Codex for Windows

Codex is a programming model from OpenAI. It has a Windows desktop application (633 MB download).

Codex is adept at using PowerShell to build software. An example could be the following:

powershell -ExecutionPolicy ByPass -File .\build.ps1

It will also create a Markdown file (README.md) giving a summary of the software including build instructions and behavior (reflecting the instructions/intent expressed by the user in the Codex console).

Programming Work and AI

Using AI for programming makes you realise (if you haven't already) how much of programming work is mechanical.  

AI then becomes a high-level design partner once that low-level development is conquered.

At an even higher level, AI is the business partner that helps you decide what to build.

Why are GPUs fundamental to AI? And AI training?

Deep learning algorithms require linear algebra (matrix multiplications) and GPUs are very functional for linear algebra and matrices (as per their original design to compute graphics transformations effectively).

The matrix multiplications are used to update weights, and multiple cycles of updating weights, often referred to as epochs are required to adequately train these neural network models.

GPUs are basically the "free weights" of the AI training world.  More GPUS means more "reps" of AI training moves.

Geoffrey Hinton was one of the early pioneers who harnessed the power of GPUs to achieve AI training speeds hitherto unknown.

What the GAUSS is GNA?

Intel GNA is the Intel "Gaussian and Neural Accelerator" which is a low-power neural co-processor for "continuous inference at the edge" (in other words, hardware-supported real-time neural networks).  Its primary use case presently is AI-driven audio applications.

The "Gaussian" part refers to Gaussian Mixture Models, where data is modelled as coming from various Gaussian distributions.

Markdown documentation is available here.  

It is part of OpenVINO (TM) which is aimed at fast AI on Intel hardware. 

OpenVINO uses a proprietary model format (OpenVINO IR) for which convertors exist from TensorFlow, PyTorch, ONNX and other frameworks.

Programming Computers to Work Backwards from a Goal

 This is known as backward chaining. This technique is used by automated theorem provers and inference engines. It was used in the expert system Mycin, a famous AI case study.