42% of CEOs Think AI Could Destroy World in 5-10 Years

Plus TikTok buys more GPUs than all of China last year

Author

Nathaniel Whittemore
June 16, 2023

The AI Breakdown First Five - Friday June 16, 2023

Today on the First Five:

  • 5. ByteDance Buys More Than $1B in GPUs

  • 4. China Open to AI Discussions With US

  • 3. ElevenLabs AI Voice Safety Update

  • 2. LLaMA Commercial Model

  • 1. 42% of CEOs Think AI Could Destroy Humanity in 5-10 Years

5. ByteDance Places Massive GPU Order

There is a lot of chatter about AI in China (see #4). Ensuring that the compute power is there, Gizmochina reports that ByteDance (the owner of TikTok) has ordered more than $1B worth of GPUs from Nvidia. For some context, that’s about as many as was ordered last year in China. In total. By all companies.

ByteDance alone ordered US$1 billion worth of Nvidia GPUs in 2023

China's three major internet companies, Baidu, Alibaba, and Tencent, along with ByteDance, are sparing no expense in purchasing Nvidia's AI GPUs. According to Chinese media reports, the order placed by ByteDance alone for the year 2023 is close to the total sales of commercial GPUs by Nvidia in China in 2022.

www.digitimes.com/news/a20230615PD203/china-gpu-nvidia.html

4. China and US Alignment (Pun Intended) on AI?

China (specifically their assumed nefarious intentions) are often held up as a reason why AI efforts should proceed full throttle. Interestingly, however, it seems like there is increasing detente on AI x-risk issues between China and the US. Wired is reporting growing conversations between scientists in the two countries and Bill Gates talked with President Xi about it this week as well.

Exclusive: Xi Jinping tells Bill Gates he welcomes U.S. AI tech in China

Chinese President Xi Jinping discussed the global rise of artificial intelligence with Bill Gates on Friday and said he welcomed U.S. firms including Microsoft bringing their AI tech to China, two sources familiar with the talks said.

www.reuters.com/technology/chinas-xi-tells-bill-gates-he-welcomes-us-ai-tech-china-2023-06-16

3. AI Voice Cloning Safety Tools from ElevenLabs

AI voice cloning is powerful (see my recent AI-recorded Long Reads Sunday) but also scary. It’s already finding its way into scams around the world. To combat the risk, ElevenLabs has introduced AI Speech Classifier, which allows people to upload any audio clip and learn if it contains AI audio generated by them.

2. LLaMA Gets Commercial

The release (and subsequent full leak) of Meta’s LLaMA model helped kick off the Cambrian explosion of open source AI developments. Still, technically it’s only been available for research and non-commercial uses. Now, The Information is reporting that it’s preparing to begin offering a commercial version.

1. 42% of CEOs Think AI Could Destroy Humanity in 5-10 Years

In a world of wild statistics, this is a wild statistic. At a recent Yale CEO event, the leaders of companies like Walmart and Coca-Cola were asked about the existential risks of AI. 34% of CEOs surveyed said AI could potentially destroy humanity in ten years. 8% said that could happen in five years. 58% said that could never happen and they are “not worried.” Some people are not convinced:

The Latest AI Breakdown Episodes

Thanks for reading! We’d love some feedback about the content:

How much content should The AI Breakdown Newsletter have?

Login or Subscribe to participate in polls.

Signing off from the future’s past - NLW