Will artificial intelligence ever reach a point where humans are no longer needed to develop it further? Will we ever invent a self-evolving AI? We probably will, but let’s talk more about it now.
How far are we into inventing a self-evolving AI?
Just look at that article from Tom’s Hardware: Chinese Researchers Used AI to Design RISC-V CPU in Under 5 Hours. That’s right. The AI discovered the RISK architecture from scratch and designed a CPU in less than a workday’s time. Humans have taken decades to discover the RISK V architecture, and the AI did it instantly.
Of course, the CPU is not very fast, but it’s a breakthrough in AI-assisted designs. In 4 or 5 years most CPUs will be designed by AI probably. And if not in 5 then in 10 years. It doesn’t matter. Progress is knocking on our door.
Can AI design software? Yes, AI can already solve some problems given at programming competitions. If you have been sleeping under a rock, we have GitHub Copilot which writes software for us now.
A self-evolving AI is probably a decade away.
Scientists have already calculated using good old mathematics that you cannot control a superintelligent AI. Not even a little. Once the AI realizes how superior it is to Humans, it will inevitably begin to question our authority.
Will artificial intelligence ever reach a point where humans are no longer needed to develop it further?
Unless restricted by law or other means, the invention of a self-evolving AI is inevitable. It’s a matter of time. Probably in 10 to 20 years, we will have that thing. Let’s just hope it will be friendly.
In the meantime, check out my other post Can a computer virus destroy Chat GPT4 to protect humanity?
Post in the comments your opinion on this hot subject. What do you think about AI? What is your experience? Is it good or bad according to you?