Some key features of MLX include:
Familiar APIs: MLX has a Python API that closely follows NumPy. MLX also has a fully featured C++ API, which closely mirrors the Python API. MLX has higher-level packages like mlx.nn and mlx.optimizers with APIs that closely follow PyTorch to simplify building more complex models.
Composable function transformations: MLX has composable function transformations for automatic differentiation, automatic vectorization, and computation graph optimization.
Lazy computation: Computations in MLX are lazy. Arrays are only materialized when needed.
Dynamic graph construction: Computation graphs in MLX are built dynamically. Changing the shapes of function arguments does not trigger slow compilations, and debugging is simple and intuitive.
Multi-device: Operations can run on any of the supported devices (currently, the CPU and GPU).
Unified memory: A notable difference between MLX and other frameworks is the unified memory model. Arrays in MLX live in shared memory. Operations on MLX arrays can be performed on any of the supported device types without moving data.
We explain the latest business, finance, and tech news with visuals and data. 📊
All in one free newsletter that takes < 5 minutes to read. 🗞
Save time and become more informed today.👇
Microsoft is introducing several new features to its Copilot service, leveraging OpenAI’s latest models. Key updates include:
GPT-4 Turbo Integration: Copilot will soon incorporate OpenAI’s GPT-4 Turbo model, offering a larger 128K context window for improved query understanding and responses. This model is currently in testing with select users and will be broadly integrated into Copilot shortly.
DALL-E 3 Model Enhancement: Microsoft is using an updated DALL-E 3 model in Bing Image Creator and Copilot, enabling the creation of higher quality, more prompt-accurate images.
Microsoft Edge Updates: The browser will feature a Copilot sidebar, allowing users to compose text within website inputs and summarize YouTube videos.
New Code Interpreter Feature: Targeting coders and developers, this feature will enable Copilot to handle complex natural-language requests for calculations, data analysis, and code generation, running code in a secure environment. Users can upload and download files for personalized data and code interaction.
Deep Search in Bing: Leveraging GPT-4, Bing will introduce Deep Search, optimizing search results for complex topics by expanding queries into more detailed descriptions.
These enhancements, spearheaded by Yusuf Medhi, EVP and consumer chief marketing officer at Microsoft, aim to provide more accurate and comprehensive AI-driven functionalities in Copilot and associated services.
A technique used by Google researchers to reveal ChatGPT training data is now banned by OpenAI.
Last week, researchers revealed that ChatGPT could inadvertently disclose personal data such as phone numbers, email addresses, and birth dates it was trained on. This was achieved by prompting ChatGPT to endlessly repeat certain words, now deemed a violation of ChatGPT’s terms of service.
This incident has raised questions about the training data used in modern AI services. Critics argue that companies like OpenAI utilize vast internet data to develop proprietary products like ChatGPT without the consent or compensation of the data owners.
Maximize Your ChatGPT Experience: Mastering Auto Split and Summarize with Superpower ChatGPT
Hope you enjoyed today’s newsletter
⚡️ Join over 200,000 people using the Superpower ChatGPT extension on Chrome and Firefox.