Artificial General Intelligence (AGI) News and Discussions

[Tadasuke]

This might be the truth:

[weatheriscool]

For the most part A.i is at the same level as VR. It is about 20-30 years too early and needs toto gain a few elements so it can consider and form opinions within its own "mind" and control its self within a rational way. Until that day it won't be too useful and sure as hell won't be a tool that can move society ahead like something with 200, 300 or 1000 iq would be expected too be capable of.

[Tadasuke]

I agree.

I asked ChatGPT 10 questions and it responded with wrong answers to all of them. Turns out it's not useful to me at the moment.
I also consider search more important than generating "new" content from all the learned human content.

Google search works much worse than it did. I even have screenshots to compare how Google worked in the past and how it works today. I screenshoted in order to see how large of an improvement Google would do in the next few years. Instead of improving, they are making it worse and worse every year. Stuff that I could easily find is completely gone from results or very very far and difficult to see. I know, because I have a notepad with links. Sorry to write this. It's not what I expected, it's not what I've been waiting for. I'm very disappointed. I think that many digital things are just overhyped, overmarketed and not as good or useful as some think or want others to think. Microsoft Edge even underlines 'ChatGPT' with red, like it was a non-existing word...

As of now, I consider VR, MR, AR & AI as completely unworthy of my time. There is still other stuff to do.

[wjfox]

I asked ChatGPT

Which version? If you're still on 3.5, maybe it's time to upgrade...

[funkervogt]

For the most part A.i is at the same level as VR. It is about 20-30 years too early and needs toto gain a few elements so it can consider and form opinions within its own "mind" and control its self within a rational way. Until that day it won't be too useful and sure as hell won't be a tool that can move society ahead like something with 200, 300 or 1000 iq would be expected too be capable of.

I agree that AGI is 20-30 years away. However, I strongly disagree with your second claim that the narrow AIs we will have between now and then "won't be too useful and sure as hell won't be a tool that can move society ahead like something with 200, 300 or 1000 iq would be expected too be capable of."

LLMs are already better than most humans at a variety of tasks, and could become substantially better with mere tweaks to existing programming, more training data, and faster chips to run on. The random order in which human skills are being overtaken by machines (medical image interpretation, legal research, speechwriting, artistic drawing) is striking and shows how the job market is being rapidly altered in weird ways that affect people at nearly every level. The whole job market could get rocked by narrow AIs over just the next 10 years.

It's also easy to envision narrow AI being sufficient to create military drones that can fight with the same or better skill as humans. Having your nation's potential military strength no longer handicapped by the size of your young male population would revolutionize warfare. A small country not known for its fighting skill but rich with money like Qatar could suddenly become a military power.

[Cyber_Rebel]

Not so certain I'd compare AI to VR regardless of timelines. AI is not just a different monitor or new "interface" competing with TVs and monitors, it's a whole lot more general in the domains where it affects, be it programming, automation, knowledge use cases, entertainment, etc.

Even when it comes to GPT-4, I think a lot of people are becoming complacent thinking that it's the pinnacle of the field rather than the beginnings. The internet back in the AOL days when we were still trying to figure out how to best use it is a good analogy.

Also, worth noting that seemingly only a small portion of people even have a Chat GPT plus subscription: https://www.reddit.com/media?url=https% ... 2ae19a4472

0.1% Before subscriptions reopened, which means only a small amount are being exposed to the current SOTA the public currently has access to. I'm curious how perceptions would change if the full GPT-4V model were actually the free standard one.

[wjfox]

[spryfusion]

[Powers]

.

[weatheriscool]

AI companies will need to start reporting their safety tests to the US government

Source: ABC News/AP

January 29, 2024, 5:01 AM

WASHINGTON -- The Biden administration will start implementing a new requirement for the developers of major artificial intelligence systems to disclose their safety test results to the government.

The White House AI Council is scheduled to meet Monday to review progress made on the executive order that President Joe Biden signed three months ago to manage the fast-evolving technology. Chief among the 90-day goals from the order was a mandate under the Defense Production Act that AI companies share vital information with the Commerce Department, including safety tests.

Ben Buchanan, the White House special adviser on AI, said in an interview that the government wants "to know AI systems are safe before they’re released to the public — the president has been very clear that companies need to meet that bar.”

The software companies are committed to a set of categories for the safety tests, but companies do not yet have to comply with a common standard on the tests. The government's National Institute of Standards and Technology will develop a uniform framework for assessing safety, as part of the order Biden signed in October.

Read more: https://abcnews.go.com/Technology/wireS ... -106755590