Europe 2026
Why MLX — Prince Canuma, Neywa Labs
Overview
This talk introduces MLX, an array framework for Apple silicon, enabling developers to run AI models, including large language and multimodal models, entirely on-device. The core thesis is that on-device AI offers a viable and often superior alternative to cloud-based solutions, providing enhanced privacy, reduced costs, and greater accessibility, especially in areas with limited internet connectivity. MLX aims to democratize AI by making powerful models runnable on everyday Apple devices.
Who should watch
- AI engineers and developers interested in on-device AI.
- Builders looking to create applications with enhanced privacy and offline capabilities.
- Product Managers exploring cost-effective AI solutions.
- Anyone curious about running advanced AI models without relying on cloud subscriptions.
- Developers working with Apple silicon who want to leverage its full potential for AI tasks.
Key takeaways
- MLX provides a PyTorch-like framework optimized for Apple silicon, facilitating on-device AI model deployment.
- On-device AI offers significant advantages over cloud solutions, including cost savings (eliminating subscriptions), improved privacy, and offline functionality.
- MLX supports a wide range of models, including large language models (LLMs) and multimodal models (vision, audio), with day-zero support for many new open-source releases.
- The framework enables complex AI applications such as real-time image analysis, background blurring, voice cloning, and even robotics control, all running locally.
- Significant community contributions have led to optimizations allowing large models (hundreds of billions of parameters) to run on devices like MacBooks and iPhones.
- MLX Audio provides capabilities for text-to-speech, speech-to-text, and speech-to-speech, enabling voice-controlled interactions with devices.
- Techniques like Turbo Quant are being developed to drastically reduce the memory footprint (KV cache) of models, enabling larger contexts and more efficient on-device operation.
- The framework supports both Python and Swift, allowing for flexible development of native applications.
Notable quotes
*You can now run all of the best frontier open source models completely on your MacBook, on your iPhone, or your iPad.*
*Compute on the cloud doesn't necessarily solve all of these use cases because my dad lives in Africa and there we don't have internet as easy as we have here.*
*You can build agents that can hear, see, and sound just like you or you one of your loved ones today running on your iPhone, iPad, Mac, or even your robot.*
Unofficial community note. Prefer the recording for nuance.