“I Trained ChatGPT on My Journals to Talk to My Inner Child.”
“ChatGPT Can Negotiate Comcast Bills Down For You.”
“Since humans can’t manage fusion the US puts millions into AI-powered creation.”
These are just a few of the myriad of headlines that have been published since ChatGPT was released recently. I will link all of the articles below, and if you read them, you’ll see that these aren’t simple “gimmicks” or half-truths to get clicks. The young woman who trained it on her journal entries actually experienced some meaningful healing from chatting with her inner child via ChatGPT; the negotiation of the Comcast bill down was actually successful, although if it’s a bot talking to a bot, it can get stuck in a loop; and the AI managing fusion, things like this are already happening, so it’s not a stretch at all to imagine how much progress we will be able to make, far more than we could without it.
After years of people poo-pooing AI progress, claiming we were decades away from AGI still, belittling and dismissing those of who thought we could see it within the next 10 years or so, even the skeptics have to admit that what we are witnessing now with ChatGPT is some next level stuff. I didn’t even mention the coding capabilities. As someone who spent the greater part of the last 15 years writing code for a living, I am pretty blown away by just how robust its coding skills are getting. You can simply describe something you would like the code to do, and it will whip up a little method to accomplish it. And as someone who has never been the greatest at SQL, I am EXTREMELY impressed and pleased with the query creation capabilities. This has gone beyond something that might work, but is really more trouble than it’s worth, to something I could see being used by developers on a daily basis. Oh, and then there is the aspect where it actually *explains* in plain language what the code is doing. Wow.
And Google is planning to release something soon that could make even ChatGPT look quaint. Its new PaLM (Pathways Language Model) is looking to increase the number of parameters to 540 billion, which will unleash all kinds of new capabilities, and even deeper, more accurate reasoning and comprehension.
https://ai.googleblog.com/2022/04/pathways-language-model-palm-scaling-to.html
The vision statement for Pathways is
“Enable a single AI system to generalize across thousands or millions of tasks, to understand different types of data, and to do so with remarkable efficiency.”
Is that not the very definition of AGI, Artificial General Intelligence?
2023 could very well be the year – the year the oracle fully awakens.
Links to stories referenced at top of article:
https://www.businessinsider.com/i-trained-ai-chatbot-on-my-journals-inner-child-2022-12
https://www.vice.com/en/article/wxn955/chatgpt-can-negotiate-comcast-bills-down-for-you
Perpetual Mystic 