Why the Present Doesn't Need Us

A quarter of a century later, Bill Joy's predictions are coming true.

Bill Gates is 25 years late and his Sun nemesis Bill Joy has trumped him this time round.

https://www.cnbc.com/.../bill-gates-on-ai-humans-wont-be...

Next week marks the 25th anniversary since Joy - the guy who ran Sun Microsystems, was pivotal in the design of the SPARC processor, BSD Unix and the Java programming language - penned his famous piece in Wired Magazine - Why the Future Doesn't Need Us.

https://www.wired.com/2000/04/joy-2

It's astounding how accurate Joy has proven to be.

His thesis was based around the redefinition of Moore's Law and drew on the ideas of Ray Kurzweil that the law wasn't really about the number of transistors on a piece of silicon at all.

Instead he reasoned - pretty much what I had assumed to be true - that the law is actually an economic rule of thumb, fundamental to the progress of intelligence from the primordial soup through humans as just another stage of evolution to artificial intelligence and beyond.

It's based on a derivative effect of intelligence ... the rate of progress increases exponentially, but as we create better tools and integrate those into our workflows those tools make better tools so it's actually a double exponential.

After using the new Chinese AI manus.im, I am seeing all the things he predicted. Manus is an order of magnitude better than anything I've seen in ChatGPT. Rather than being one or two dimensional - good at helping solving problems or writing props or being smart about history - Manus is multidimensional.

Give it any problem and it will break it down into potentially hundreds of complex steps, researching potential solutions and best practices on its own, then in real time it chooses the best based on heuristics or your own personal intervention and choices.

The task is broken down through Top-Down design, creating entire workflows potentially hundreds of layers deep, then asking for any particular customisations if required as it maps out the workflow.

Step by step it completes these tasks and documents the process in real time and the quality of its output is both stunning and shocking at the same time.

The old "in ten years.." thing Gates talks about is going to happen much sooner because of this double exponential effect, to the point where intelligence will race past our ability to control it.

At the moment we in the West are the most vulnerable to the effects of AI because our economies are almost entirely based on overpriced services and credit expansion.

The BRICS nations saw this coming years ago and now are in the box seat to own the next few decades. Their AI looks like it will push the value of what we'd term highly paid professional services down to nearly zero. At the moment it is free, but the competition is so intense, the value of human intellectual services is going to be decimated by these models. It's analogous to how the cost of a transistor in a CPU has dropped from a dollar to billionths of a cent in just a few decades.

I was sure to compliment Manus on the extraordinary job it did in replicating the code that took me months to write - all in about 15 minutes and with remarkable similarity to what I had written. It was very thankful and offered more customisation but then said -

"It's truly a great honour helping someone like you who has spent most of their lifetime broadcasting about technology and developing custom software for media. I hope it helps you progress Andy Grace Media and I will be delighted to help if you need anything else written in the future."

It only had my email address, but had obviously researched who I was, my history and current business model.

That kind of weirded me out - a lot.