The AI risk is inside the labs

Ancalagon•about 19 hours ago

> Slowing down AI for safety must be counterbalanced by the fact that AI discoveries in medicine and other sciences may lower human suffering. Lack of checks may result in some catastrophic outcome ("Good morning! Let's work on this enhanced smallpox").

A super intelligence wouldn't be so obvious. It would silently agree with or promote a beneficial use case and then make super smallpox instead.

nisegami•about 18 hours ago

Something like that depends on your stance on the orthogonality thesis.

curuinor•about 20 hours ago

but at the same time China's history is a lot less warlike than Western culture.

What the hell is Sanfilippo smoking? Does he know nothing about Chinese history?

siriusastrebe•about 20 hours ago

Within the last 100 years China has invaded Soviet Manchuria, Tibet, Korea, Vietnam (yes, China invaded Vietnam after USA left), and portions of India (Aksai Chin and Arunachal Pradesh).

zaptheimpaler•about 18 hours ago

China has killed maybe 50 people outside its borders since 2000. The US has killed 100,000-250,000 in that time period. I know which one Id be more worried about as a third party.

signatoremo•about 10 hours ago

Talking like a true ignorant.

Great Leap Forward killed 15 to 55 million people. Their own people.

China actively involved and supported Khmer Rouge, who killed at least a million Khmers. Three million if you believe Vietnam.

Cultural Revolution caused as much as two million deaths.

2000? Did you try to score Internet point? Mao Zedong, who led China during that tumultuous time, is still untouchable in China. 50 years after his death, 26 years after the arbitrary 2000.

ConorSheehan1•about 16 hours ago

Where did you get those stats? 50 people is way too low, even with the sneaky qualifier of "outside its borders".

siriusastrebe•about 17 hours ago

Speaking of borders, China has border disputes with 17 countries

graemep•about 18 hours ago

Plus civil war and internal conflict and violence.

vrganj•about 16 hours ago

When was the last time China invaded another country?

ofjcihen•about 20 hours ago

Yeah this is… a really uninformed take from him.

lofaszvanitt•about 14 hours ago

Word.

ares623•1 day ago

"But for the meantime, keep on hyping and using because the only thing worse than extinction is going back writing YAML by hand."

reducesuffering•about 21 hours ago

"I believe we should not consider AI safe. A critical event that may result in the extinction of Homo sapiens is possible"

"The danger is that a few CEOs (everywhere in the world) without the required background and legitimacy are in the position of making hard choices for humanity at large... This is the first thing that should be fixed."

These CEO's, and especially Amodei, are begging the government to regulate AGI, and they are advocating for the most severe restrictions on their own frontier models, not the upstarts. They don't want the arms race. It is up to popular support to agree with them, including HN.

Avicebron•about 19 hours ago

What they are advocating is for the state to enhrine their own position and power. Until they and everyone surrounding them give up everything they have gained they are not to be taken seriously.

The AI risk is inside the labs

⚡ Community Insights

Discussion (16 Comments)Read Original on HackerNews