On AI

There's a 25% chance that things go really, really badly.
— Dario Amodei, CEO, Anthropic

What if flying in an airplane was basically free, but 1 out of every 4 flights crashed and everybody onboard died. Would you fly?

Believe it or not, this is the actual gamble that AI companies publicly state we are taking by continuing to develop AI in the current manner. Some experts even think the chances of catastrophe might be upwards of 100%.

Because the risk to the public— to YOU— is so extreme, I advocate it should be the primary concern of all voters in upcoming elections. We need to educate ourselves, we need to discuss the dangers broadly, and we need to vote for educated candidates that will band together globally and push extremely hard to mitigate this threat.

An Inconvenience

One of the most powerful metaphors I've heard on the topic goes like this— if you were in charge of constructing a skyscraper, and you found out there was an ant hill in the way, would you stop construction, or would you just proceed with pouring concrete? With this line of reasoning, super-powerful AI need not be evil at all to destroy us— as long as we are sufficiently insignificant, and in the way of some of its goals. At that point, could you blame it?

More updates coming shortly.