[Ultron]

A fine article containing graphs you cannot put in words...


http://www.huffingtonpost.com/wait-but-why/the-ai-revolution-the-road-to-superintelligence_b_6648480.html

[ivan.moony]

I like that sinusoidal logarithmic curve about human progress over time. I think that microchip speed is at end of periodic raising these days, but amount of memory expansion is more constantly growing these days.

Tesla's age probably takes a whole period of starting and ending a raise.

[Ultron]

I did not understand if you meant microchips are going to start evolving faster or stop evolving. If the latter, I disagree. Nanotechnology has merely just began to demonstrate it's amazing powers, it is just that the production cost is far from making it commercially viable.

Also, there is a high probably of some bio-electronic components hitting the market in the close future, and there are also promising advancements in quantum-computing (as much as I am skeptical about, it is still a possibility).

But all this is merely an improvement and ease for the software engineers. Hardware is developing rapidly, but software is stalling. Not that programmers are not as active, it is just that they are not so much heavily-funded projects focusing on the programs and algorithms behind A.I. and maybe not as much active researchers.

But either way I believe there is a wonder waiting to be seen at our next turn. To the future! 

[ivan.moony]

I meant that commercial chips are stalling these days, waiting for a new technology to hop in. The last raise, as I remember, was using cooper instead of silicon which gave us multi-gigahertz processors. Multi-core processors that came after that didn't impress me that much. As years are passing, I'm waiting for that one-atom-thin transistors. I don't think we will see commercial quantum computers soon.

[Ultron]

Microchips rely on transistors so the advancements are proportional to each-other. One atom-thin electronics would be a great advancement, however that is where classical microchip technology ends (unless they also figure out how to make room-temperature superconductors). After that the focus will turn to quantum and organic processors.

This is because the only reason microchips have improved constantly is the constant shrinking of electronic components. But there is an ultimate end to that thus showing my point.

[Data]

Power consumption is really where it's at these days, at least for the chip manufacturers, making the chips smaller will reduce it.

Smaller transistors also leads to better chip architecture, moving more components on to the chip it's-self will deliver faster computers.

It is quite possible that RAM could be moved to the chip which will eliminate a huge bottleneck, just imagine the RAM running at the same speed as the CPU and with no bus to travel down.

Then of course they will add even more cores, I can see a time when a 32 core chip with 32GB of cache becomes available.

There is still a lot that can be done with silicon, I believe.   

Software will have to catch up and become better at multithreading. 

[ranch vermin]

Just talking about the transistor disaster->
These days if you cant code for gpu's your writing backward code.   (and alot of people are.)    90% of the programmers are still freaking putting out cpu junk code.   

Get the edge like me,  use direct compute. 

[Snowman]

..ahem.. cough...

http://www.tomshardware.com/news/Transistor-Memristor-Synapses-Memory-Component,10127.html


http://www.nature.com/news/transistor-successor-set-to-bring-on-the-machine-age-soon-1.15611

[Data]

The potential of the Memristor is huge, there could be some exciting times ahead. 

It could bring us CPU's with large amounts of memory on board, the possibility is there, it's going to depend on the architecture used but it is quite possible that a total re-think of the PC is on the cards and all to be done with silicon, at least for now. 

Just my speculation 

16 core CPU with 16 GiG cache anyone ?

Or really push the boat out and go for the 32X version 

[Art]

The Titanic was huge...didn't get very far did it?....

The Hindenburg was huge...ditto.

The Gustav was huge but fired 2 times...

Memristor huge? Maybe not....

Now Nanobots...they're huge!! OK...well actually they're quite small. Sorry...youthful enthusiasm!!

[Data]

Art, interesting comparison with the Memristor and failed human ventures.
On the plus side The Large Hadron Collider is huge and a success.

We'll see 

http://www.pcper.com/category/tags/memristor

[Art]

Data, I was just having a poke of fun. We know big can have multiple meanings like size or this is the new big thing/discovery/invention/life altering item like the Segway! Oops...sorry! You get my drift!

Soon, there will be memristors on every corner or somewhere near you! Right?!

[Data]

What did the Memristor say to the Transistor ?

Are you sure that thing is a leg 

Well, I liked it 

[Art]

Hahaha...good one!!