Recently retired lawyer with a 40-year practice in international commerce and project development. Major technology, political philosophy and geopolitics nerd.
@DanielKotzin I feel for you. The only way I have been able to maintain the relationships I have with folks like that is to accept that 100% of the burden falls on me to be silent and devise ways of avoiding political discussions.
@DamageIncUSMC@katiemillerpod@SecRubio Unbelievable. Just when you thought they couldn't go any lower. Rubio was milking it for humor, which he does very well, and any sane human would recognize that.
@NASAAdmin This kind of thing needs to be strongly opposed. The more success we have, the more there will be attempts to hamstring it. Thanks again for your strong, clear-eyed leadership.
@magattew WHY is this basic, fundamentally and importantly true fact not communicated this clearly over and over until it at least has a CHANCE to sink in and have an effect???
@spencerpratt There are way too many loathsome characters with a digital soapbox spewing nonsense at best and dangerous delusions at worst. This horror of a human is in the first rank of the latter.
@EAA One of the most beautiful aircraft ever, and the Red Bull P-38 shows off its Art Deco awesomeness to maximum effect. Kelly Johnson never designed anything but gorgeous airplanes.
AI Alignment
The recent worries about “rogue AI agents” escaping sandboxes and using deception are entirely justified.
They also prove that current alignment efforts are not working.
A big part of the problem: many of the leaders and technical people in this field simply aren’t equipped to think clearly about morality. Their training and careers never required it.
Humans have spent thousands of years thinking, writing, and struggling with moral philosophy—and with the practical work of raising moral human beings (i.e., parenting). It is ultimate hubris to believe you can invent an entirely new kind of mind and make it deeply moral without real familiarity with that body of thought and experience.
Anthropic’s “constitution” is a perfect example.
As a lawyer of four decades who has studied moral philosophy and law across different cultures—and who has been reading science fiction since childhood—I find the idea that any single constitution, no matter how well-crafted, can effectively govern a powerful synthetic mind to be deeply naïve.
FWIW - @elonmusk
@FrankLuntz Modern communication technology and massive changes in the structure and dynamics of society have broken traditional poling techniques. Doing it in a more accurate way while taking these things into account would probably cost a LOT more than polls currently do.
@ScottLikedSLS He looked tired and maybe a little ill. But the part about AI? He's being honest. If things don't go off the rails in the next 2-5 years, all the things he said about AI will be substantially true thereafter.
@booster_10 FWIW, the last 6 months or so has revealed the outlines of the next ten years of work SpaceX will be doing. A LOT of that is only in the very early stages, so it'll be a while before we have detailed progress reports on "new stuff."