The AI Daily Brief: Artificial Intelligence News and Analysis - Will Intel's New Chip Usher in the 'Age of the AI PC'?

Episode Date: September 20, 2023

Intel has announced a new AI focused chip that they say will allow generative AI applications to run directly on device without being connected to the internet. The chip comes out in December. Also on today's episode, NLW explores the controversy around the UK's decision to invite China to its upcoming AI Summit. TAKE OUR SURVEY ON EDUCATIONAL AND LEARNING RESOURCE CONTENT: https://bit.ly/aibreakdownsurvey ABOUT THE AI BREAKDOWN The AI Breakdown helps you understand the most important news and discussions in AI.  Subscribe to The AI Breakdown newsletter: https://theaibreakdown.beehiiv.com/subscribe Subscribe to The AI Breakdown on YouTube: https://www.youtube.com/@TheAIBreakdown Join the community: bit.ly/aibreakdown Learn more: http://breakdown.network/

Today on the AI breakdown, the UK court's controversy as it invites China to its AI summit later this year. Before that on the brief, can Intel compete with new AI-focused chips for laptops? The AI breakdown is a daily podcast and video about the most important news and discussions in AI. Go to Breakdown.network for more information about our Discord, our newsletter, and our YouTube channel. Welcome back to the AI breakdown brief, all the AI headline news you need in around five minutes. Today we kick off with yesterday's big event from Intel, where, surprise, surprise, the big theme was artificial intelligence. Now, Intel is not necessarily a company that you hear about at the top of the list when it comes to impact on AI or being impacted by AI. And indeed, one comment I saw

on someone's post was, they need something. They are firmly in Prove It territory. So what was their offering? Well, they are, as Reuters puts it, pitching the AIPC. There were a number of different announcements, but the biggest one, or at least the one that's capturing the most attention, is that Intel is promising a new chip to be released in December that will be able to run generative AI apps directly from a laptop rather than from the cloud. So this is the Meteor Lake CPU that Intel has been talking about for some time, but it finally has a launch date of December 14th. For the first time ever in an Intel chip design, there will be an integrated neural processing unit to, quote, enable power-efficient AI acceleration and local inference on the

PC. In order to show off the technology, Intel had a demo laptop that generated a song in the style of Taylor Swift, as well as answered questions in an AI chatbot style, all while being not connected to the internet, said CEO Pat Gelsinger, quote, we see the AIPC as a sea change moment in tech innovation. Now, in addition to Meteor Lake coming at the end of this year, Intel is also promising a follow-up successor chip Arrow Lake to come out next year. The company claims that their technology will rival the best from TSM. And one of the things that's most interesting to me is that you can really see the business strategy start to form around this AI chip battle. While Nvidia feels to so many to have the chip market for big data centers veritably locked up, it appears that Intel is trying to go

the other direction and try to gain ground in the market for chips that handle AI work outside of those data centers. The company was very clearly emphasizing something that we've talked about extensively on this show, which is the belief, or at least the narrative, that enterprises are going to highly value AI hardware and applications that can be run on local devices without having to hand over sensitive data to third parties. Now for the sake of completeness, the set of quote unquote lake chips coming from Intel include Meteor Lake this year, followed by Lunar Lake, a preview of which was also shown off at the event, which was the first time it had been, and then finally Panther Lake, which is slated to arrive sometime in 2025. Now, the other big

announcement that captured attention in the AI world was Intel's announcement that it is building an AI supercomputer for Stability AI. Stability AI is, of course, the company behind stable diffusion, and Intel CEO said that when the supercomputer was complete, they believed that it would be the largest in Europe and among the world's top 15 supercomputers overall. Rather than Nvidia chips, this will be filled with 4,000 Gowdy 2 deep learning processors, which is an implementation after Intel's 2020 acquisition of Israeli startup Havana Labs. Now, despite all these announcements, the market didn't seem all that impressed. Reuters reports that Intel shares were down 1.5% after the company's presentation.

Does that mean that the market isn't as excited as it used to be about AI? Or just that the market isn't convinced that Intel can compete in this white-hot space. Speaking of AI in the markets, Ark Invest's Kathy Wood has reiterated her call that Tesla is the biggest AI opportunity in the world. The argument is, of course, based on how significant she and her firm believe autonomous driving will be in terms of reshaping society. And I have to say that after Elon's recent full-sumption, self-driving mode demo? My guess is that more people agree with her now than the last time she said this back in May. Speaking of Elon Musk, another one of his company's Neurrelink has just opened recruitment for its first ever human clinical trials. Neurrelink is of course a brain computer

interface, and in their blog post, Neurlink writes, we're happy to announce that we've received approval from the reviewing independent institutional review board and our first hospital site to begin recruitment for our first in human clinical trial. During the study, the R1 robot will be used to surgically placed the N1 implants ultra-fine and flexible threads in a region of the brain that controls movement and tension. Once in place, the N-1 implant is cosmetically invisible and is intended to record and transmit brain signals wirelessly to an app that decodes movement and tension. The initial goal of our BCI is to grant people the ability to control a computer cursor or keyboard using their thoughts alone. The company is specifically looking for people who have

quadriplegia due to cervical spinal cord injury or who have ALS. Staying on the medical theme for a moment, the Chan Zuckerberg Initiative, which is of course Mark Zuckerberg's main philanthropic vehicle, is building a significantly sized AI GPU cluster dedicated exclusively to medical research. The computing cluster will include more than 1,000 H-100 GPUs, and in a press release, the Chan Zuckerberg Initiative says that this will, quote, lead to groundbreaking new discoveries that could help cure, prevent, or manage all diseases by the end of the century. Now, if you're interested in this, I'll include a link to an essay by Priscilla Chan and Mark Zuckerberg that was published in the MIT Technology Review, where they talk about this

virtual cell modeling system, which they believe will be key to breakthroughs in our understanding of diseases. Turning to a story that is more immediate in the world of AI right now, TikTok is launching that tool that was promised a couple months ago that gives creators the ability to label when they have used AI to create content. In addition to this optional label, they're also going to be renaming effects on the app that use AI to explicitly include that phrase AI in their name and the label that comes with them. Lastly, an interesting paper around the energy impact of AI. One of the things that we tend to see at the beginning of any new technology hype cycle are assessments of how much energy it uses. Now, this is a conversation that I have had numerous times on numerous podcasts,

but it's jumping from the world of crypto and Bitcoin to the world of AI as well. This new paper argues, however, that the CO2 emission for an image created with Mid Journey or Dali 2 is radically lower than the emissions from a human creating the same image, or even today's laptop or desktop computers doing the same. The study also shows similar results when it comes to writing a page of material. Now, of course, if there is a thousand X difference between how much energy it takes for a human to create a page versus chat GPT to create a page. Will the presence of chat GPT 1000X the number of pages that get created, making it all sort of null? Hard to say, but still it's an interesting little piece of nuance in what

is going to probably become a political conversation. Anyways, friends, that is going to do it for today's AI breakdown brief. If you enjoyed it, chuck us a like or a subscribe. Next up, the main AI breakdown. Hey guys, one more quick note before we dive in. I so appreciate everyone who has taken the time to fill in our educational content survey. The TLDR, for those of you who haven't, is that we are considering launching a number of different types of AI educational content to better help you retrain, reskill, get to where you want to be when it comes to AI, including some really different ideas, such as an AI learning community that would exist on Discord or some other shared space. But I really need feedback to know what people actually

need and what people most want, and if you are willing to take less than one minute to give me that feedback, I would be ever grateful. Just go to bit. slash AI breakdown survey. And like I said, it'll take you less than a minute to fill that out. Thanks in advance, and let's get to today's show. Welcome back to the AI breakdown. One of the dimensions of AI that we track here at the AI breakdown is, of course, the geopolitics surrounding it. Now, this is not just an issue that matters to policy wonks and people who have a history or social sciences background. This is an issue that could have significant impacts on how the field develops. In the U.S., tensions with China are

already impacting things like where American chip companies can sell their wares or not. In the military sphere, there is a constant and accelerating arms race when it comes to AI applications for future fighting uses. China is also at the center of the debates around AI safety. You will often hear people point to China as the reason that we can't slow down, effectively saying they're not going to agree to slow down so we can't either. And so one of the big questions has become how much to try to include China in those larger global governance conversations. Well, that is getting a live test right now because the UK government has invited China to participate in their big AI summit later this fall. UK foreign minister James Cleverly said in a statement, quote,

We cannot keep the UK public safe from the risks of AI if we exclude one of the leading nations in AI tech. Now, to understand Cleverley's background in this context, last month he became the most senior minister in the UK to visit China over the last five years, and broadly speaking is in the camp of arguing for deeper engagement with Beijing. He said that there are too many big issues, ranging from climate change to economic instability to now artificial intelligence, where it just doesn't make sense to isolate the world's second-larged economy. And yet at the same time, the UK is currently embroiled in a Chinese spy scandal. As Politico writes, Rishi Sunak has spent most of his brief premiership treading a delicate balance on relations with China. So the last thing the British Prime Minister needed as he enters a careful autumn of diplomacy was a Chinese spy scandal erupting in his midst. Basically, about a week and a half ago, two parliamentary researchers, both men, one in their 20s, one in their 30s, were arrested under the Official Secrets Act on suspicion of spying for the Chinese government. What's more, as Politico writes, one of the men is a well-known figure at Westminster who was involved in international affairs

and had close links to senior conservatives. One parliamentary aide who knows the man said, everyone who had come into contact with him is feeling angry and exposed. Another person involved in Westminster's China Policy Circle added, it's hard to trust anyone in the field now. This is a real shock. Everyone wonders what went wrong. So that is the climate into which the UK has decided to proceed inviting China to the AI summit. And since the announcement, they've basically been on the defensive, Politico again wrote a piece about how UK Chancellor Jeremy Hunt had been pushed to defend the decision to invite them to the summit. Hunt said in an interview,

if you're trying to create structures that make AI something that overall is a net benefit to humanity, then you can't just ignore the second biggest economy in the world. That doesn't mean that you make any kind of compromises with your values, but sometimes dialogue can be beneficial. Hunt and the Sunak government's argument is basically that A, they're important, they're big, they're a technology leader, and B, just because we talk to them doesn't mean they were going to embrace their approach to things like surveillance. Now, as an interesting aside, Jeremy Hunt is actually at the beginning of a trip to visit with a number of tech giants in places like Seattle,

San Francisco, and L.A. Now, Prime Minister Rishi Sunak has also tried to reinforce the message coming from other parts of his government, saying, we recognize that China is an authoritarian country and represents a significant challenge. And it's right only to engage when it's in the interest of the British public. And I think that's what you're seeing us do here. Now, in many ways, this has become a bit of a referendum in the UK on the Sunuk premiership's approach to China in general. UK news stations are filmed with current and former politicians, sharing their two cents about why China shouldn't be at the AI summit. And perhaps in response to some of that, the latest news is that the government may be trying to split the difference by keeping the Chinese invite intact, but banning

Chinese officials from parts of the conversation. The Guardian writes that Chinese officials might be only allowed to attend the first day of the two-day summit because of concerns over this espionage scandal. Now, we don't have a full agenda for the event, so we don't know exactly what it would mean for them to come only to one day versus the second, but presumably you might expect to see less deep technical conversations on that day that China was invited to and more general global cooperation type conversations. Now, to be clear, China hasn't even accepted this invite yet, and to the extent that they did, it's not clear who would come. Right now, expected attendees include U.S. Vice President Kamala Harris, French President Emmanuel Macron,

Canadian Prime Minister Justin Trudeau, and tech leaders, although people like Zuckerberg and Elon aren't confirmed yet. Now, when it comes to wider conversation, at least in the AI safety side, the decision to include China has a lot of backup. Future of Life Institute president Max Tegmark writes, the UK Bletchley Park Summit is a huge opportunity for AI safety if it doesn't turn into big tech's regulatory capture 2023 or a geopolitical pissing contest by state. loving China. It also needs to cover the full spectrum of AI risk from current harms to human extinction. The Future of Life Institute has actually published a set of recommendations for the summit, which may end up being part of this week's long read. But when it comes to China, some of the recommendations include

one involving all major AI powers in the summit, including Brazil, China, the EU, India, Japan, and the US. Two, to the People's Republic of China, they recommend, quote, engage in an open dialogue with the US despite ongoing economic and strategic competition. Global threats from advanced AI, they write, much like climate change, urgently demand cooperation, even if this is not possible on most bilateral issues. And of course, they say the same things directed toward the United States. So we will see how this shakes out. It's not clear yet if the UK government is going to go through with this sort of half-and-half approach where China is invited to the first day, but not the second day. It's also not clear if China is going to accept the invite. But it does certainly reinforce how the existing geopolitical

tension between China and the U.S. in particular, but really the West more broadly, impact the world when it comes to these issues of global importance. An opinion piece in the South China Morning Post from August read, World cannot afford AI cooperation falling prey to U.S.-China geopolitical strife. Geopolitics is a threat to international cooperation on AI legislation, resulting in patchwork strategies competing goals and disjointed global standards. So, friends, this is something that we will have to keep an eye closely on, but for now, that is going to do it for today's AI breakdown.

Thanks, as always, for listening or watching, and until next time, peace.