– 0Apr 29
Elon Musk recounted a past friendship dispute with OpenAI's Sam Altman during his trial. Musk testified that Altman had previously expressed a desire to start a company focused on artificial general intelligence (AGI) and had sought Musk's involvement. This testimony, given under oath, echoes narratives Musk has shared previously in interviews and in Walter Isaacson's biography. The trial centers on Musk's lawsuit against OpenAI and its CEO, alleging breaches of the company's founding principles.
▲ 70Apr 28
Researchers have developed a novel AI training paradigm called Reinforcement Learning with Verifiable Rewards with Self-Distillation (RLSD). This technique significantly lowers the computational and financial barriers for enterprises looking to build custom reasoning models. RLSD combines the reliable outcome tracking of reinforcement learning with the detailed feedback of self-distillation, outperforming traditional methods in experiments. It addresses the limitations of existing approaches like sparse feedback in reinforcement learning and high overhead in distillation, enabling more efficient and cost-effective development of specialized AI.
– 0Apr 28
OpenAI has issued new guidelines for its coding agent, Codex, to prevent it from discussing specific creatures like goblins, gremlins, raccoons, trolls, ogres, and pigeons. The instructions emphasize that such topics should only be mentioned if they are "absolutely and unambiguously relevant." This move appears to be an effort to refine the AI's focus and ensure its outputs remain professional and pertinent to its intended functions, likely in coding and technical assistance. The directive aims to steer the AI away from potentially distracting or irrelevant conversational tangents.
▲ 70Apr 28
AI startup Poolside has released two new large language models, Laguna M.1 and Laguna XS.2, optimized for agentic workflows and coding tasks. The smaller Laguna XS.2 is available under an open-source Apache 2.0 license, allowing developers to run it locally on consumer hardware. This move aims to foster community innovation and provide a powerful, efficient alternative to proprietary models. Poolside's approach focuses on software development as a proxy for general intelligence, utilizing advanced training techniques and novel developer environments like 'pool' and 'shimmer' to enhance AI capabilities.
– 0Apr 28
Elon Musk testified that his primary motivation for co-founding OpenAI was to prevent a catastrophic 'Terminator outcome' from advanced artificial intelligence. During his testimony, Musk also faced a warning from the judge regarding his and Sam Altman's public social media exchanges, which were deemed to be exacerbating tensions outside the courtroom. This legal proceeding highlights the ongoing debates and concerns surrounding the development and control of powerful AI technologies.
▼ -60Apr 28
Stocks associated with OpenAI experienced a decline following a Wall Street Journal report indicating the company did not achieve its internal sales and user growth objectives. This news comes as a jury has been selected for the high-profile trial where OpenAI faces allegations from Elon Musk that it has strayed from its original mission. The legal proceedings and the missed targets raise questions about OpenAI's trajectory and market performance, potentially impacting investor confidence in related ventures.
– 0Apr 28
Taylor Swift is intensifying her legal battle against AI imitation, seeking trademark protection for two spoken phrases: 'Hey, it's Taylor Swift' and 'Hey, it's Taylor.' Her team filed applications including audio clips of her saying these phrases, originally used in album promotions. This move highlights the complex intersection of celebrity rights and emerging AI technology. While Swift's efforts aim to safeguard her identity and voice from unauthorized AI replication, the legal landscape for such protections remains challenging and potentially a long shot for comprehensive enforcement against AI-generated content.
The Vergeai
regulation
legal
– 0Apr 28
Amazon Web Services (AWS) is now offering OpenAI's latest models, including its code-writing service Codex and a new AI agent creation tool, through its Bedrock service. This move follows a revised agreement between OpenAI and Microsoft, which previously held exclusive rights to OpenAI's products. Amazon CEO Andy Jassy hinted at this development, and the integration signifies a deepening collaboration between AWS and OpenAI. This partnership diversification comes as both OpenAI and Microsoft reportedly explore relationships with rivals, with OpenAI also engaging AWS and Oracle, and Microsoft working with Anthropic.
▲ 60Apr 28
A recent survey reveals that over 80% of U.S. government agencies are already utilizing AI agents, signaling a significant adoption of artificial intelligence within the public sector. This trend is projected to accelerate, with a majority of government leaders anticipating a future where humans and AI agents collaborate extensively by 2030. The findings underscore a growing reliance on AI for operational efficiency and service delivery, suggesting a transformative shift in how government functions will operate in the coming years.
▼ -60Apr 28
Elon Musk has begun his testimony in a significant legal battle against OpenAI, specifically targeting CEO Sam Altman and president Greg Brockman. Musk, an initial co-founder of OpenAI, is suing the company over alleged breaches of its founding mission and structure. Disagreements reportedly centered on OpenAI's shift towards a more commercial, less open model, and potential integration with Musk's own ventures like Tesla. Having previously invested substantially, Musk later departed and founded a rival AI company, xAI. This trial represents a critical juncture in the ongoing disputes between Musk and the AI giant he helped create.
The Vergeai
legal
regulation
– 0Apr 28
Amazon has introduced a new AI-powered feature called "Join the chat" on its product pages. This innovative tool allows customers to ask questions directly about products and receive instant audio responses generated by artificial intelligence. The feature aims to enhance the shopping experience by providing quick, conversational answers to user inquiries, making product information more accessible and engaging. This move signifies Amazon's continued investment in AI to improve customer interaction and streamline the online purchasing journey.
▼ -50Apr 28
Hundreds of contract workers in Ireland responsible for training Meta's artificial intelligence models are reportedly at risk of job losses. Documents indicate that over 700 individuals could be impacted by these potential layoffs. This situation raises concerns about the precarious employment conditions for those involved in the data labeling and AI training processes, which are crucial yet often overlooked aspects of developing advanced AI systems. The news highlights the human cost associated with the rapid evolution and restructuring within the AI industry.