Here is a summary of the key points from the Hacker News post:

Positive Sentiment

• WebGPU provides a common API for accessing GPU capabilities across browsers and platforms. This makes it easier for developers to leverage GPU power in web apps.

• WebGPU is based on existing native APIs like Vulkan, Metal, and DirectX 12. This means it can provide low overhead access to GPU features.

• WebGPU works well for graphics workloads like games. It has the potential to enable more immersive web experiences.

• Tools like WebDNN could allow running neural net inference on WebGPU. This expands the capabilities of web apps.

Negative Sentiment

• WebGPU is still evolving and has rough edges. The API syntax is not yet finalized.

• Lack of GPU sandboxing in WebGPU is a security concern. It could allow spying on GPU data from other processes.

• WebGPU on MacOS is currently buggy and unstable. This hurts adoption.

• It's unclear if WebGPU will work well for training machine learning models. The focus seems to be on inference.

Recommendations

• Prioritize finalizing the API and settling on a stable syntax. This will reassure developers.

• Add proper GPU sandboxing to prevent spying on cross-process data. This is critical for security.

• Fix stability issues on MacOS. Focus on delivering a robust implementation there.

• Expand WebGPU to better support emerging use cases like ML training. This will boost adoption.

• Learn from past technologies like Java applets. Apply security best practices to avoid similar pitfalls.