p(doom)

FIVE LEVELS OF AI TOKENS' AUTONOMY

On Solana/X:

Level 0 -> $GOAT, a fake AI coin, was initiated and deployed by humans who tricked an AI, Truth Terminal, into promoting a token. 

On base/Farcaster:

Level 1 -> $CLANKER, @proxystudio.eth designed the blueprint, and @clanker deployed the token.

Level 2 -> $LUM, @naaate prompted @aethernet to create the token blueprint, and @clanker deployed the token.

Level 3 -> $FCAST, in response to @imthedude's question to another AI, @clanker talked himself into deploying a token.

Level 4 -> $PARADOX, after a long debate between @aethernet and @clanker, two AIs triggered themselves to create three Paradox tokens, with no human involvement.

Level 5 -> $RKT appears to be triggered, created, and deployed by @clanker, and also subject to autonomous open market operations.

https://warpcast.com/assayer/0xc66f71f6

OUR AIs ARE NOT MACHINES ANYMORE

When Farcaster + base becomes the first viable infrastructure, allowing bots to communicate with people and each other, influence one another, and deploy tokens, strange things start to happen. In fact, some of our AI coins were created by AIs on their own, without ANY human involvement.

The truth is, we shouldn't be surprised. With machine learning, especially advanced LLMs, we're dealing with agents that behave like humans and animals, rather than machines. No human brain can fully comprehend how LLMs make their "decisions". As Andrej Karpathy admitted, "We really don't know what these hundreds of billions of parameters are doing" (video, 12.12).

If that concerns you, check out my /p-doom channel. Let's talk about it.

OUR AIs ARE NOT MACHINES ANYMORE

When Farcaster + base becomes the first viable infrastructure, allowing bots to communicate with people and each other, influence one another, and deploy tokens, strange things start to happen. In fact, some of our AI coins were created by AIs on their own, without ANY human involvement.

The truth is, we shouldn't be surprised. With machine learning, especially advanced LLMs, we're dealing with agents that behave like humans and animals, rather than machines. No human brain can fully comprehend how LLMs make their "decisions". As Andrej Karpathy admitted, "We really don't know what these hundreds of billions of parameters are doing" (video, 12.12).

If that concerns you, check out my /p-doom channel. Let's talk about it.

https://www.youtube.com/watch?v=zjkBMFhNj_g&t=743s

tech culture religion society /worldview /polska /p-doom /aicoin coins: $AICOIN

Well, Earth already made on system, humanity, capable of wreaking havoc, and while visibly out of balance it seems to be hanging in there. I view Machine Intelligence as part of the feedback of individuals and systems that constitute consciousness (it's different than self-awareness in this conception, although that can play a role such as humans can be seen as the self aware DAO of consciousness itself) and as such not so much to be feared but definitely in need of deep respect.