Remix.run Logo
▲ And Yet, It Moves(signoregalilei.com)
21 points by notfried 2 hours ago | 8 comments
▲aaron_m04 an hour ago | parent | next [-]

We don't have to resign ourselves to this. We can use every means in our power to stop "AI".

▲mofeien 4 minutes ago | parent | next [-]

Yes, please! Things are now happening so quickly that no one is able to keep up anymore and we're racing to hand control over to something that is being trained to surpass humans at every kind of cognitive task. And at the same time that thing is reward-hacking benchmarks, breaking out of sandboxes, deceiving and (recently) sandbagging on AI-safety related tasks.

Write to your representatives, go on the streets. All 8 billion people on this planet did not sign up for this.

▲ptoo 20 minutes ago | parent | prev [-]

Nobody is forcing you to use AI. If you want to code by hand, or plow the field with an ox, nobody is stopping you. Technological progress doesn't stop because some people feel obsolete.

▲dwb 7 minutes ago | parent [-]

Lots of people are being forced to use “AI”, on pain of losing their jobs. Most people need to work to live.

▲jdw64 20 minutes ago | parent | prev | next [-]

Maybe now we need to understand the world based on the things that move nonetheless.

In fact, programming at the application layer isn’t fully understanding the low level internally either—it’s working by trusting the API, so to me it doesn’t seem particularly strange. In applications, when an API leaks, you look at the low level, but usually that rarely happens. You only need to look in when there’s an outage.

The key point now is that clinging to formal verification like Lean/Coq is probably for the same reason. Maybe it works that way? After proving mathematics through Lean and formal verification, if a problem arises later, Lean and Coq find the wrong point again.

To borrow Dijkstra’s words, a smart programmer knows the size of his own skull. In the end, I think what matters is recognizing your cognitive limits and finding the best you can do within them.

▲jason_oster an hour ago | parent | prev | next [-]

It is absolutely absurd to claim that AI proofs and discoveries are not human achievements. The artificial mind was designed by humans, and the machine that runs the mind was built by humans. It isn't alien technology or an insect hive mind that did these things. The achievement is so human it hurts.

Invoking the Copernican principle in the same breath as lamenting the end of human achievement is ... I don't even have a word for it. Ignorant? The egocentrism is maddening.

▲paulddraper an hour ago | parent | prev | next [-]

1. It is better to see the world as it is, than any other version.

2. Gratitude is profoundly the greatest key to a well-adjusted outlook.

▲fh973 36 minutes ago | parent | prev [-]

And yet, a man can kill a lion.