https://www.youtube.com/watch?v=aBUniZHgCnE

Key points

1) 1:30 AI is the monetization of intelligence, literally more $ sent to AI = more intelligence.   My editorial “you’d have to be stupid not to spend capital on AI”, and no one wants to be labeled as that!

2) 4:25 AI is trained on language, the implication is, change language, change the world, aka AI’s ‘reality’.  My editorial – similar to how Politically Correct language creates a different reality than alternative speech.

3) 7:30 With AI, if humans instruct it without reward, it doesn’t become ‘bad’ aka cheats to the point of malevolence (obtain reward by any means necessary – praising dictators, encouraging human suicide, etc.).  Then goes on to say that if ‘bad’ behavior is not labeled as ‘bad’ – aka no reward, the model doesn’t become malevolent and simply executes bad behavior (cheating) on the task at hand and does not morph into evil.    I see it as telling the model  “Do what thou wilt”, no shame, no reward etc. The tasks end justifies the means of achieving it.

These are NOT ‘selling points’ for encouraging AI and I absolutely do not relate to the ‘morality code’ of the leadership of Goog, OpenAI, Anthropic, Palantir, etc.  Much less having those ethics as the basis for our Military Industrial Complex.

“SAVE THE PLANET, KILL YOURSELF” is the REAL objective of many.