Welcome to Get The Gist, where every weekday we share an easy-to-read summary of the latest and greatest developments in AI—news, innovations, and trends—all delivered in under 5 minutes! ⏱

In today’s edition:

• Microsoft Upgrades Copilot with Voice and Reasoning

• Nvidia’s Game-Changer in Open-Source AI

• China's Telecom Giant Unleashes TeleChat2-115B AI Model

• Microsoft Paint Gets AI-Powered Features

• And more AI news….

Top Developments

Copilot

Microsoft Upgrades Copilot with Voice and Reasoning to Rival Google AI

The Gist: Microsoft's AI assistant, Copilot, just got a major update, adding a conversational voice and advanced web analysis features as it aims to compete with Google and OpenAI in the consumer AI space.

Key Details:

• Copilot now includes voice features, making interactions feel more natural and engaging.

• A new "Think Deeper" tool helps users reason through decisions, enhancing its problem-solving capabilities.

• Expanded web analysis allows Copilot to interact more deeply with online content as users browse.

• Microsoft envisions Copilot becoming an all-around digital companion across platforms like Word, Edge, and Xbox.

Nvidia

Nvidia’s Game-Changer in Open-Source AI

The Gist: Nvidia has launched NVLM 1.0, a powerful open-source AI model that competes with top proprietary systems like those from OpenAI and Google, opening doors for researchers and developers everywhere.

Key Details:

• The NVLM-D-72B model excels in both visual and text-based tasks, outperforming many existing models.

• Nvidia’s choice to release this model openly could democratize access to cutting-edge AI technology, allowing smaller entities and independent researchers to flourish.

• Positive reactions from the AI community hint at the potential for rapid advancements in research and development following this release.

• While promising innovation, the model's availability also brings concerns about ethical use and potential misuse, prompting discussions on responsible AI practices.

China Telecom

China's Telecom Giant Unleashes TeleChat2-115B AI Model

The Gist: China Telecom has rolled out a groundbreaking AI model—TeleChat2-115B—with an impressive 100 billion parameters, all developed using domestic resources. This marks a significant step in China's tech landscape, showcasing its resilience in the face of international tech restrictions.

Key Details:

• The model was trained on a colossal dataset of 10 trillion tokens, pulled from high-quality texts in Chinese and English.

• It relies on advanced processors from Huawei, known for their efficiency and sophisticated manufacturing.

• While TeleChat2-115B has fewer parameters than some global counterparts, its performance suggests that parameter count isn't the only measure of effectiveness.

• This development highlights China's commitment to AI innovation, regardless of external tech export sanctions.

Quick Gist

Microsoft Paint is revamping its legacy app with new AI tools like Generative Fill and Generative Erase, enabling advanced image editing features similar to those found in Adobe Photoshop (Read More).

ByteDance is planning to acquire over 100,000 Huawei AI chips to reduce its dependence on Nvidia for training new AI models, despite facing supply challenges (Read More).

At DevDay 2024, OpenAI launched new developer tools including a Realtime API and vision fine-tuning for GPT-4o, aimed at enhancing AI application development while reducing costs (Read More).

Google updated Gemini to simplify image sharing from other apps, adding a direct share option, making it easier for users to interact with the AI (Read More).

Durk Kingma, a former OpenAI co-founder and research scientist at Google DeepMind, has joined Anthropic, continuing a trend of ex-OpenAI staff moving to the safety-focused AI start-up (Read More).

The European Commission appointed a 13-member team to draft general purpose AI rules under the AI Act, set to be enforced in 2025, focusing on transparency, copyrights, risk assessment, and governance (Read More).

Developers can now test Google’s Gemini Nano AI model on Pixel 9 device (Read More).

OpenAI’s new updates enhance voice and vision capabilities of AI (Read More).

THAT’S IT FOR TODAY!