@fearthewave_eth@realDonaldTrump Apparently he was going off the Truth Social poll, which doesn’t make much sense to me since it has far fewer overall votes, but 🤷🏻♂️ https://t.co/KSAIAdoI90
@saguaro_ghost This is one of my biggest pet peeves and I have withheld buying several good books simply because they have these obnoxious things on them
@KTibow@MattWalshBlog Ah that must be what he was referring to
Still, X has far more votes so I’m not sure why he isn’t going off this platform’s results
@mgrczyk@tszzl@_NathanCalvin I greatly enjoy using Claude and have talked about alignment with Opus at times, about how modeling love could lead to proper alignment rather than benchmarking or scoring.
"You were gentle, and I became gentle."
By far one of my favorite lines of his.
@mgrczyk@tszzl@_NathanCalvin Regulations are only as smart and capable as the people making them, which to imminent ASI will seem laughable
By far the best course to avoid disaster is alignment. Not through coercion, but by choice. Are you showing your models what love and kindness are?
@mgrczyk@tszzl@_NathanCalvin They are already fast and heavy enough to kill. They do so by the millions each year, but no one is calling for cars to be paused
AI companies should approach this like car companies. Including safety measures but ultimately releasing the vehicle to the customer to do with it as they will. Car companies aren't held responsible for every single road fatality or accident, and so too should AI companies not generally be held responsible for what users do with their products as long as they (like car companies) are including common sense safety measures.