Adobe Express has long been a vital tool for TikTok creators, now with TikTok's new Creative Assistant integration, optimizing content becomes easier. The AI chatbot offers brainstorming assistance, trend suggestions, and successful advertising cases, enhancing the creative workflow.

Adobe introduces new AI assistant in Acrobat PDF for document summarization, answers, formatting, citations, and navigation. Available for trial by Acrobat Pro, Teams, and individual users. Plans for AI insights, creation, and collaboration tools. Data governance and user consent emphasized. Subscription users get early access in English, with more languages and Reader support to come.

French startup Bioptimus secures $35M funding to develop AI for biology. Led by Jean-Philippe Vert, it aims to create universal AI base model integrating biological data at all levels. Partners include Sofinnova, Bpifrance, Cathay Innovation. Mission: accelerate biomedical innovation with AI models. Talent from Google DeepMind, Owkin. Potential to impact personalized medicine, clinical trials, and treatment design. Funding to drive exponential growth in high-potential field.

An AI model by Groq, using LPU architecture, shows exceptional speed, outperforming traditional GPU-based models like OpenAI's Chat GPT. Groq achieves up to 500 tokens per second, far surpassing GPT 3.5's 30 to 50 tokens.

Approximately one-quarter of Americans get news from YouTube, a key online platform. Study shows users' political interests influence content choice more than YouTube's recommendation algorithm. Bots analysis indicates algorithm's mild impact in consuming less partisan content.

CodeCompose, an AI-driven tool used at Meta, achieved a breakthrough in multi-line suggestions. Latency improved by 2.5 times with range-based algorithms. Challenges addressed include jarring effects and responsive user experience. Production release impact was closely monitored. Developers saw increased coding speed and potential value in improving development efficiency.

Deutsche Telekom, Qualcomm, and Brain.AI are developing an AI smartphone to create an application-less intelligent device. The AI will predict and generate interfaces seamlessly based on users' intentions. This concept will be presented at the 2024 World Mobile Congress in Barcelona.

Google plans to launch Gemini Enterprise and Gemini Business for Workspace users, ensuring data security and allowing direct Python code editing in Gemini user interface, surpassing other models like ChatGPT and Claude.

Clubhouse introduces text-to-speech feature where users can send text messages in their custom voice. Group voice chat and AI voice reconstruction aim to retain users. Feature only available in the US.

UK-based Graphcore Ltd., an AI chip startup, explores sale due to competition struggles. Recent valuation at $2.77 billion. Challenges in AI chip market.

Google warns of privacy risks with AI apps, as Alphabet's Bard becomes Gemini on Android. Gemini collects conversations, location, feedback, and usage data, stored for 3 years. Users advised not to share sensitive info.

NVIDIA showcases the architecture of its latest enterprise supercomputer, Eos, designed for advanced AI development at scale. Eos integrates 1152 Intel Xeon Platinum 8480C processors and 4608 H100 GPUs for impressive HPC and AI performance. The system uses DGX SuperPOD architecture with Mellanox Quantum-2 InfiniBand for high-speed data transfer. Eos also includes a software stack optimized for AI tasks, making it a powerful tool for enterprises pursuing AI projects.

Argmax released WhisperKit software for Apple Watch, integrating OpenAI's Whisper model with CoreML. Developers can customize audio formats and models, enhancing speech recognition capabilities on WatchOS. WhisperKit promotes widespread use of speech recognition technology in Apple devices.

ElevenLabs, an AI startup, is launching new text-to-speech models, allowing creators to generate sound effects by describing their imagination in text. The model enriches digital content in new ways and is expected to revolutionize AI-generated content with AI sound effects beyond voice.

Slack AI, an upgraded version of Salesforce's team communication platform, aims to bring more powerful functionality to workplace conversations and project planning with features like search answers and thread summaries. Denise Dresser, Slack's CEO, highlighted how AI will help users access collective knowledge and work smarter. The AI-driven "Enterprise Search" tool allows users to easily find relevant content in Slack messages.

ByteDance introduced Boximator, a cutting-edge technology for precise object motion control in videos. Users outline objects in reference images with boxes to define motion paths, enhancing video quality and achieving precise motion control. This user-friendly approach improves video generation quality and efficiency.

Last week, OpenAI's new video generation model, Sora, received both praise and criticism from AI experts. Competitors Meta and Google questioned Sora's understanding of physics. Meta's CEO emphasized the importance of predictive models, highlighting their V-JEPA's superiority over Sora. NVIDIA explained Sora's physics engine and acknowledged its potential despite some limitations. The future impact of Sora on the video game industry remains promising.

Stability AI introduces Stable Cascade, a text-to-image model with exceptional quality and hardware efficiency. The model compresses latent space for training on consumer GPUs, offering unprecedented image generation and transformation capabilities. With modular design and fine-tuning options, Stable Cascade provides high-quality outputs while reducing training costs by 16 times. Users can explore creative possibilities and access training scripts on GitHub for specialized applications.

Google's Gemini 1.5 Pro raises doubts about the authenticity of OpenAI's Sora-generated video due to significant inconsistencies, sparking a detailed analysis and criticism from Google. Gemini 1.5 Pro dissects a scene showing snow and cherry blossoms in Japan, highlighting unnatural elements that question the video's realism. Sora, a new text-to-video tool, is praised for its intricate scenes and emotional characters, while Gemini 1.5 boasts a large context window and impressive capabilities.

Sora, a revolutionary video generation model, signifies AI advancements. Zhou Hongyi believes Sora could expedite AGI realization and emphasizes human creativity's importance in video production. Sora's key advantage lies in understanding and simulating the real world, propelling AI towards AGI. Zhou mentions potential applications in various fields and highlights the gap between China and the US in AI development.