#code

Apple’s release notes detail that RDMA integrates with the Thunderbolt framework to enable zero-copy data transfers, meaning data moves directly from one device’s memory to another’s without intermediate buffering. This eliminates bottlenecks associated with TCP/IP protocols, which Thunderbolt previously emulated. Insiders note that while Thunderbolt 5 offers peak speeds, real-world performance depends on factors like cable quality and device compatibility—only M4 and later chips fully support this enhanced mode.

Diving deeper into the technical specifics, Apple’s developer documentation explains that RDMA over Thunderbolt is exposed through new APIs in the macOS networking stack. Developers can initialize clusters using Swift or Objective-C calls that negotiate memory mappings directly over the Thunderbolt bus. This is a departure from traditional Ethernet-based RDMA, which relies on Infiniband or RoCE (RDMA over Converged Ethernet), adapting instead to Thunderbolt’s point-to-point topology.

For those building apps, the update introduces protocols for fault-tolerant clustering. If a device drops out—say, due to a disconnected cable—the system can redistribute workloads dynamically, minimizing disruptions. Testing scenarios outlined in the notes suggest latency as low as microseconds for small transfers, rivaling dedicated high-performance computing setups.

Security is paramount in such a powerful feature. Apple’s notes emphasize built-in encryption for RDMA transfers, preventing unauthorized memory access. A separate 9to5Mac report on the update’s patches reveals fixes for kernel vulnerabilities that could have been exploited in clustered environments, ensuring that the feature doesn’t become a vector for attacks.

Looking at adoption, early sentiment on X suggests enthusiasm among AI researchers. One thread discussed collaborative model training, where multiple users contribute compute power via clustered Macs, democratizing access to high-end AI tools. This could disrupt markets dominated by cloud providers, offering cost savings for startups avoiding subscription fees.

RDMA lets the Macs all act like they have one giant pool of RAM, which speeds up things like massive AI models.

exo connects all your devices into an AI cluster. Not only does exo enable running models larger than would fit on a single device, but with day-0 support for RDMA over Thunderbolt, makes models run faster as you add more devices.

Run, manage, and scale AI workloads on any AI infrastructure. Use one system to access & manage all AI compute (Kubernetes, 20+ clouds, or on-prem).

The Ultimate Express. Fastest http server with full Express compatibility, based on µWebSockets.

This library is a very fast re-implementation of Express.js 4. It is designed to be a drop-in replacement for Express.js, with the same API and functionality, while being much faster. It is not a fork of Express.js. To make sure µExpress matches behavior of Express in all cases, we run all tests with Express first, and then with µExpress and compare results to make sure they match.

npm install ultimate-express -> replace express with ultimate-express -> done

µWebSockets.js is a standards compliant web server written in 10,000 lines of C++. It is exposed to Node.js as a simple-to-use, native V8 addon and performs at least 10x that of Socket.IO, 8.5x that of Fastify. It makes up the core components of Bun and is the fastest standards compliant web server in the TechEmpower (not endorsed) benchmarks.

Easily fine-tune 100+ large language models with zero-code CLI and Web UI

we in- troduce SWE-bench, an evaluation framework consisting of 2,294 software engineering problems drawn from real GitHub issues and corresponding pull requests across 12 popular Python repositories. Given a codebase along with a description of an issue to be resolved, a language model is tasked with editing the codebase to address the issue. Resolving issues in SWE-bench frequently requires under standing and coordinating changes across multiple functions, classes, and even files simultaneously, calling for models to interact with execution environments, process extremely long contexts and perform complex reasoning that goes far beyond traditional code generation tasks.

Easy to use, well documented fine-tuning. NVIDIA optimized with AMD support and Apple M support in the works.

continually updating a model's parameters with new data, often leads to “catastrophic forgetting” (CF), where learning new tasks sacrifices proficiency on old tasks. Researchers traditionally combat CF through architectural tweaks or better optimization rules. However, for too long, we have treated the model's architecture (the network structure) and the optimization algorithm (the training rule) as two separate things, which prevents us from achieving a truly unified, efficient learning system.

By defining an update frequency rate, i.e., how often each component's weights are adjusted, we can order these interconnected optimization problems into "levels." This ordered set forms the heart of the Nested Learning paradigm.

We observed that many standard optimizers rely on simple dot-product similarity (a measure of how alike two vectors are by calculating the sum of the products of their corresponding components) whose update doesn't account for how different data samples relate to each other. By changing the underlying objective of the optimizer to a more standard loss metric, such as L2 regression loss (a common loss function in regression tasks that quantifies the error by summing the squares of the differences between predicted and true values), we derive new formulations for core concepts like momentum, making them more resilient to imperfect data.

In a standard Transformer, the sequence model acts as a short-term memory, holding the immediate context, while the feedforward neural networks act as long-term memory, storing pre-training knowledge. The Nested Learning paradigm extends this concept into what we call a “continuum memory system” (CMS), where memory is seen as a spectrum of modules, each updating at a different, specific frequency rate. This creates a much richer and more effective memory system for continual learning.

"Nested Learning" extends the traditional two-tier memory concept of "attention layers" (short-term memory / context window) and "feed-forward network layers" (long term memory) into a spectrum of modules that update at different rates, some very frequently (like attention), some rarely (like FFNs), and others at various points in between.

Whalebrew creates aliases for Docker images so you can run them as if they were native commands. It's like Homebrew, but with Docker images.

Common Expression Language (CEL) is an expression language that’s fast, portable, and safe to execute in performance-critical applications. CEL is designed to be embedded in an application, with application-specific extensions, and is ideal for extending declarative configurations that your applications might already use.

A llama.cpp-based app for running local models.

A great open source alternative that I used for running llms locally without having to use llama.cpp directly.