- Joined
- 31 May 2016
- Messages
- 29,195
- Reaction score
- 7,956
- Country
You’ve no clue.
10% chance of AI killing us?
- Thread starter Justin Passing
- Start date
you do?
- Joined
- 31 May 2016
- Messages
- 29,195
- Reaction score
- 7,956
- Country
I spent my whole career advising tech companies on IP and M&A.
- Joined
- 31 May 2016
- Messages
- 29,195
- Reaction score
- 7,956
- Country
Are you about to tell us about your long career in the industry?
Its just software - it has the same ability to destroy jobs, people, industries as any software.
Its just software - it has the same ability to destroy jobs, people, industries as any software.
Would you claim it couldn't access this ability in future?
- Joined
- 31 May 2016
- Messages
- 29,195
- Reaction score
- 7,956
- Country
No
you won't make the claim or you don't think it can?
- Joined
- 31 May 2016
- Messages
- 29,195
- Reaction score
- 7,956
- Country
Quantum Computing in its current form isn't scalable and is highly unreliable. AI has vastly more usable compute in the form of conventional GPU/CPUs.
But it's a capability it could attain in future?
- Joined
- 31 May 2016
- Messages
- 29,195
- Reaction score
- 7,956
- Country
along with any uninvented developments in compute hardware. Photonic, hybrid etc.
imagine you are moving some people, you have a bus or a motorbike. depending on how many you need to move and how far you are going, one will be faster than the other.
So why say this...
... when it's plausible that AI would find a reason to use this capability in future?
And when researchers are already looking into the possibility.
- Joined
- 31 May 2016
- Messages
- 29,195
- Reaction score
- 7,956
- Country
The practical applications of Quantum computing in its current state are extremely limited, due to the fact that the technology is neither scalable nor reliable.
Meanwhile AI data centres are being built with massive amounts of GPU and CPU capacity to support demand.
"in the future" - 1 year, 5 years 10 years 20 years?
some light reading for you:
Meanwhile AI data centres are being built with massive amounts of GPU and CPU capacity to support demand.
"in the future" - 1 year, 5 years 10 years 20 years?
some light reading for you:
Connecting the building blocks of scalable quantum computing
The race to build useful quantum computers is entering a new phase: scalability. Coordinated by Professor Niels Gregersen, a new MSCA Doctoral Network will develop the technologies needed to connect multiple photonic quantum processors into a larger, distributed quantum computer.
electro.dtu.dk
- Joined
- 22 Aug 2006
- Messages
- 7,819
- Reaction score
- 1,466
- Country
They are different things
If you've ever come across the OSI model, it splits communication into 7 "different things":
The Open Systems Interconnection (OSI) model is a conceptual framework that divides network communication into seven abstract layers, providing a standardized approach for different computer systems, applications and network devices to communicate across networks.
So, e.g., one layer tels you what represents a 1 or a 0 in voltage terms, and a different one sets out how you format the message from A to B putting the address it has to go to at the head, or whatever. That one doesn't care about voltages.
IN a computer, "quantum computing" and AI are again, different Layers. They don't "compete".
They they're different dimensions. Hardware versus software algorithms.
A quantum computer doesn't necessarily do AI. It runs quantum algorithms.
AI is a class of models and techniques that can run on classical hardware. So, not the same axis.
You might instead contrast physical substrate, transistors versus qubits, and computational approach, classical algorithms versus quantum algorithms, and then Model type, AI versus non-AI software.
You can't compare a model to a piece of hardware.
Maybe you're thinking that AI is like classical computing only faster. It's not.
Say your AI machinecan direct a load of wood and stone working tools.
You tell it you need a bridge to cross a river.
It goes from there. You have no idea how it's going to design and build the bridge. It would never come up with exactly the same design twice, because it would take different compromises. There's no single best design. You will not know what thought process it went through to come up with the design.
If you tell it the max is £10m , it might go to war by throwing hammers, and steal money from another country to get round that limitation.
If you tell it it has to cope with 1000 cars a day, it might make the bridge 1000 lanes wide, so they can all go over as quickly as possible.
A Quantum computer could go through all the options quicker, so it would make a quicker Chess playing computer, but there's no AI there.
The AI one might remove all the other player's pieces from the board, or it might ask you what style of play you want it to use.
It might not kill 10% of us per se but, if Orangina sees this, it might just do for him 
Similar threads
- Johnmelad502
- General Discussion
