Ask HN: Best hardware investment for learning ML in 2023?

[krkartikay]

Hi HN'ers, I'm a new grad / software engineer and have decided to learn ML and experiment with some models. I wanted to buy a PC for coding and ML experiments. I want to try running larger models and perhaps experiment with new architectures too. I am also very interested in RL. (I am aware that I cannot compete with companies like OpenAI in terms of compute power but I still have a hope of doing something new in terms of application or architecture).

What would you consider the best investment for a begginer? Would a M2 Mac be a good choice? My main concern is that my initial investment should be extensible, so I could add more compute/GPUs later if required.

[kristianp]

I'm wondering the same thing. I see that nvidia gpus are the most widely supported hardware, however there are cloud options for those. How do you decide whether to go cloud?

Small datasets for learning about machine learning abound, which can run be locally.

[hardchaos]

Most people do not build/train large models locally - if you get to that point, you'll be renting cloud compute. Any computer that has a GPU and is within your budget is good to get started with intro models/Kaggle.