August 31, 2024: Amazon Acquires Covariant Founders and Talent - Amazon has hired the founders of AI robotics startup CovariantPieter Abbeel, Peter Chen, and Rocky Duanalong with about 25% of its employees. The company also acquired a non-exclusive license for Covariants robotic AI models, aimed at improving warehouse tasks. This move parallels a similar hiring strategy used in June with AI startup Adept, allowing Amazon to gain new technology and expertise amidst antitrust scrutiny. Covariant will continue operations under new CEO Ted Stinson, focusing on industries such as apparel, health, and pharmaceuticals.

August 31, 2024: NightCafe Thrives as AI Art Hub - NightCafe, co-founded by Elle and Angus Russell, provides AI-generated art tools without in-house AI capabilities, profiting by aggregating top models like Stable Diffusion and DALL-E alternatives. The platform boasts 25 million users and generates $4 million annually. Emphasizing community and accessibility, it hosts art challenges and offers fine-tuning. Despite facing moderation challenges and potential legal issues, NightCafe continues to succeed by focusing on UI, community engagement, and providing diverse model access at competitive prices.

August 30, 2024: LAION Cleans CSAM from Training Dataset - LAION re-released its dataset as Re-LAION-5B, removing child sexual abuse materials (CSAM) and more, following recommendations from various organizations. The new dataset, available in standard and Safe versions, targets research use. This follows findings from a Stanford report which led to temporary removal of LAION-5B. LAION urges migration to the updated dataset for remaining users.

August 30, 2024: Amazon Taps Claude LLM for Alexa Upgrade - Amazon plans to launch a premium version of Alexa, powered by Anthropics Claude LLM series, in October. Named Remarkable Alexa, this paid service will remember user interactions and autonomously perform tasks such as ordering food and setting alarms. It may use multiple Claude models to balance performance and cost, with pricing anticipated between $5 to $10 per month.

August 30, 2024: Google Adds AI Safeguards for 2024 Election - Google is implementing additional safeguards on its generative AI products, such as Search AI Overviews and YouTube AI-generated summaries, to prevent the spread of misinformation ahead of the 2024 U.S. presidential election. The restrictions will prevent AI from engaging with election-related topics. Additionally, Google Search will feature voter registration information, and YouTube will highlight credible election information and voting reminders. Google Play will also mark apps from official government agencies to ensure reliable election data.

August 30, 2024: Alibaba Unveils Advanced Qwen2-VL AI Model - Alibaba Cloud announced Qwen2-VL, an AI model that excels in video analysis and multilingual conversations. Capable of understanding videos over 20 minutes long, it can summarize content, answer questions, and support live chat. Utilizing the Vision Transformer model with 600 million parameters and Multimodal Rotary Position Embedding, it outperforms many closed-source models. However, it cant extract audio and is trained up to June 2023. Available open-source, Qwen2-VL is a vital step toward comprehensive vision-audio AI models.

August 29, 2024: IBM Cloud Adopts Intels Gaudi 3 AI Chips - IBM Cloud will offer Intels Gaudi 3 AI accelerators early next year, incorporating them into its WatsonxAI and data platform. Gaudi 3 aims to rival Nvidia and AMDs AI chips, featuring new Ethernet connectivity and pairing with Intels Xeon 6 processors. Despite promising benchmarks, Intel faces challenges against Nvidias dominance and financial setbacks, including a $1.6 billion Q2 loss and upcoming job cuts.

August 29, 2024: Codeium Raises $150M, Hits $1.25B Valuation - Codeium, a GitHub Copilot competitor, raised $150M in a Series C round led by General Catalyst, valuing it at $1.25B. Founded in 2021, the startup now boasts 700,000 users and 1,000 enterprise customers. Codeiums AI-powered coding assistant supports 70 programming languages and integrates with popular development environments. With advanced data security measures and a client-centric approach, Codeium aims to reduce coding inefficiencies and enhance code security. Revenue hit eight figures this year, and the startup plans to expand its team to 120 by 2025.

August 29, 2024: Llama AI Models Reach 350M Downloads - Metas Llama AI models have surged in popularity, nearing 350 million downloads, a tenfold increase over last year. Following the release of Llama 3 and 3.1, monthly downloads hit 20 million, and token usage soared ten times from January to July. The models, praised for their multilingual and diverse capabilities, are driving innovation in startups and enterprises alike. The open-source approach has led to 60,000 derivative models on Hugging Face, with corporate adoption from major players like AT&T, Goldman Sachs, and Zoom.

August 29, 2024: IBM Integrates Intels Gaudi 3 AI Chip - IBM will offer Intels Gaudi 3 AI processors in its public cloud early next year. Positioned as a rival to Nvidias H100 GPU, Gaudi 3 features enhanced matrix multiplication and parallel computation capabilities, promising improved efficiency and speed, especially in large language models. IBM will integrate Gaudi 3 into its Cloud Virtual Servers for VPC and watsonx tools, enhancing AI model development and deployment.

August 29, 2024: Nvidia, Apple Eye OpenAI Funding Round - Nvidia and Apple may join OpenAIs multibillion-dollar funding round, potentially boosting the companys valuation beyond $100 billion. Thrive Capital will lead with a $1 billion investment, while Microsoft might increase its stake. Apple plans to integrate ChatGPT into its devices, while Nvidia looks to enhance on-premises AI deployments through potential technical collaborations. OpenAI also explores custom machine learning accelerators to diversify chip suppliers.

August 28, 2024: Adobe Unveils AI-Powered Workfront Update for Marketers - Adobe announced a major update to Workfront, introducing Workfront Planning with AI capabilities to centralize and streamline marketing workflows. The update offers visual tools for campaign planning, timeline mapping, and data visualization. A new AI assistant can create dynamic campaign records, marketing briefs, and auto-update plans based on user input.

August 28, 2024: Agency: Debugging Tools for AI Agents - San Francisco startup Agency, founded from AI hackathons, offers AgentOps, a platform for monitoring and debugging AI agents. Co-founders Alex Reibman, Adam Silverman, and Shawn Qiu shifted focus from building AI agents to developing tools to improve their reliability. AgentOps helps teams understand and manage AI behavior. The startup, now with $2.6 million in pre-seed funding, partners with AI developers like Cohere and Mistral.

August 28, 2024: Midjourney Ventures into Hardware with New Team - Midjourney, the AI image platform earning over $200 million without VC backing, announces its entry into hardware. The new team based in San Francisco includes Ahmad Abbas, an ex-Neuralink and Apple Vision Pro engineer, and CEO David Holz, co-founder of Leap Motion. Despite ongoing lawsuits over its AI training methods, Midjourney continues to develop AI models, hinting that the hardware could support video and 3D generation efforts.

August 28, 2024: Google Restores Geminis People-Generating Feature - Google has reintroduced Geminis people-generating feature for users on premium plans after addressing historical inaccuracies. The updated model, Imagen 3, incorporates diverse training data and safety filters. While regular Gemini users also get access to Imagen 3, they wont receive people generation yet. Gemini now includes Gems, custom AI experts for various tasks, available in 150 countries.

August 28, 2024: Midjourney Announces Launch into AI Hardware Sector - Midjourney, renowned for its AI image generation, has announced its venture into hardware devices. The company is building a team in San Francisco, led by Ahmad Abbas from Apples Vision Pro team. The device will be unique and non-wearable, possibly an orb form factor. This move comes amid rising competition in AI image models and legal challenges. Despite previous AI hardware failures, Midjourney aims to innovate and push the boundaries of current form factors.

August 28, 2024: Apples AI Clean Up Tool Transforms Photos - Apples iOS 18.1 update introduces Clean Up, an AI-powered feature for removing objects from photos without altering the background, including handling shadows and reflections. Users can easily select or brush over unwanted objects to delete them. This new tool follows Apples roll-out of various Apple Intelligence features, rivaling Googles Magic Eraser.

August 27, 2024: Fortinet Expands SASE with AI and Sovereignty - Fortinet announced updates to its unified SASE solution, adding sovereign options and generative AI integration. The new features enhance data control for regulated industries and optimize SD-WAN management, reducing complexity and operational burdens. The updates offer a single operating system, console, and data lake for streamlined security and visibility, positioning Fortinet as a leader in truly unified SASE solutions.

August 27, 2024: Googles AI Struggles with Hindi Overviews - Googles AI overviews in Hindi are suffering from inconsistencies and inaccuracies, ranging from poor translations to misleading answers. The tools inability to properly interpret and summarize web content has led to errors, such as incorrect information about food and holidays, and culturally sensitive topics like menstruation and pregnancy. Despite Googles claim that the feature is rooted in its core search quality systems, the results are disappointing, especially in comparison to English queries. With over 830 million internet users in India, Google must enhance its AIs language and content accuracy to maintain trust.

August 27, 2024: Nvidia Unveils AI Blueprints for Enterprise Apps - Nvidia has launched NIM Agent Blueprints to help enterprises build generative AI apps easily. These blueprints offer ready-to-use templates and tools for applications such as customer service chatbots, PDF data extraction, and drug discovery, aimed at accelerating app development and deployment. Nvidias AI Enterprise toolkit, including NeMo and NIM microservices, supports this initiative. Initial partners include Accenture, Cisco, Dell, and others. This move aligns with Gartners forecast that 80% of conversational offerings will embed generative AI by 2025.