Mind-blowing interview of Jordan Peterson with scientist Brian Roemmele:
https://m.youtube.com/watch?v=S_E4t7tWHUY&pp=ygURUGV0ZXJzb24gcm9lbW1lbGU%3D
Key points:
Future application of #ChatGPT: The primary innovation of ChatGPT is its function as a reasoning engine. In its current implementation this is bundled with a predefined dataset, i.e. information until 2021 that is openly acessible via the Internet. Additionally, it lacks privacy. The future killer application in a corporate context will likely be "private AI based on a configurable dataset" so that one can control the dataset that the reasoning engine is applied upon while ensuring #privacy. Ultimately, this leads to the merge of the human with the AI creating a #cyborg.
#Immortality and #learning: With upcoming memory technology, it is imaginable to encode all visual and textual information that a human is exposed to in his lifetime. Conversing via #AI based on this dataset would feel like a having a conversation with this actual human being. By extending the model of the self with a model of the to-be-learned, AI could become a mentoring companion to every human.
Efficient interaction: ChatGPT's statements are sometimes ambiguous and evasive. In order to efficiently work with the system, you have to address the question from different angles. Intelligent engineering of prompts is key. ChatGPT allows you to toss around an idea in a human-like interaction while it is still in a nebulous state.
#Threat of AI: The dystopian view of AI is that has the potential to logically foresee all of a human's potential actions and calculate countermeasures within seconds. It is the technological foundation of a big brother worse than the one imagined in Orwell‘s in #1984. On top of this, time is only finite for humans, not for machines. The battle will be faught over the data and metadata that defines you as a person as this is the base information for exerting control over you.
#grownostr
Marco
npub1y3…qup90
2023-05-16 16:39:11
Author Public Key
npub1y3e675favj057kw4chsxcf84ml5yfqmwze5qanpwsy0t5etmfufs8qup90Published at
2023-05-16 16:39:11Event JSON
{
"id": "ea92893c6f3dad963f8d5abe4cc9802360e5b2ea8ae086c27cda54d3e5cd1a23",
"pubkey": "2473af513d649f4f59d5c5e06c24f5dfe844836e16680ecc2e811eba657b4f13",
"created_at": 1684255151,
"kind": 1,
"tags": [
[
"t",
"chatgpt"
],
[
"t",
"privacy"
],
[
"t",
"cyborg"
],
[
"t",
"immortality"
],
[
"t",
"learning"
],
[
"t",
"ai"
],
[
"t",
"threat"
],
[
"t",
"1984"
],
[
"t",
"grownostr"
]
],
"content": "Mind-blowing interview of Jordan Peterson with scientist Brian Roemmele:\n\nhttps://m.youtube.com/watch?v=S_E4t7tWHUY\u0026pp=ygURUGV0ZXJzb24gcm9lbW1lbGU%3D\n \nKey points:\n \nFuture application of #ChatGPT: The primary innovation of ChatGPT is its function as a reasoning engine. In its current implementation this is bundled with a predefined dataset, i.e. information until 2021 that is openly acessible via the Internet. Additionally, it lacks privacy. The future killer application in a corporate context will likely be \"private AI based on a configurable dataset\" so that one can control the dataset that the reasoning engine is applied upon while ensuring #privacy. Ultimately, this leads to the merge of the human with the AI creating a #cyborg.\n \n#Immortality and #learning: With upcoming memory technology, it is imaginable to encode all visual and textual information that a human is exposed to in his lifetime. Conversing via #AI based on this dataset would feel like a having a conversation with this actual human being. By extending the model of the self with a model of the to-be-learned, AI could become a mentoring companion to every human.\n \nEfficient interaction: ChatGPT's statements are sometimes ambiguous and evasive. In order to efficiently work with the system, you have to address the question from different angles. Intelligent engineering of prompts is key. ChatGPT allows you to toss around an idea in a human-like interaction while it is still in a nebulous state.\n \n#Threat of AI: The dystopian view of AI is that has the potential to logically foresee all of a human's potential actions and calculate countermeasures within seconds. It is the technological foundation of a big brother worse than the one imagined in Orwell‘s in #1984. On top of this, time is only finite for humans, not for machines. The battle will be faught over the data and metadata that defines you as a person as this is the base information for exerting control over you.\n \n#grownostr",
"sig": "41b2c545c4cabfb4f37cd5012084e2e56136928e77b2e5fc291d675a1ef44e3eb49a46cba9398ffa7e218d711e16db5f2d181fef6566808cd265ff3ebf722da3"
}