machine learning

Looking to accelerate the training of AI models? Lightning AI's Thunder compiler, developed in collaboration with Nvidia, is here to help. With the ability to utilize multiple GPUs, Thunder offers a 40% speed-up in training large language models compared to unoptimized code. Compatible with Nvidia's suite of products and led by PyTorch core developer Thomas Viehmann, Thunder promises to maximize GPU potential and expedite workflows for organizations. Find out more about this exciting new offering from Lightning AI.

Discover the significant data security risks posed by the rapid surge in enterprise AI adoption. Zscaler's report reveals a 600% increase in adoption, with a corresponding rise in blocked AI/ML transactions. Learn how adversarial AI is being weaponized by attackers, and how CISOs can protect their organizations. Explore the dominance of data protection, AI data quality management, and privacy concerns in the industry. Find out which industries are most unprepared for AI-based attacks and how leading cybersecurity vendors are leveraging AI to stay ahead. Gain insight into the emergence of a more lethal AI threatscape and the need for risk management.

Foundational, a startup focused on data quality and AI readiness, has raised million in seed funding. Their platform automatically analyzes data teams' code to identify potential issues and prepare data for AI applications. With public companies already signed on as customers, Foundational aims to solve the growing concern of data quality as organizations scale up their data capabilities. By providing actionable suggestions and fixes within developers' workflows, Foundational helps maintain complex data pipelines. Their comprehensive code analysis approach positions them as a key player in unlocking the full potential of data-driven initiatives in the digital age.

Introducing the groundbreaking arrival of KL3M, the first Large Language Model (LLM) to receive a "Licensed Model (L) Certification" from Fairly Trained. Developed by startup 273 Ventures, KL3M was trained on legally obtained data, challenging the assumption that copyrighted materials are necessary for AI training. Designed for the legal industry, KL3M offers impressive performance and broader applications. This development marks a significant milestone in the AI industry, prioritizing ethical and lawful AI development. Find out more about KL3M and its potential impact.

Discover the unseen cyber threat of adversarial AI and understand the dangers it poses to organizations. Learn about the significant gap between IT leaders' intentions and actions in securing AI systems, with 77% experiencing AI-related breaches. Explore the three classes of adversarial AI attacks and the importance of proactive defense measures. Bridge the gap between intention and action to protect your AI systems from manipulation and breaches.

Discover how artificial intelligence (AI) is assisting DevOps teams in mitigating supply chain attacks in five key ways. From automating security processes and enhancing threat detection capabilities to streamlining analytics and automating patch management, AI is proving to be a powerful ally in the battle against supply chain attacks. Learn how leading vendors in cloud-native application protection platforms (CNAPPs) and extended detection and response (XDR) are leveraging AI to strengthen supply chain security and safeguard the integrity of software supply chains.

Discover the concept of "Antagonistic AI" proposed by MIT and University of Montreal researchers. This new approach challenges the current overly-sanitized and agreeable nature of AI systems, advocating for purposeful combativeness, criticism, and even rudeness. Explore how antagonistic AI can foster personal and collective growth, build resilience, promote self-reflection, and diversify ideas. Learn about the three types of antagonism and the responsible and ethical practices necessary for implementation. Embrace the idea that AI should reflect the full range of human values, including honesty, boldness, eccentricity, and humor. Join the discussion on the future of AI and its potential to engage and challenge users in meaningful ways.

Discover Alibaba's groundbreaking development in artificial intelligence with EMO, a system capable of animating portrait photos and generating lifelike videos of individuals talking or singing. By utilizing a direct audio-to-video synthesis approach, EMO captures subtle motions and identity-specific quirks associated with natural speech, surpassing traditional methods. Backed by extensive training and a diffusion model, EMO outperforms existing techniques in video quality, identity preservation, and expressiveness. Explore the vast implications of this technology, while considering the ethical concerns surrounding its potential misuse. Alibaba's EMO system has the power to revolutionize industries such as entertainment, marketing, and personal communication.

Looking to break out of a slowdown, venture capitalists are turning their attention to emerging opportunities in artificial intelligence (AI). Pitchbook's latest report highlights the challenges faced by VCs, but also identifies three areas of potential growth: AI data centers, local large language models (LLMs), and domain-specific foundation models. Despite a decline in deal activity, there are exceptions, and the report highlights Nvidia's role in driving the AI market. With breakout growth potential and opportunities to partner with startups, VCs are positioning themselves for success in the AI market.