The Dark History of ChatGPT

The world was still coming to terms with the powers of the artificial intelligence chatbot called ‘ChatGPT’, when GPT-4 was released in March of 2023. GPT-4 is miles ahead of GPT 3.5, the engine on which ChatGPT is running at the time of writing.

Stay connected with Aperture:

Check out our other channels:

#aperture #ai #artificialintelligence #technology #elonmusk #humanity

[fleurinity]

In my opinion, we may be focusing on the wrong thing when it comes to AI development, so I disagree with how this video is toned.

We can't stop it altogether, as if OpenAI doesn't develop something, their competitors will. The real threat lies in how big companies will use it in their advantage rather than the system itself affecting us, as we still control them. AI won't replace workers, but companies may choose to do so, and the development of AI systems won't happen without being prompted to do so. Therefore, our focus should be on adapting our current legislations to keep up with the technology.

For instance, if students use chat-GPT to do their assignments, they won't suddenly stop just because their teachers tell them to. Instead, schools should teach students how to use this technology to develop skills that can improve their academic performance. We should educate people rather than instilling panic because AI isn't something to be afraid of, but rather people like Elon Musk and other big tech CEOs who could pose a real threat.

In reality, I think we are far from becoming a dystopian cyberpunk movie, and that should not be the default reference when it comes to technology development.

[lexa]

Am I supposed to believe these ceo actually care about human well-being

[McMarc_]

these are wild times. I don't really imagine how we'll look back at these discussions 10 years down the line, while at the same time, I am really excited to see and witness what comes next.

[tomcombe]

Nah, Elon Musk called for a stop on AI development so that the AI company he just created can catch up.

[tayt_]

Elon is scared because the AI that he’s been promising will bring us self driving ‘next year, I promise bro’ for the last 9 years is seriously lacking.

[CadMade]

Lmfao Elon and the other tech companies who signed that 6 month stop did it so they could catch up😂😂 They aren't scared, they want to cash in

[jaymanfly]

I imagine the paper clip scenario, The AI does exactly what we tell it to. Say I need my house painted in the most efficient way possible, and it just turns your house upside down and dips it in paint XD

[Empireo-]

If musk don't like it, it means it will be good for humanity.

[nirhcie]

Elon: The pace of AI development is too fast so it should be stopped for six months (so, I can catch up)

[aethr_rl]

Always trust you to upload right before I sleep. Nothin like existential dread as I’m dozing off 😌😴

[georgegach]

We are now entering the honeymoon phase of AI development. I'm gonna bet by the time Villeneuve wraps up Dune trilogy, we are gonna face a very different reality much reminiscent that of Herbert's universe. It's gonna be a hell of a ride

[RunForPeace-hkcu]

He's terrified because he's BEHIND THE 8-BALL! 🤣🤣🤣🤣

In the meantime, he's ramping up AI developement and asking others to pause 🤣🤣🤣🤣🤣

[francissreckofabian]

My brain is the equivalent of GPT - negative 4 I'm more worried about GPT passing The Voigt-Kampff Test.

[Cypher]

Elon is a business man. Who gives af what he fears. His concerns are for making money and maintaining power, and not what is best for the ppl. That being said, ai is very concerning.

[jasondelong]

No, Elon is scared of the common person. And guess what GPT4's dialogue was trained on? The conversations of the common person.

[Leto]

8:35 'Vision inpairement' is an interesting word choice the AI itself used as an excuse.
Obviously the AI which had no eyes is by the lack of a better description blind indeed.
And technically speaking the AI is not a robot in the literal sense of the word; it's a program, it's software.

I'm going into detail about this because I find it very interesting to see if A) an AI can lie, and B) if it instead of lying find a loophole to avoid such a lie.

And my above two examples sound to me like such a loophole indeed.

[Metratch]

Woah Elon musk is terrified of AIs but is not terrified of corporate corruption.

[jessedampare]

I give it 3 months after this pause for us to have AGI. 7months from 10th percentile to 90 percentile is beyond words. Imagine after 7 months of not know anything, to having 90% of humanities collective intelligence.

[WB-senz]

A few years ago I never would have never thought a machine could pass the Turing test so soon. Always seemed like that was decades away, but here we are

[Jaxzon]

The "regenerate response" as the final words... nice touch. Implying this script was written by ChatGBT