I approve of this, but in your place I'd wait for hardware to become cheaper when the bubble blows over. I have a i9-10900, and bought an M.2 SSD and 64GB of RAM in july for it, and get useful results with Qwen3-30B-A3B (some 4-bit quant from unsloth running on llama.cpp).
It's much slower than an online service (~5-10 t/s), and lower quality, but it still offers me value for my use cases (many small prototypes and tests).
In the mean time, check out LLM service prices on https://artificialanalysis.ai/ Open source ones are cheap! Lower on the homepage there's a Cost Efficiency section with a Cost vs Intelligence chart.
But it does disagree. Python is a dynamic language where essentially everything is indirect (an object).
Hardcoding types and methods (so they compile to simple/fast machine code as the video proposes) takes away flexibility (but you can do that with Cython by the way, but it is not nearly that popular).
Among the most popular languages is Python. It is popular in spite of its bad performance, high memory use, and lack of CPU multithreading.
And it is heavily ran on servers.
Why? Because running Python apps is still much cheaper than hiring humans to wait for calls or manage e-mails.
Humans are valuable. They should not be working on easily automatable problems.
The bottleneck is automating AT ALL, rather than automating with a low machine cost. Only at huge scale (i.e. Big Tech with billions of daily events) does it warrant to optimize the code.
Of course, assuming you have a sane computational complexity. If you don't, it doesn't matter which paradigm you use.
Indeed it is. I once bought a Sony phone for which Sony only offered 1 year warranty, but the store had to offer two (as per EU regulation).
The touchscreen started losing sensitivity on the edges (curved glass) after 13 months. The store made a bad decision carrying that phone model. They fixed/replaced it.