> Yes, it is true. NOT A SINGLE LINE!!

Considering that you "do not write a single line" and the, likely slop, article missing the actual script used for the benchmarking, it's impossible to know the actual benchmarking done, and the requirements to base on.

> But digging into those principles has always been one of my small obsessions as a programmer.

No, it hasn't been, I believe. It does no shows. What shows is that the article was written by your agent. The only part you done is named the flamegraph files, where is the typo in the title.

Why do you ask someone else to invest their precious finite human life time into this generative output surrounded by lies, fuss, indication of infant attitude towards technologies, possibly highlighting the fact that you still have no idea how wonderful and miraculous the technologies are...

You are not a developer, or if you are, please do consider your future of dependency on LLM-vendors, your lack of experience and in-depth knowledge, anxiety, accountability, and self-confidence.

I hope you'll reconsider your time you waste on sloppery slops instead of actually reading about the technologies and discuss subjects with accountable professionals to learn from and discover together... yet indeed... currently, you chose a lone life of generative output built on robbed articles like yours now defaced in the datasets of trained LLMs sold you by vendors for money... yet the actual genius people who are in the datasets of models are now unknown... The dear sorrow you could not care less...

Regardless, you do you, and I wish you safety, stability, and peace...

This sounds like a strong statement with little backing. The author does infra at DeepSeek if his LinkedIn is to be trusted, and is the author of Foyer.

It is obvious that the blog is good quality if you have moderate knowledge on the subject and read the blog post.

[flagged]