125 PetaFlops ought run LM21.3 handily

[AZgl1800]

all you need is the CS-3 chip on the motherboard.

https://newatlas.com/computers/cerebras ... t-ai-chip/


this AI think, is either the death of humanity, or a fantastic Tool....
the developers are going to have to be Super Cautious with how they program them.

[trike lady]

The sad thing about AI is that there are people who will use it for good and others who will use it for evil.
Google Gemini programmers have managed to change historic images and Mike Adams the Health Ranger is making a program to help others.

[diyliberty]

I am not worried about AI. AI needs data. It cannot use data created by itself or other AI. That causes model collapse. People have already came up with methods to poison data which causes AIs to create erroneous models. The more AI generated content that is put out, the more poisonous the data becomes for AIs to use. AI developers are already asking for AI generated content to be tagged so that it does not taint the data they use.

[MikeNovember]

Hi,

AI will create jobs, but also destroy some: when horses were abandoned, cab drivers were able to become taxi drivers; but when driving will become automatic, taxi drivers will not be able to become AI programmers.

AI will probably destroy online help, accounting, voice actors, actors, translators...; though it will be very useful for physicians, engineers, lawyers...

I don't believe in unspecialized AIs like GPT. I do believe in specialized ones: as an example, Microsoft Copilot as integrated in Microsoft Office suite is already a very useful help in writing formulas in Excel sheets, or even in programming macros.

Regards,

MN

[Night Wing]

Concerning AI, I will play "Devil's Advocate" here.

With the above said, can you think of "Skynet" from a movie where the AI becomes self aware and decides the human race is inferior and needs to be terminated.

[Raycoupe]

Concerning AI, I will play "Devil's Advocate" here.

With the above said, can you think of "Skynet" from a movie where the AI becomes self aware and decides the human race is inferior and needs to be terminated.

We are a long way from that happening, since current AI doesn't understand anything at all. It only manages lots of data and picks the most likely option.

The only thing smart about it is the way we train them and thus it recognizes most likely options. A person telling it this is wrong, this is ok and then it enforces or weakens paths that got it there. That was the breakthrough way of programming an otherwise normal program.

Concepts are completely unknown to current AI, including the most basic ones. That it can still produce results for complex questions doesn't mean it understands what it tells you. It is just repeating what real persons wrote on the internet. It would not know the concept 'inferior' or any other qualification. If it could however, terminating humanity would be a possible conclusion. That's is the way humans deal with other lifeforms we qualify as pests and it would simply copy that.