🔥New AI Chip Collab

China's ByteDance partners with Broadcom to develop advanced AI chip

Key Points 

• The collaboration is expected to reduce procurement costs and ensure a steady supply of high-end chips.

• The processor, a 5-nanometer application-specific integrated chip (ASIC), will comply with U.S. export restrictions and be manufactured by Taiwan's TSMC.

☕News - China's ByteDance is collaborating with U.S. chip designer Broadcom to develop an advanced AI processor, according to two sources familiar with the matter. This effort aims to ensure that the TikTok parent company has enough high-end chips despite U.S.-China tensions.

The processor, a 5-nanometer application-specific integrated chip (ASIC), will comply with U.S. export restrictions and be manufactured by Taiwan's TSMC. Notably, this is the first known collaboration between Chinese and U.S. companies on 5nm technology since the U.S. imposed export controls on advanced semiconductors in 2022, with previous deals typically involving less advanced tech.

ByteDance's collaboration with Broadcom is expected to reduce procurement costs and ensure a steady supply of high-end chips, the sources said. However, TSMC will not start manufacturing the new chip this year, as the design phase is still ongoing and the "tapeout" has not yet begun.

👨🏻‍💻What's more? Besides TikTok and its Chinese counterpart Douyin, ByteDance also runs several popular apps, including Doubao, a chatbot service with 26 million users. To support its AI initiatives, ByteDance has been stockpiling Nvidia chips, a separate source mentioned.

New from Synthesia: Make Apple-style presentations with AI avatars using only your webcam or phone 

Key Points 

• The company's new product updates will allow users to create Apple-style key presentations with AI avatars using just a laptop webcam or phone.

🎥News - On Monday, Synthesia, a British AI startup backed by Nvidia, unveiled several new product updates. These updates will allow users to create Apple-style key presentations with AI avatars using just a laptop webcam or phone.

According to the company, the updates are aimed at transforming its platform into a comprehensive video production suite for large companies, moving beyond its initial focus on AI-generated avatars.

🤩So what's new? 

1. AI Avatars from Webcams or Phones: Users can now create AI avatars by recording less than five minutes of footage using a webcam or phone, eliminating the need for in-person studio visits.

2. Full-Body Avatars: Synthesia now offers full-body avatars that can move their hands and arms, unlike the previous portrait-only avatars. However, these avatars need to be created in a studio with multiple cameras, sensors, and lights.

3. Language and Voice Cloning: AI avatars can now speak in multiple languages, including English, French, German, and Chinese. Users can also clone their voices for avatars to use in different languages.

4. AI Video Assistant: A new AI video assistant can generate summaries of entire articles and documents, useful for creating quick explanatory videos, like company benefits packages.

5. Screen Recording Tool: A new screen recording feature allows AI avatars to guide users through on-screen content.

Future updates will enable the creation of region-specific avatars, such as a Nigerian avatar for regional tutorials.

🤓Interested in experimenting with AI avatars? Here are some applications to consider:

• Enhance customer support by creating interactive tutorials and troubleshooting guides.

• Generate virtual presenters for online events or conferences to deliver engaging presentations.

• Create interactive educational content for students, making learning more immersive and accessible.