AI: Major Emerging Existential Threat To Humanity

[Etern4l]

On 3/11/2024 at 9:01 AM, serpro69 said:

 

Well, nvidia for one is definitely being valued like there's no tomorrow. Won't be surprised to see it take on M$ by year's end.
https://www.reuters.com/technology/nvidia-cusp-overtaking-apple-second-most-valuable-company-2024-03-08/

 

There is certainly a FOMO aspect to that. BTW If there is no tomorrow than the value of NVidia stock should be close to $0. I don't think any adverse scenarios are factored into the pricing.

 

On 3/14/2024 at 10:42 AM, Eban said:

I just came across this youtube, how to run your own private chatgpt like LLM on a home pc. Feed it all your own data.

I havent even finished watching the video yet but had to share this..........still collating

 

 

This is not particularly new, but does dismantle the "AI safety guarantee" supposedly provided by containment of advanced AI models in big datacentres which can be powered off or nuked. 

 

BTW While we have seen Geoffrey's Hinton's talks here before, they are so good I would like to share another: a short lecture where he develops and delivers the fairly natural premises behind his AI safety concerns:

 

 

Worth watching the whole thing, it's not overly technical, but fastforwarded to the punch line for those time-constrained. 

[serpro69]

9 hours ago, Etern4l said:

There is certainly a FOMO aspect to that. BTW If there is no tomorrow than the value of NVidia stock should be close to $0. I don't think any adverse scenarios are factored into the pricing.

Yeah, the FOMO part is more or less what I meant ;)

"To do something like there's no tomorrow" is a phrase AFAIK , didn't mean it literally :)
(UPD: yep, it is a phrase - https://idioms.thefreedictionary.com/like+there's+no+tomorrow . Glad I didn't make that one up in my head  )

[Eban]

14 hours ago, Etern4l said:

 

This is not particularly new, but does dismantle the "AI safety guarantee" supposedly provided by containment of advanced AI models in big datacentres which can be powered off or nuked. 

 

 

 

It's not new? oh...that I didn't know 

[ryan]

I think after using hal in the 90s and seeing its progression this new generative ai is impressive, as for dangerous. Anything can be dangerous in the wrong hands, the real question is, is the polarity of AI leaning towards help or hurt, I think NSFW Ai can be modified for evil intents but major companies like microsoft and nvidia have given it a positive spin. The real censoring should be treating NSFW like a suitcase A bomb

 

no one here but if someone in the future visits this forum, perhaps aliens visiting earth chat gpt 4 passed the turing test

 

Can AI Be More Human than Us? - GPT-4 Passes Turing Test (msn.com)

[Etern4l]

On 3/18/2024 at 12:18 AM, serpro69 said:

Yeah, the FOMO part is more or less what I meant ;)

"To do something like there's no tomorrow" is a phrase AFAIK , didn't mean it literally :)

 

Of course, although the phrase also makes a lot of literal sense when applied in reference to hedonistic behaviour. I don't know too many people that would say that the last thing they want to do before the world ends is buy some NVidia stock lol

 

On 3/18/2024 at 5:22 AM, Eban said:

 

It's not new? oh...that I didn't know 

 

Well, this is going more mainstream now.

 

--

BTW I'm not sure any current LLM chatbot could pass an informed Turing test by an interviewer who is aware of LLM's shortcomings.

 

--

Meanwhile, uninformed government bureaucrats buckle under the FOMO pressure and embed the technology into the infrastructure:

 

NYC's Government Chatbot Is Lying About City Laws and Regulations

--

More good news from our "friends" at M$ and ClosedAI:

 

Microsoft, OpenAI Plan $100 Billlion 'Stargate' AI Supercomputer

I mean this is getting completely out of hand now. Remember any time you use the current data-harvesting M$ software and services, or otherwise give them any money: you are funding that.

 

--

Compared to the above, this next piece of minor significance and par for the course (in part because the technology is out there already):

 

OpenAI Reveals AI Tool To Recreate Human Voices

So, Closed are "responsibly" releasing a tool which can clone human voice. What could go wrong? This stuff should be criminalised ASAP. Why isn't it, America?

 

---

I saved this real gem for last:

Larry Summers, Now an OpenAI Board Member, Thinks AI Could Replace 'Almost All' Forms of Labor

 

So, Closed are getting bolder fast: previously they would publish those reports cautiously claiming about 20% or so of the jobs would be affected, but now they hired Larry Summers (a famous and outspoken economist) to sit pretty on the board and essentially profess the end of capitalism and human civilisation - clearly an excellent business opportunity in the era of AI-pumped disaster capitalism. 

 

Stay strong and keep thinking about the future guys, while we still can do something about it.

 

Actually another batch - I don't do these very often so hopefully this will be excused. Here is an update from a company which is building a humanoid robot in partnership with Closed, reportedly using some version of GPT under the hood, probably 5:

 

 

The voice is pretty uncanny, especially the human-like pauses. The robot is a bit slow, but clearly it's just a prototype. It also doesn't need breaks, doesn't waste time talking to co-workers, won't join a union etc.

 

Meanwhile, Sam A. is hoping to put some more creatives and film industry professionals out of work ASAP:

 

https://arstechnica.com/ai/2024/03/openai-shows-off-sora-ai-video-generator-to-hollywood-execs/

 

That's a follow up to the earlier story of some film industry projects already being cancelled:

 

Tyler Perry halts $800m studio expansion after being shocked by AI

 

[Etern4l]

Bad Apple news today:

 

Apple wouldn’t let Jon Stewart interview FTC Chair Lina Khan, TV host claims

 

What does this have to do with AI? Well, turns out Jon Stewart decided to join the chorus of high profile humans speaking out about the risks and consequences of AI... and Apple didn't like that, since they are reportedly planning to jump on the LLM bandwagon. 

 

 

In other news:

 

UK and US Sign Landmark Agreement On AI Safety 

Have not had a chance to scrutinise, but expecting max big tech sponsored BS along the lines of the below quotation:

 

UK tech minister Michelle Donelan said it is "the defining technology challenge of our generation." "We have always been clear that ensuring the safe development of AI is a shared global issue," she said. "Only by working together can we address the technology's risks head on and harness its enormous potential to help us all live easier and healthier lives."

 

Lastly, the artists are fighting on:

 

Billie Eilish, Pearl Jam, 200 artists say AI poses existential threat to their livelihoods