According to a report from Reuters, the U.S. government is finalizing regulations that will prohibit investments in AI in China, as indicated by a recent government update. The regulation will restrict specific outbound investments to China in areas such as AI, semiconductors, microelectronics, and quantum computing.

According to the report, the rule is currently under review at the Office of Management and Budget, suggesting that it is expected to be released within the next week or so.

The rules will reportedly require U.S. investors to notify the Treasury Department about certain investments in AI and other sensitive technologies in China.

The report highlighted that the rules are based on President Joe Biden’s executive order from August 2023, aimed at safeguarding American knowledge and preventing its application in China’s military advancements.

According to the proposals released last year, the U.S. Treasury Department highlighted that the military, intelligence, surveillance, and cyber-enabled uses of these technologies and products pose risks to U.S. national security, particularly when developed by countries of concern such as the PRC.

Citing former Treasury official Laura Black, the report suggested that the rule is expected to be released before the U.S. election. Black also noted that the Treasury office responsible for overseeing regulations typically allows for a minimum 30-day period before the actual implementation of the rule.

According to the report, the Treasury Department released proposed rules, including some exceptions, and invited public comments in June. Black expected that the final rules will provide further clarity on the scope of coverage for AI and the thresholds for limited partners.

Additionally, the report noted that the proposed exceptions include publicly traded securities such as index funds and mutual funds, along with certain limited partnership investments and specific syndicated debt financings.

Read more

• [News] U.S. Urges South Korea to Tighten Chip Export Controls to China
• [News] China Bypasses Restrictions to Acquire NVIDIA Chips, with Cloud Service Costs Even Lower Than in the U.S.

(Photo credit: istock)

Apple introduced the Apple Intelligence feature at WWDC24 back in June. However, according to Mark Gurman from Bloomberg, some Apple employees believe the company is two years behind its competitors in AI development.

The Apple Intelligence includes several features, such as AI notification summaries, intelligent prioritization for important alerts, a new Siri with personal context, Image Playground, Genmoji, and more.

According to a report from 9to5Mac, one of the main challenges facing Apple Intelligence is that it mostly depends on models that can run on-device, which also means that Apple Intelligence has quite high requirements for device.

To run Apple Intelligence, users will need a device equipped with at least an A17 or M1 chipset, and it must have a minimum of 8 GB of memory to support the personal intelligence system, as highlighted by 9to5Mac.

The report from 9to5Mac noted that since the models primarily operate on-device, the volume of information they can process is inherently limited. This is also why Apple has chosen to integrate Siri with ChatGPT.

According to the announcement by OpenAI, Apple will support GPT-4o across iOS, iPadOS, and macOS. GPT-4o will be integrated into Siri and Writing Tools, with the expectation that this integration will help reduce the knowledge gap.

Regarding the knowledge gap, according to Mark Gurman from Bloomberg, Apple’s internal research shows that ChatGPT is about 25% more accurate than Siri and can answer about 30% more questions, leading some Apple employees to believe the company is two years behind its competitors in AI development.

Nonetheless, Gurman believes that Apple still has the potential to catch up, as they can often successfully catch up with competitors in areas where they seem to be lagging behind, demonstrated by its experience with Apple Maps.

Furthermore, Gurman also pointed out that Apple has another advantage in its efforts to catch up: the ability to deploy features across a vast array of devices. Apple can quickly equip its current products with the technology required to support new software. We can expect to see this again soon with the rollout of the M4 Macs, which will accelerate Apple’s AI tasks, as indicated by Gurman.

According to Gurman, Apple Intelligence will be available on nearly every Apple device with a screen by 2026. The iPhone SE is set to receive this feature in March, along with the A18 chip, while the entry-level iPad is expected to be updated later in 2025.

Read more

• [News] Apple Intelligence Is Expected to Arrive on October 28th with the Release of iOS 18.1
• [News] iPhone 16 “Supercycle” in Doubt? Concerns Rise Over Slow Rollout with its AI Platform

(Photo credit: Apple)

Amid the wave of global AI investments, China’s TikTok has confirmed the layoff of employees from its Malaysia division. According to a report by Reuters, over 700 employees have been laid off as the company pivots to increased use of AI for content moderation.

The information was first revealed by local media the Malaysian Reserve last Thursday, as the report notes that over 500 individuals were let go after receiving emails from the company. Most of the workers, according to the report, are said to be involved in content moderation on the platform.

The job cuts come as global tech companies are facing growing regulatory pressure in Malaysia, where the government mandates social media operators to apply for an operating license to combat cyber offenses.

According to a previous report by the South China Morning Post, social media and internet messaging platforms with over 8 million users in Malaysia must obtain a license by New Year’s Eve next year or face penalties under current cyber laws.

It is worth noting that more layoffs might be on the way, according to Reuters. Citing sources familiar with the situation, Reuters suggests that the Chinese tech giant is planning further layoffs next month as it aims to streamline some of its regional operations.

In response to inquiries from Reuters, TikTok confirmed the layoffs last Friday, but refused to provide an exact number of employees affected in Malaysia. According to a TikTok statement cited by Reuters, the company is implementing these changes as part of its ongoing efforts to enhance the global operating model for content moderation.

TikTok utilizes a combination of automated detection and human moderators to review the content posted on its platform, Reuters notes.

Furthermore, the Chinese tech giant reportedly plans to invest USD 2 billion in trust and safety this year and will continue to focus on improving efficiency, noting that 80% of content that violates guidelines is now removed by automated technologies, according to Reuters.

According to the company’s website, ByteDance, TikTok’s parent company, has over 150,000 employees based out of nearly 120 cities worldwide.

Read more

• [News] Despite Geopolitical Tensions, China’s AI Startups Break into the U.S. with Products like Chatbots
• [News] Broadcom Reportedly Developing 5nm AI Chip with ByteDance, Manufactured by TSMC

(Photo credit: TikTok)

In light of slow revenue growth in the Chinese market, China’s AI startups are seeking to enter the U.S. market in pursuit of overseas growth opportunities, following the successful model of the short video platform TikTok, according to a report by Commercial Times.

According to a report by the Financial Times, despite U.S. restrictions on chip exports to China, the country still maintains a competitive advantage in launching products like chatbots that do not require substantial computing resources.

The report highlights that Chinese AI companies, including MiniMax, TikTok’s parent company ByteDance, and 01.AI, have introduced AI products overseas, particularly targeting the U.S. market, which boasts a larger base of high-end consumer users.

These Chinese app companies have seen significant success. For instance, the report in Commercial Times points out that the majority of MiniMax’s sales stem from its chatbot application “Talkie,” which has gained immense popularity among American teenagers.

Shanghai-based MiniMax has made major breakthroughs in the past year, and the company plans to reach a sales target of USD 70 million in 2024, Commercial Times notes.

The Financial Times report also highlights that Chinese company ByteDance has launched several AI apps internationally. ByteDance’s photo editing application “Hypic,” along with Zuoyebang’s homework assistant “Question AI,” both made it into the top 20 downloads internationally.

However, according to the Financial Times report, Chinese companies still face certain challenges. Firstly, the opportunities for revenue growth are limited due to the high costs associated with training language models, which can negatively impact the potential of these companies. Additionally, to mitigate external risks, Chinese firms need to take measures such as placing servers outside of China to avoid the potential crisis of TikTok being banned in the United States.

Lisa Su of AMD recently celebrated a decade as CEO, a remarkable journey that has transformed the company from a challenging position into one of the most influential players in the AI landscape.

10 years ago I had the honor of a lifetime to be named @AMD CEO. It’s been an incredible journey with so many proud moments. Today I want to thank the global @AMD team for all you do. As amazing as the last 10 years have been, the best is yet to come. 🥰 pic.twitter.com/WuQRt3xlX3

— Lisa Su (@LisaSu) October 8, 2024

When Lisa Su first joined AMD, the company was on the brink of bankruptcy, having laid off nearly a quarter of its workforce, while its stock price had plummeted to just USD 2.

Su’s achievements at AMD are numerous, particularly her revamp of the product portfolio. The reforms she implemented as CEO significantly shaped the company’s future success.

Notably, as reported by CNN, when Su first took the position as CEO of AMD, she decided not to pursue technology for mobile phones or Internet of Things sensors, recognizing that while these areas are promising, they are not AMD’s core competencies.

(Photo Credit: AMD)