0.10.0. Since our previous announcement, we’ve rebuilt llamafile from the ground up, following an approach that…
When Shipping Becomes Too Easy
When the hardest part of building shifts, so does leadership We have gotten very good at building software. We have not gotten equally good at…
Federate Phishing Detection: Training a URL Classifier without Sharing Browsing Data
At Mozilla.ai, we believe useful machine intelligence shouldn’t require centralizing sensitive user data. Federated learning offers a practical path toward collective intelligence without surveillance-style data…
Owning Code in the Age of AI
Software engineering is going through a shift that feels small on the surface but changes something fundamental: code is no longer scarce. For decades, writing…
The Star Chamber: Multi-LLM Consensus for Code Quality
Every AI model has blind spots. It might overlook context, lean toward certain patterns, or fill gaps with confident guesses. When you’re using an AI…
any-llm in the Wild: Three Integrations as We Grow Our Ecosystem
A core part of building any-llm is making sure it is present where developers already are. Over the past few months, we’ve integrated any-llm into…
Run OpenAI, Claude, Mistral, llamafile, and more from one interface, now in Go!
Go where the models are When we released any-llm v1.0 last year, the goal was simple: one interface to use any model, cloud or local,…
Integrating Alinia into any-guardrail for Multilingual AI Security
Introduction In Evaluating Multilingual, Context-Aware Guardrails: Evidence from a Humanitarian LLM Use Case, we explored how guardrails responded to the same policies and prompts in…
Evaluating Multilingual, Context-Aware Guardrails: Evidence from a Humanitarian LLM Use Case
Effective large language model (LLM) evaluation needs to be context-, language-, task-, and domain-specific. As developers gravitate towards custom performance benchmarks, they are also increasingly…
The Misunderstood Small Model Market: A Look Beyond OpenRouter’s Data
The recent State of AI report by OpenRouter and Andreessen Horowitz (a16z) offers compelling insights into the growing adoption of open-weight LLMs. It categorizes models…
Octonous: Making AI useful for everyday work
Most teams don’t wake up asking for “more AI.” They just want less busywork and fewer tabs open. In practice, that usually means one thing:…