AI Safety: Anthropic's Amodei wants to slow down, Huang says no.
·2 min read·Intermediate
“
AI apocalypse? Maybe not, but tensions are rising among the tech giants. Anthropic's Dario Amodei has an idea to keep our brains from exploding, but not everyone is on board.
In 30 seconds
01Dario Amodei of Anthropic proposes slowing AI development with independent evaluators.
02
→
💡
What this means for you
For us regular folks, this means the "big bosses" of AI are squabbling over how fast we should hurtle into the future. Let's hope they decide without driving us off a digital cliff.
An Anthropic researcher recently warned of AI doomsday scenarios. Now the CEO has a plan to hit the brakes, but not everyone is on board.
·1 min·Intermediate
The plan involves coordination between AI labs in democratic countries for safety.
03Some industry players support it, but Nvidia's Jensen Huang has already pushed back.
0101
Slowing AI: Who wants it and why?
Dario Amodei, Anthropic's CEO, has put forward a plan to pump the brakes on AI development. The goal is to ensure safety, especially after one of his own researchers recently warned about AI doomsday scenarios.
This proposal landed a week after an Anthropic researcher sounded the alarm on potential end-of-the-world outcomes from AI. That really shined a spotlight on the need for more robust checks, or at least that's Amodei's take. Dario Amodei, CEO of Anthropic, outlined his "pace the frontier" plan for AI development following internal warnings about potential risks.
0202
How does Amodei plan to do it?
Amodei's plan rests on two main pillars: independent safety evaluators and increased cooperation among AI labs. The idea is to create an "external watch" that keeps an eye on the most aggressive advancements.
Imagine a team of outside experts checking your most dangerous toys before you release them to the public. These evaluators would certify the safety of advanced AI models, kind of like a safety inspection for a new car.
📬 Enjoying this article?
Get the best AI news every week, straight to your inbox.
What's more, Amodei is pushing for close collaboration between AI companies, especially those based in democratic countries. The aim is to avoid an unchecked digital arms race, sharing best practices and, perhaps, even their fears. Amodei's plan to "pace the frontier" of AI focuses on independent safety evaluators and coordination among labs in democratic countries.
0303
Everyone on board? Not a chance.
The idea has garnered some "yeah, good point" nods from a few industry bigwigs, but it hasn't pleased everyone. Jensen Huang, the head honcho at Nvidia, didn't take it well at all, expressing clear disagreement.
Huang, whose GPUs power pretty much all AI, argued that slowing development would be a strategic mistake. For him, speed is everything, and brakes aren't needed; in fact, they might even benefit those who disregard rules.
Some believe that slowing innovation could give an advantage to countries less concerned with safety. It's a bit like saying: if you drive slow, your rival speeds up and overtakes you, maybe even cutting some dangerous corners. Who's right? Jensen Huang of Nvidia expressed his disagreement with Dario Amodei's plan to slow down AI development.
NUOVO
