Elon Musk Adds Microsoft to Lawsuit Alleging OpenAI’s Shift to Profit-Driven Model

A Thought Leadership platform to help the world navigate towards Artificial General Intelligence We are committed to navigate the path towards Artificial General Intelligence (AGI) by building a community of innovators, thinkers, and AI enthusiasts.

Introducing the GEN Matrix: Your Essential Guide to Generative AI Trailblazers!

Dive into the forefront of Generative AI with the GEN Matrix—your ultimate resource for discovering the innovators, startups, and organizations leading the AI revolution. Our platform features three categories spotlighting:

• Organizations: Early adopters advancing GenAI in production.

• Startups: Pioneers across diverse GenAI layers (chips, infrastructure, applications, etc.).

• Leaders: Key figures driving GenAI innovation and adoption.

Know someone making strides in GenAI? Nominate them to be featured in the GEN Matrix! Whether you're a business seeking AI solutions or a developer looking for tools, explore GEN Matrix to stay at the forefront of AI excellence.

TheGen.AI News

Elon Musk Adds Microsoft to Lawsuit Alleging OpenAI’s Shift to Profit-Driven Model

Elon Musk, known as the U.S. government's efficiency czar, has expanded his lawsuit against OpenAI, now led by Sam Altman, accusing it of abandoning its non-profit origins. The amended complaint, filed in a California district court, adds new defendants, including Microsoft, LinkedIn co-founder Reid Hoffman, and former OpenAI board member and senior Microsoft executive Dee Templeton. Additionally, new plaintiffs have been named, such as Neuralink executive and former OpenAI board member Shivon Zilis, along with Musk's AI company, xAI.

The lawsuit alleges that OpenAI, originally founded by Musk as a non-profit dedicated to safety and transparency, has transformed under Altman, Brockman, and Microsoft into a for-profit subsidiary of Microsoft. Musk's legal team claims OpenAI is “actively working to eliminate competitors,” including xAI, by pressuring investors not to fund rival projects.

The filing accuses OpenAI of orchestrating a transition that took it from a tax-exempt charity to a $157 billion for-profit powerhouse in just eight years, labeling this shift as unprecedented and unlawful. The lawsuit claims this transformation involved misleading donors, members, regulators, and the public.

Zilis, who stepped down from OpenAI’s board in 2023, is cited in the lawsuit as an “injured employee” under California Corporations Code. Zilis has close ties to Musk, having previously worked as a project director at Tesla and as a research director at Neuralink.

The complaint further alleges that OpenAI’s partnership with Microsoft laid the foundation for its for-profit transition. It claims that this collaboration created a web of opaque, for-profit affiliates that siphoned intellectual property, employees, and relationships originally developed under OpenAI’s non-profit status, using Musk’s name and contributions to bolster its reputation and philanthropic image.

While Musk has expressed respect for Microsoft CEO Satya Nadella, he argued that Microsoft’s and OpenAI’s values diverge from his vision. Musk believes AI should be decentralized and open, given its existential risks, while Nadella and Microsoft co-founder Bill Gates reportedly downplayed Musk’s concerns as overblown and premature. Microsoft has yet to issue a statement regarding the amended lawsuit.

InVideo Unveils GenAI Video Creation Tool Backed by Tiger Global

Indian video editing platform InVideo is introducing a generative AI-powered feature for video creation, allowing users to generate video clips using simple prompts. Branded as InVideo v3.0, this new release enables users to create videos in various styles, such as live-action, animated, or anime. The platform also supports exporting videos in formats optimized for YouTube, Shorts/Reels, and LinkedIn. Users can refine their videos by prompting specific edits or additions to particular parts of a clip.

Rather than developing its own AI model, InVideo leverages a pipeline of existing models to interpret prompts and generate videos. According to co-founder and CEO Sanket Shah, while there’s no theoretical limit to video length, users are restricted by the number of generation credits available. The company does not offer a free tier.

To fund this feature, InVideo has introduced a Generative Plan, priced at $120 per month for 15 minutes of video generation. Additional minutes can be purchased for $8 to $10 per minute across all plans.

Previously, InVideo had launched a prompt-based feature that combined licensed stock footage and images into videos, but its output appeared rudimentary. The new tool, while more advanced and capable of producing videos several minutes long, still displays inconsistencies in style and quality within the same clip. The startup acknowledged this and is working to enhance the tool’s capabilities.

InVideo competes with platforms like Captions, D-ID, and Lightricks, which also offer AI-powered video creation tools. Unlike its competitors, which often cater to filmmakers, ad agencies, and sales teams, InVideo targets individuals, content creators, and small businesses. The company highlights its user-friendly interface as a key differentiator.

Currently, InVideo boasts 4 million monthly active users and reports that 7 million videos were generated in the past month. Shah noted that the platform’s AI features are designed to inspire users to imagine and create more ambitious videos, ultimately encouraging them to invest in premium AI tools.

The company raised $35 million in funding in 2021 from investors like Greenoaks, Tiger Global, and Peak XV and still retains $25 million in funds, with minimal cash burn. Shah stated that InVideo is on track to generate $50 million in revenue this year and is exploring potential partnerships with new investors in the coming quarters.

“InVideo v3.0 represents a bold leap towards our mission,” Shah remarked, emphasizing the company’s long-term vision and sustained investor interest.

Gen AI Users Ready to Pay Premium for High-Speed, Secure Connectivity

A report by Ericsson reveals that one in four users of generative AI (GenAI) applications are willing to pay up to 35% more for fast, secure connectivity to support high-capacity applications.

The study predicts a 2.5-fold increase in the number of smartphone users accessing generative AI apps weekly over the next five years. Emerging markets like India, Thailand, and Saudi Arabia demonstrate a much stronger interest in differentiated connectivity than countries such as France and Spain.

Rising Demand for Enhanced Connectivity

Jasmeet Sethi, head of ConsumerLab at Ericsson, noted that the growing adoption of AI-powered applications is driving user demand for better connectivity. This reflects user expectations for advanced capabilities in areas like image, audio, or video generation and their willingness to invest in reliable, high-quality connectivity to support these features.

Opportunities for Communication Service Providers (CSPs)

The report highlights a significant opportunity for Communication Service Providers (CSPs) as users increasingly seek faster and more reliable connections for AI applications. CSPs adopting performance-based business models could benefit from tailored subscriptions, potentially boosting 5G ARPU (average revenue per user) by 5-12%.

Sethi emphasized that one in three 5G smartphone users are prepared to shift 10% of their mobile app budget toward apps with built-in elevated connectivity. By leveraging Quality on Demand (QoD) network APIs, CSPs can enable premium app experiences, creating new revenue streams through high-performance applications.

Key Findings

1. Willingness to pay more: Globally, 35% of 5G users are willing to pay extra for guaranteed performance connectivity.

2. Assurance seekers: A segment comprising 20% of users, termed ‘assurance seekers,’ actively seeks superior connectivity for critical applications and is ready to pay a premium.

3. Generative AI app growth: Weekly usage of generative AI apps is expected to increase 2.5 times over the next five years. A quarter of current users are already open to paying 35% more for responsive connectivity.

4. Regional trends: Countries like India, Thailand, and Saudi Arabia have twice the number of smartphone users interested in differentiated connectivity compared to France and Spain.

5. CSP evolution: The report outlines a five-stage roadmap for CSPs to transition from standard mobile broadband to performance-driven, platform-based models. Network APIs are identified as key enablers for developers to build customized app experiences.

As demand for AI-powered applications rises, CSPs are well-positioned to capitalize on this trend by offering differentiated connectivity solutions, opening the door to new revenue opportunities while meeting evolving consumer expectations.

TikTok Unveils Gen AI Video Tools to Simplify Brand Marketing

TikTok is introducing a new AI-driven video generation feature that can create TikTok-style video clips using a product description or URL. The system pulls images directly from the URL to incorporate into AI-generated videos.

While the results may vary, the goal is to simplify TikTok content creation for brands, eliminating the need for video editing expertise or familiarity with the platform. These videos are designed to align with popular TikTok trends.

According to TikTok, brands can input their product details and assets or import them from a URL. The Symphony Creative Studio then generates multiple video options with unique layouts and scripts. These videos are inspired by TikTok's top-performing content and include licensed assets like videos, images, sounds, and avatars sourced from partners such as Billo and Getty Images, all pre-approved for commercial use.

Additionally, brands can use digital avatars to narrate video scripts, create content in multiple languages for global markets, and customize the videos before publishing.

This tool adds to TikTok's growing suite of generative AI features for marketers, which already include script generation, idea prompts, and image creation. The demo versions of these tools look promising, while TikTok's digital avatars have proven successful in China, enabling brands to run 24/7 live-stream sales.

Though some trial and error may be required to achieve the desired outcome, testing these tools could be worthwhile to determine their effectiveness in promoting your brand.

TikTok has taken a measured approach with its AI innovations, focusing on practical tools rather than launching chatbots like Meta. This strategic restraint has resulted in potentially valuable AI features like this one, which could encourage more brands to leverage TikTok trends for marketing success.

TheOpensource.AI News

Near Protocol Aims to Build 1.4 Trillion Parameter Decentralized AI Model

Near Protocol has announced an ambitious initiative to develop the world’s largest open-source artificial intelligence (AI) model during its Redacted conference in Bangkok, Thailand. The planned model, featuring 1.4 trillion parameters, will surpass Meta’s open-source Llama model by a factor of 3.5.

A Crowdsourced Approach to AI Development

The model will be built through competitive crowdsourced research involving thousands of contributors via the newly launched Near AI Research hub. Starting November 10, participants can begin training a smaller 500-million-parameter model. The project will scale progressively across seven models, with contributors advancing based on their performance.

To maintain privacy and incentivize contributors, the project will utilize encrypted Trusted Execution Environments. This setup will reward participants and ensure continuous updates as technology evolves.

Funding and Monetization

The project’s estimated $160 million cost will be covered through token sales rather than issuing or selling Near Protocol’s existing tokens. Contributors will be compensated through a new token created for each model. Token holders will recoup investments through revenue generated from AI model usage, creating a self-sustaining loop for funding future developments.

Near Protocol co-founder Illia Polosukhin emphasized that while the budget is significant, it is achievable within the cryptocurrency space. Polosukhin, who co-authored the transformer research that laid the groundwork for ChatGPT, sees this as a pivotal project for decentralizing AI innovation.

Challenges and Technological Breakthroughs

The project faces substantial hurdles, including the need for “tens of thousands of GPUs” for training. Co-founder Alex Skidanov, who previously worked at OpenAI, explained that decentralized compute networks currently lack the necessary interconnect speeds for such tasks. However, emerging research, such as studies from Deep Mind, suggests these challenges can be overcome.

The Importance of Decentralized AI

Polosukhin stressed the significance of decentralizing AI to prevent any single entity from controlling the technology and the broader economy. He argued that decentralized AI is essential for maintaining the philosophical relevance of Web3 principles.

“This is probably the most important technology of our time,” Polosukhin said, warning against the dangers of centralized AI. His remarks were echoed by conference guest speaker Edward Snowden, who warned that centralized AI could lead to a pervasive surveillance state. Snowden advocated for the creation of independent, mathematically enforced systems to safeguard digital sovereignty and civil rights.

A Vision for the Future

Near Protocol aims to create a sustainable and decentralized AI ecosystem that preserves user privacy, rewards contributors, and democratizes access to cutting-edge technology. By addressing these challenges, the project seeks to set a new benchmark for open-source AI development while aligning with the core values of Web3.

Indosat and GoTo Launch Sahabat-AI to Empower Local Languages

Indosat Ooredoo Hutchison, in partnership with GoTo, has announced the launch of an open-source language model tailored for local Indonesian languages. Named Sahabat-AI, the model is designed to generate responses in Bahasa Indonesia and various regional languages across the archipelago.

The open-source initiative aims to enable Indonesians to develop AI services and applications suited to their unique cultural and linguistic contexts—an area often neglected by generalized Western AI models.

Empowering Local Communities Through AI

“By creating an AI model that speaks our language and reflects our culture, we empower every Indonesian to harness the potential of advanced technology,” said Vikram Sinha, President Director and CEO of Indosat. He emphasized that the project represents a critical step in democratizing AI for growth, innovation, and empowerment across Indonesia’s diverse society.

Research by Omdia indicates growing demand in the Asia Pacific for AI solutions aligned with local languages and cultural nuances, highlighting a shift away from Western-centric approaches. Sahabat-AI seeks to address this gap by providing a comprehensive AI ecosystem for businesses, research institutions, and government agencies.

Collaboration and Technology

The development of Sahabat-AI is supported by Nvidia, with AI Singapore and Tech Mahindra utilizing Nvidia’s AI Enterprise software and NeMo training platform to fine-tune the model for local languages. At the Indonesia AI Day launch event, Nvidia CEO Jensen Huang, alongside Sinha, GoTo CEO Patrick Walujo, and Indonesia’s Minister of State-Owned Enterprises Erick Thohir, celebrated the project’s potential.

“Sahabat-AI marks Indonesia’s entry into the AI landscape, showcasing how language models can be customized to meet unique linguistic and cultural needs,” Huang stated. He also highlighted how Indonesia’s culture of ‘gotong royong,’ or mutual collaboration, exemplifies the synergy between industry, researchers, and the public sector in advancing AI for national development.

Accessible AI for All

Sahabat-AI will launch in two model sizes, 8 billion and 9 billion parameters, making them relatively smaller than conventional language models. This design ensures affordability, allowing users to leverage AI services without incurring the high costs of large-scale models. Nvidia will continue supporting the development of the Sahabat-AI model family.

Among the early adopters is Hippocratic AI, a startup focused on safety-oriented AI for healthcare, which plans to integrate Sahabat-AI into its offerings for Indonesian users.

“Our vision for Sahabat-AI is to empower everyone in Indonesia with AI,” said Walujo. “By operating in Bahasa Indonesia, the model bridges critical gaps in context and cultural references left by global language models.”

Although Sahabat-AI is promoted as an open-source initiative, no confirmation has been provided regarding the publication of the training data, a key requirement under the recently defined open-source AI standards by the Open Source Initiative (OSI).

TheClosedsource.AI News

OpenAI’s o1 Model Redefines the Race for Smarter Systems

Artificial intelligence companies, including OpenAI, are tackling delays and challenges associated with scaling ever-larger language models by developing training techniques that allow algorithms to "think" in a more human-like manner.

A group of AI scientists, researchers, and investors told Reuters that these innovative methods, utilized in OpenAI’s recently launched o1 model, could redefine the AI arms race and significantly influence the resources AI companies rely on, such as energy and specialized chips.

Moving Beyond "Bigger is Better"

Following the launch of the viral ChatGPT chatbot two years ago, tech companies have largely endorsed the notion that scaling models with more data and computational power leads to better AI. However, prominent AI experts are now questioning this philosophy, citing its limitations.

Large model training, or “training runs,” requires vast amounts of computing power, costing tens of millions of dollars and relying on hundreds of chips running simultaneously. These runs are prone to hardware failures due to their complexity, and researchers often cannot predict the models' performance until the training is completed—a process that can take months. Additionally, the models consume enormous amounts of data, and readily available datasets worldwide are becoming scarce. Energy shortages further complicate these resource-intensive processes.

Embracing Test-Time Compute

To address these issues, researchers are exploring test-time compute, a technique that improves model performance during the "inference" phase when the model is in use. For instance, rather than selecting a single answer immediately, the model evaluates multiple possibilities in real-time, choosing the optimal solution. This approach allows models to allocate more computational resources to complex tasks like math, coding, or problem-solving that require advanced reasoning.

OpenAI researcher Noam Brown highlighted this method's efficiency, sharing at the TED AI conference that having a bot take just 20 seconds to evaluate a poker hand delivered performance boosts equivalent to scaling a model 100,000 times larger and training it 100,000 times longer.

OpenAI's o1 Model

OpenAI’s o1 model, previously known as Q* and Strawberry, leverages this human-like multi-step reasoning capability. It also incorporates curated data and feedback from PhDs and industry experts. The model is built on top of base models like GPT-4, with plans to expand this approach to larger models in the future.

Other AI labs, including Anthropic, xAI, and Google DeepMind, are also developing similar techniques, according to sources familiar with their work. Kevin Weil, OpenAI’s Chief Product Officer, emphasized the speed of these advancements, stating, “By the time others catch up, we’ll be three steps ahead.”

Impact on the AI Industry

This shift could reshape the AI hardware market, which has been dominated by the demand for Nvidia’s AI chips. Venture capital firms, such as Sequoia and Andreessen Horowitz, which have invested billions in AI labs including OpenAI and xAI, are closely monitoring the transition and evaluating its impact on their investments.

OpenAI declined to comment on the developments, while Google and xAI did not respond, and Anthropic had no immediate comment. However, these advancements mark a pivotal shift in AI development, prioritizing smarter training techniques over sheer scale to achieve greater efficiency and innovation.

Don’t miss out on the insights driving the future of Artificial Intelligence! Join a community of researchers, developers, and AI enthusiasts to stay ahead of the curve in Generative AI. Each edition delivers exclusive updates, expert analysis, and thought-provoking discussions straight to your inbox. Subscribe today and be part of the journey toward AGI innovation.

Contact us for any paid collaborations and sponsorships.

Unlock the future of problem solving with Generative AI!

If you're a professional looking to elevate your strategic insights, enhance decision-making, and redefine problem-solving with cutting-edge technologies, the Consulting in the age of Gen AI course is your gateway. Perfect for those ready to integrate Generative AI into your work and stay ahead of the curve.

In a world where AI is rapidly transforming industries, businesses need professionals and consultants who can navigate this evolving landscape. This learning experience arms you with the essential skills to leverage Generative AI for improving problem-solving, decision-making, or advising clients.

Join us and gain firsthand experience in how state-of-the-art technology can elevate your problem solving skills using GenAI to new heights. This isn’t just learning; it’s your competitive edge in an AI-driven world.