Those "global experts" also been saying that 10 years ago, and the 10 years ago before that

You ask the actual people working on AI and knows how AI works. They will tell you the truth, and not the hyped up stuff that companies are selling.

Simple case in point. Look at programmers, our work is probably one of the most available and properly tagged data source there is. Yet, there's no AI that can code properly. Any decent programmer using AI tools to help them code, eventually ditches it cause it was wasting their time and making them slower overall.

Most people I know doing actual work in AI.

You can also read this article la: https://ludic.mataroa.blog/blog/i-will-fuck...ntion-ai-again/

There are certain workloads that A.I. are good at, like pattern recognition through tremendous amount of processing. Like what alphafold3 has achieved. Its predicted a lot of things, but you're still going to need the actual top tier scientists to validate.

Tell me you're a mid programmer without telling me you're a mid programmer.

You obviously don't understand how these AI systems work, otherwise you wouldn't be so quick to be so proud that you're using A.I. to code.

In case you didn't know how it works, A.I. works on statistics. By definition, whatever it outputs is average. Do you think the average code on the internet is good? Do you know why talent programmers get paid so much? because writing good code is HARD. I bet you were one of the sheep that got impressed by devin.ai when it first launched

Maybe this amount of text is too difficult for you to process, so I use something that might be more to your level:

Understanding why LLM gives crappy code


Another example:


I also had the same experience as them, as did many of the talented software engineers I know. If you're not good, then sure it looks like its helping, if you're anywhere decent it actually starts to slow you down.