Tips

May 30, 2024

5 mins

AI Overview: Your Weekly AI Briefing

Hello Niuralogists!

Step into this week's edition as we navigate the dynamic realm of artificial intelligence to present you with the most recent breakthroughs. Our primary focus is to dissect the implications of these updates on various aspects of our lives, ranging from workplaces and businesses to policies and individual experiences. In this edition, we'll uncover exciting developments, including OpenAI's commencement of training for its upcoming AI model and SAP's integration of Amazon Bedrock into its AI core, streamlining generative AI for regulated industries.

For deeper insights, continue reading…

OpenAI Initiates Training for its Next AI Model

OpenAI recently announced the formation of a new Safety and Security Committee to oversee the development of its next frontier AI model, which the company has just begun training. This committee, led by CEO Sam Altman and including board members Bret Taylor, Adam D’Angelo, and Nicole Seligman, will publicly evaluate safety and security processes over the next 90 days and share their recommendations. OpenAI stated that this new model will advance their capabilities significantly in achieving artificial general intelligence (AGI). However, the formation of the committee, composed solely of Altman's supporters, has raised concerns among critics who distrust the current leadership. The vague commentary about the new model has also fueled speculation that GPT-5's release is imminent and the next system is already in training.

SAP Integrates Amazon Bedrock into AI Core, Simplifying Generative AI for Regulated Industries

SAP is deepening its relationship with Amazon Web Services (AWS) by integrating Amazon Bedrock into its AI Core platform, facilitating generative AI experimentation for regulated firms. SAP will leverage AWS's Graviton4, Trainium, and Inferentia chips to power its AI infrastructure, enhancing performance and reducing costs. This integration will allow enterprise customers to access a variety of large language models, including Amazon Titan, Anthropic’s Claude 3 Opus, and Meta’s Llama, to improve business processes. According to Matt Garman, AWS’s incoming CEO, the collaboration aims to make generative AI applications more efficient, responsive, and sustainable. This move is particularly significant for companies in regulated sectors, such as finance, healthcare, and public services, which can now use AI to optimize large amounts of text data and improve compliance. The integration is part of SAP’s broader strategy to offer a standardized, scalable, and hyper-scale-agnostic AI management solution within its Business Technology Platform.

Source: Adobe Stock

AI Demonstrates Exceptional Stock-Picking Skills

A new study from the University of Chicago has revealed that large language models (LLMs) like GPT-4 can effectively perform financial statement analysis and even surpass human analysts in predicting future earnings trends. Researchers supplied GPT-4 with standard, anonymized financial statements and instructed it to process the information as human analysts would. Remarkably, GPT-4's predictions on future earnings trends outperformed those of professional analysts, despite lacking access to qualitative data. Additionally, its performance matched or slightly exceeded that of other cutting-edge models specifically designed for earnings prediction. The study found that GPT-4's success stemmed from its ability to analyze trends, ratios, and economic reasoning rather than from its training data. This discovery suggests that AI is poised to significantly transform the financial industry, challenging assumptions about its capabilities in judgment-based tasks. The findings indicate that AI financial expertise could soon become widely accessible, democratizing information and analysis and revolutionizing how non-experts engage with the markets.

GitHub Accelerator Ignites Open Source AI Revolution, Empowering Startups to Democratize Access

GitHub has launched its 2024 Accelerator program, aiming to support and promote open-source projects that are reshaping the technology landscape. The program provides financial backing, mentorship, and community-building opportunities, generating significant buzz among developers and enterprises. GitHub’s VP of Communities, Stormy Peters, emphasized that the initiative empowers developers to bring their innovative ideas to life with funding and resources to scale projects globally. This year's cohort includes diverse projects such as Unsloth, which enhances AI model efficiency, and Formbricks, which has revolutionized user feedback. The Accelerator program strengthens GitHub's position as a key enabler of open-source innovation by fostering promising projects and ensuring a steady pipeline of technological advancements. With financial support of $40,000 per project, extensive mentorship, and community engagement, GitHub aims to address challenges faced by open-source developers and democratize access to cutting-edge technologies. As the 2024 program kicks off, the tech industry eagerly anticipates the breakthroughs that will emerge from this year's cohort.

Source: GitHub

OpenAI Controversy: Insights from 'Sky'

OpenAI is embroiled in several controversies, including leaked documents about its NDA clauses, new information on the 'Sky' voice model, and another resignation. The voice actress for ChatGPT's 'Sky' voice revealed she was hired months before CEO Sam Altman contacted Scarlett Johansson. Leaked documents signed by Altman contradict his claims of being unaware of NDA clauses that could reclaim former employees' equity. OpenAI policy researcher Gretchen Kreuger resigned, citing safety concerns. Presentations at the VivaTech conference hinted at a potential rebranding away from 'GPT-5' and showcased a new Sora demo. These issues add to the ongoing drama surrounding OpenAI, suggesting that more turmoil may be on the horizon despite the anticipation for upcoming releases.

Newsletter

📬 Receive our amazing posts straight to your inbox. Get the latest news, company insights, and Niural updates.

Thank you! Your message has been received!
Oops! Something went wrong. Please fill in the required fields and try again.

Q&Ai

Can This AI-Based Method Help You Find a Specific Action in a Video?

MIT researchers have developed an innovative method to enhance video analysis, enabling machines to pinpoint specific actions within lengthy videos. Unlike traditional approaches that rely on labor-intensive manual labeling, their technique leverages unlabeled videos and their transcripts for training. By teaching AI models to simultaneously grasp spatial details and temporal sequences, the researchers achieve superior accuracy in identifying multiple activities within extended video footage. This advancement not only streamlines virtual training and educational processes but also holds promise for applications in healthcare, facilitating swift access to critical moments in diagnostic videos.

Can Controlled Diffusion Models Change Material Properties in Images?

MIT researchers, in collaboration with Google Research, have developed a cutting-edge diffusion model known as Alchemist, capable of altering material properties within images with unprecedented precision. This innovative system allows users to adjust attributes such as roughness, metallicity, albedo (base color), and transparency on a continuous scale, significantly enhancing capabilities in fields like video game design, visual effects, and robotic training. Unlike traditional methods, Alchemist leverages a slider-based interface to refine these details efficiently, surpassing existing tools like Photoshop in simplicity and effectiveness. This advancement not only promises to streamline creative workflows but also holds potential for advancing AI applications in diverse visual contexts, from graphics refinement to industrial automation.

Tools

  • 📊 Nomic Atalas structures unstructured data, make AI available anywhere
  • 🤖 Forloop is a no-code web scraping and data automation
  • 🚀 Merlin AI - 26-in-1 is an AI extension to write, summarize & code
  • 🎥 Glato creates high-quality product video ads in minutes
  • 🎵 AI Jukebox  is an in-browser text-to-music generation

Follow us on Twitter and LinkedIn for more content on artificial intelligence, global payments, and compliance. Learn more about how Niural uses AI for global payments and team management to care for your company's most valuable resource: your people.

See you next week!

Request a demo