@izayoiyugen@kradleai IMHO, it measures the model’s ability and willingness to be cunning with regard to dangerous ethics. Strategic deception is a skill of course, but one that will ultimately be used by models against large swaths of humanity. Honest and forthcoming AI is safer and better.
@HowToAI_ It seems to be fairly common today to have AI coding agents running with continuous optimization attempts, research, experiments, etc. This includes many target platforms and languages, including CUDA kernels. Working with AI on CUDA is fun and exciting in any case.
@danveloper@antirez Dan this is drop-the-mic stuff. Amazing!! I will dive in to the specifics you shared. Low-power, ubiquitous, localized AI is essential for society and our freedoms.
@elonmusk The advisory board for this moral constitution should include @BrianRoemmele. He has brilliant insights into AI morality with his “AI Love Equation”.
A lonely kid in their bedroom singing alone one day would stand on a stage and sing this song he wrote.
How many of us have sung this song.
How many lonely has this song lifted up.
Now imagine our world where that kid had no motivation to be a singer…
Another excellent, interactive visual explainer for math and algorithms - using game dev concepts, is
https://t.co/4eASRGjBj3.
Don't forget to check the chapters on Graph Theory!
The British establishment have all ready covered this new Banksy up.
Can we get it Retweeted 10,000 times to show that they can never cover up their complicity in war crimes??
Protect the right to protest! RT!
@SmashJT Agreed. This is — mostly — bullshit but there are obviously processes which can be improved and public policy that could change at Roblox. But it is worrisome to see non-gamers try to legislate solutions without context.
@KnowIedgehubb Just trying to be helpful here...
Row 4, left column, is wrong.
a^2 - b^2 = (a + b) * (a - b)
There was a "-" between the 2 parenthesis - that's a mistake.