Willow. New computer working. 10,000,000,000,000,000,000,000,000 times faster.

[Justin Passing]

From Google.

Meet Willow, our state-of-the-art quantum chip

Our new quantum chip demonstrates error correction and performance that paves the way to a useful, large-scale quantum computer.

blog.google

The importance of this, is difficult to ponder.
I thought someone would solve Fusion reactors' stability, before the error problem in quantum computing.

(Actually the latter may help the former.)

If it's anything like they imply, it could be one of the most significant technological breakthroughs of a lifetime, or more.

Buy shares in Google, before it gets them all?

Joe Soap first saw AI when ChatGPT was announced 2 years ago. Each dollar spent then on slightly leveraged shares in the best-known chip maker (3NVD), which you can buy in your ISA , has grown by over 100x. Actual profits from AI, are still only just showing.

I bought 3x google 2 days ago, it's doing well

There's a way to go - "Applications, usability, acceptability, and accessibility will need to undergo their own evolution,”
but the HYPE around this thing is firing people up.

 

[motorbiking]

Meanwhile large scale GPU grids get cheaper and cheaper.

Serial computing tasks need super fast processors. Most businesses applications, AI/machine learning can be processed in parallel.

But it’s good PR. We just need to invent interstellar ion drives for a good use case.

 

[Justin Passing]

You missed the point completely.
We're not doing interstellar travel.
But we are computing data.
This small trial processor is 10 to the power 25 times faster that the current best.
10 million million million million times faster.

Speed wise you could run everyone's business appications concurrently. There are only what 10million million people. Give each one a most whizzy laptop. The laptops are logistically more usable, but this is a million million times more powerful than all of them working in parallel..

It's true that we don't have the applications yet, but military is an obvious field, as is medicine, and making money is too.

 

[motorbiking]

I didn’t miss the point.

 

[bernardgreen]

Any system that creates and then erases errors has to be treated with suspicion.

 

[motorbiking]

You missed the point completely.
We're not doing interstellar travel.
But we are computing data.
This small trial processor is 10 to the power 25 times faster that the current best.
10 million million million million times faster.

Speed wise you could run everyone's business appications concurrently. There are only what 10million million people. Give each one a most whizzy laptop. The laptops are logistically more usable, but this is a million million times more powerful than all of them working in parallel..

It's true that we don't have the applications yet, but military is an obvious field, as is medicine, and making money is too.

These models are shared nothing architecture. Business apps don’t work like that. And given my recent visit to ESS. I suspect ion drives have a lot to do with colder fusion. Anyway that’s what I understand from talking to them.

 

[Justin Passing]

"These models are shared nothing architecture."
Still missing the point.

"Business apps don’t work like that."
Still missing the point.

"I suspect ion drives have a lot to do with colder fusion. Anyway that’s what I understand from talking to them."
Irrelevant, to the point.



"I didn’t miss the point."
You keep doing it, because you aren't addressing it.

If you want to discuss something else, start a different thread.

 

[Justin Passing]

Any system that creates and then erases errors has to be treated with suspicion.

Only if you don't understand it. These guys DO understand it. They've been dealing with it for 20 years or whatever.

We've been using Hamming Codes since the 50's. They work like magic - check em out.

 

[blup]

Almost reported this as spam from the thread title

 

[bernardgreen]

Hamming Codes

detect and where possible correct errors created in the communications between devices.

Willow data processing seems ( I am ignorant of the details ) to use "errors" that it generates to calculate results.

 

[Justin Passing]

Almost reported this as spam from the thread title

I can understand that. It's that boggling. These are numbers you only find in astronomy.

That amount of computing power used for AI, has the potential - it seems a certainty - to produce something more intelligent than people by another huge multiple.

What intrigues me is what it can do to civilised existence - ie wreck it.

Like the AI forex trading app nearly available already for our phones. We shall all be able to generate more money that we can use, for ourselves The apps would all compete until one computer is best and wins all the time, but people not playing/playing as well, would become paupers. Where does that end? Everyone starves except the good players then "the one" with the best computer wins.. Who also starves because the workers all died.
I daresay there are good books about the idea.

 

[Justin Passing]

detect and where possible correct errors created in the communications between devices.

Yup, you add more bits to your data so the probabillity of an error shrinks so much it's insignificant.
I forget the ratio, but the Voyager spacecaft uses "lots" in its comms. You could say use 50 bits to convey 4 bits of data, in an environment where there's a huge amount of noise around. A lot of thought has gone into errors!

 

[motorbiking]

So the point is.

Buy Google buy Google buy Google ignore the 52 week high. so the price rises and you can make a few dollars.

 

[Justin Passing]

So the point is.

Buy Google buy Google buy Google ignore the 52 week high. so the price rises and you can make a few dollars.

Non sequitur.

The point is wow look at this processor and
I wonder what the market will do in response, and where could it lead in general.
If it's another NVIDIA it would be more than a few dollars.
If/when it looks right I'll stick 100k on it to get another m, or wherever it goes.

Using Baillie Gifford America B, and Morgan Stanley US advantage at the mo, with a tech fund or twoThey're similar. Pick Hedged or not, depending what you think the dollar will do. 12% in one month , 35% in 3 on those.
For a boring Bondy thing, MAN GLG Dynamic is now called MAN Group DynInc H
Artemis Funds (Lux) GLB HY HLD HACC is 1.5% a month.

 

[motorbiking]

Well if you are going to make investment decisions based on corporate blogs, you should read them carefully and ask yourself some questions.

Is this news already factored in to the price?
They have made a small scale test, how far away are they from a commercial platform?
Is there a genuine first mover advantage?
Will others do it better?