Samsung Confirms Major Generative AI Overhaul For Bixby

Welcome to Towards AGI, your premier newsletter dedicated to the world of Artificial Intelligence. Our mission is to guide you through the evolving realm of AI with a specific focus on Generative AI. Each issue is designed to enrich your understanding and spark your curiosity about the advancements and challenges shaping the future of AI.

Whether you're deeply embedded in the AI industry or just beginning to explore its vast potential, "Towards AGI" is crafted to provide you with comprehensive insights and discussions on the most pertinent topics. From groundbreaking research to ethical considerations, our newsletter is here to keep you at the forefront of AI innovation. Join our community of AI professionals, hobbyists, and academics as we pursue the ambitious path toward Artificial General Intelligence. Let’s embark on this journey together, exploring the rich landscape of AI through expert analysis, exclusive content, and engaging discussions.

TheGen.AI News

Samsung Confirms Major Generative AI Overhaul for Bixby

Samsung has officially revealed its plans for the future of its AI assistant, Bixby. TM Roh, the Head of Mobile Business, confirmed in an interview with CNBC that Bixby will undergo significant updates "this year," incorporating generative AI technology. Roh mentioned that Samsung aims to enhance the number of generative AI features available on its Galaxy devices.

He emphasized the convenience and benefits that AI capabilities bring to consumers, stating that these features are major drivers for purchasing new products. Despite the focus on expanding Galaxy AI, Samsung will continue offering multiple AI assistant options on its devices, maintaining its existing strategy.

Details on Bixby's specific upgrades for 2024 remain scarce. However, this confirmation follows earlier rumors that Samsung would integrate generative AI features into Bixby, inspired by OpenAI's work on ChatGPT. The source of these rumors indicated that Samsung needed to redefine Bixby's purpose due to the rise of large language models and generative AI.

During the Unpacked event, Samsung also showcased its commitment to multiple AI options. Google's Rick Osterloh joined the stage to discuss AI advancements and the ongoing partnership between Google and Samsung, highlighting how their collaboration has revolutionized the mobile experience.

The latest Galaxy devices, Flip 6 and Fold 6, are set to receive new Gemini features. Osterloh praised these foldable phones as ideal platforms for innovation.

Alibaba Leverages Generative AI for Overseas Merchant Support

Alibaba's international e-commerce division is leveraging generative AI tools to aid cross-border merchants with translation, content creation, and product returns, according to group vice president Kaifu Zhang.

During his presentation at the Reuters Next Conference in Singapore, Zhang, who leads Alibaba International Digital Commerce's AI initiative, expressed optimism about AI's potential for the company, despite growing tensions between China and the U.S.

Zhang stated, "China is nearly on par with the U.S. in the pace of AI innovation," noting significant advancements in open-source AI technology made in China and Europe.

The tech conglomerate has been working to enhance the competitiveness of its global marketplaces, such as AliExpress and Lazada, following the restructuring of its international e-commerce division into an independent business.

In recent years, Alibaba's early advantages have diminished, with competitors like Shein and NASDAQ-listed PDD's Temu outperforming it in sales and growth. Zhang mentioned that Alibaba International Digital Commerce Group now serves 300 million consumers annually and has over one million merchants worldwide.

This division has become one of Alibaba's fastest-growing segments, with earnings increasing by 44% from October to December compared to the previous year.

Generative AI Startup Simplismart Secures $7 Million Funding

Bengaluru-based generative AI startup Simplismart has reportedly secured $7 million in a funding round led by Accel.

Previously, the startup had raised seed funding from investors including Anicut Capital, Sunn91, First Cheque, and Shastra VC. According to Entrackr, this round values Simplismart at approximately $35 million.

What is Simplismart's focus?

Founded in 2022 by Amritanshu Jain and Devansh Ghatak, Simplismart specializes in helping businesses create scalable, production-grade AI systems without needing to write code. The startup's online platform enables both beginners and experts to collaboratively train and monitor machine learning models across various data types and use cases.

The newly raised funds will be used to further develop the company's technology, scale its operations, and expand its team.

Additionally, the financing will accelerate Simplismart's go-to-market strategy and broaden its customer base across different industries.

The company aims to enhance its marketing and sales efforts to showcase the capabilities of its generative AI solutions, which are designed to automate tasks, reduce costs, and improve operational efficiency.

TheOpensource.AI News

Bioptimus Launches Advanced Open-Source AI for Medical Diagnostics

French startup Bioptimus has introduced H-optimus-0, an open-source AI foundation model for pathology.

The model, featuring 1.1 billion parameters, was trained on a proprietary dataset of hundreds of millions of images from over 500,000 histopathology slides.

Due to its extensive training on a diverse dataset from 4,000 clinical practices, it can perform various medical diagnostic tasks, such as identifying cancerous cells and detecting genetic abnormalities in tumors.

Bioptimus CTO Rodolphe Jenatton commented, “We are thrilled to unveil H-optimus-0, a fully open-source model, just five months after our launch. This model represents a significant advancement in pathology. We are dedicated to advancing medical research and enhancing patient care.

"By fostering collaboration and the adoption of new practices, we aim to accelerate innovations in pathology and beyond, ultimately benefiting patients worldwide." H-optimus-0’s extensive training ensures high accuracy and the ability to handle a wide range of diagnostic scenarios. The model has shown superior performance in various diagnostic tasks, setting new standards in the field.

The open-source nature of H-optimus-0 promotes widespread access and collaboration.

This model is expected to drive innovation and enhance diagnostic tools in pathology, ultimately improving patient care.

Bioptimus co-founder and CEO, Professor Jean-Philippe Vert, stated, “H-optimus-0 is just the beginning. It marks the first in a series of increasingly advanced and comprehensive models we plan to create at Bioptimus.

“Future models will be trained on even larger datasets of pathology images from Europe, Asia, and Africa, and will include other modalities such as genomics and proteomics.

"Our ultimate aim is to develop the first multiscale foundation model of biology, capable of integrating diverse biological data and scales to enable scientific discoveries and accelerate biomedical innovations.”

Patronus AI Unveils ‘Lynx’: The Open-Source Model Outsmarting GPT-4

Patronus AI, a startup based in New York, has introduced Lynx, an open-source model aimed at detecting and mitigating hallucinations in large language models (LLMs). This development has the potential to transform enterprise AI adoption, addressing the reliability issues of AI-generated content.

Lynx surpasses industry leaders like OpenAI’s GPT-4 and Anthropic’s Claude 3 in hallucination detection tasks, marking a significant advancement in AI reliability. Patronus AI claims that Lynx achieved 8.3% higher accuracy than GPT-4 in identifying medical inaccuracies and outperformed GPT-3.5 by 29% across all tasks. 

In an interview with VentureBeat, Anand Kannappan, CEO of Patronus AI, highlighted the importance of this achievement. “Hallucinations in large language models occur when the AI generates false or misleading information, presenting it as fact,” he explained. “For enterprises, this can lead to incorrect decision-making, misinformation, and a loss of trust from clients and customers.”

Patronus AI also introduced HaluBench, a new benchmark for assessing AI model accuracy in real-world scenarios. This tool is notable for its inclusion of domain-specific tasks in finance and medicine, where precision is essential. “Industries that handle sensitive and precise information, such as finance, healthcare, legal services, and any sector requiring stringent data accuracy, will greatly benefit from Lynx,” Kannappan noted. “Its ability to detect and correct hallucinations ensures that critical decisions are based on accurate data.”

The decision to open-source Lynx and HaluBench may accelerate the adoption of more reliable AI systems across various sectors. However, it also raises questions about Patronus AI’s business model. Kannappan addressed this by stating, “We plan to monetize Lynx through our enterprise solutions that include scalable API access, advanced evaluation features and workflows, and bespoke integrations tailored to specific business needs.” This strategy reflects the broader trend of AI companies offering premium services based on open-source foundations.

The launch of Lynx comes at a pivotal moment in AI development. As enterprises increasingly rely on LLMs for various applications, the need for robust evaluation and error-detection tools becomes urgent. Patronus AI’s innovation could significantly enhance trust in AI systems, potentially speeding up their integration into critical business processes.

TheClosedsource.AI News

Worldcoin Falters While OpenAI Targets Human-Level Problem-Solving

OpenAI, the American tech company known for the Large Language Model (LLM) ChatGPT, has outlined five stages towards achieving Artificial General Intelligence (AGI). According to employees who spoke with Bloomberg, OpenAI is currently at level one with ChatGPT.

As OpenAI progresses towards AGI, the Worldcoin (WLD) project, led by OpenAI founder Sam Altman, is struggling. WLD has faced selling pressure from token unlocks, losing over 25% of its value in July. A Bloomberg report noted that OpenAI's path to AGI involves five stages. While OpenAI aims for tier 2, human-level reasoning, it hasn't positively impacted Worldcoin's price.

WLD has been closely linked to OpenAI's progress and updates, as both projects are led by Sam Altman. In July, WLD is experiencing significant selling pressure due to token unlocks, increasing the supply on exchanges.

Santiment data shows that the supply on exchanges reached 43.97 million on Friday, nearing the highest level since October 2023. A high supply on exchanges typically means more tokens are available for sale, increasing pressure on the asset and potentially driving prices down.

OpenAI continues to strive towards creating the most advanced tier-five AGI, which aims to replace organizations. This recent development at OpenAI was driven by executives and senior leaders. The tier system is still being refined, with the company gathering feedback to adjust the levels over time.

With no other major market influences, traders are focusing on WLD's on-chain metrics and developments at OpenAI to predict the future of the AI token. Recovery for Worldcoin's price seems unlikely in July as token unlocks add 6.62 million WLD to the supply daily, according to Token Unlock data.

Sanofi Teams Up with OpenAI to Pioneer AI-Driven Pharma Industry

In May 2024, French pharmaceutical giant Sanofi announced a collaboration with Formation Bio and OpenAI to develop AI software for drug research.

William Tsai, Sanofi's Head of Public Affairs, Communication, and Market Access for Hong Kong and Taiwan, stated that this partnership supports the company's ambitious digital transformation goals for 2030.

Sanofi, the world's sixth-largest pharmaceutical company, operates over 70 locations globally with 86,000 employees, offering products such as vaccines and treatments for cardiovascular diseases, diabetes, and rare diseases. In Taiwan, where Sanofi has been present since 1982, the company employs around 310 people.

Digital Transformation 2030: Accelerating Drug Research with OpenAI

Tsai emphasized that drug development is typically time-consuming, costly, and high-risk, with a success rate as low as one in ten thousand. In 2023, Sanofi's research and development expenditure was EUR 6.7 billion.

The company aims to use digitization and AI technology to improve research efficiency and reduce development timelines. AI has the potential to revolutionize various stages of drug development, including disease target identification, drug creation, and clinical translation.

Sanofi's digital transformation goals for 2030 include becoming the first large-scale AI-driven pharmaceutical company, cutting research timelines in half, and dedicating 50% of its resources to innovative projects. Achieving these goals requires cross-disciplinary partners, which is why they are partnering with OpenAI.

At the announcement, Sanofi mentioned it would provide its data to create customized applications using OpenAI technology. OpenAI's COO, Brad Lightcap, expressed optimism about AI's potential to accelerate drug development, hoping the collaboration would lead to new drugs that benefit patients and their families.

Formation Bio will contribute its expertise in drug development and AI applications.

Digitizing Healthcare Globally

In Taiwan, Sanofi has been working with Health2Sync for several years on insulin management software for diabetes and is conducting clinical trials. The company has various strategic partners in other countries.

Tsai pointed out that digital healthcare is thriving in countries like Iceland, Spain, Sweden, and Austria, which are not typically known as pharmaceutical powerhouses. These digital healthcare companies often collaborate with multiple pharmaceutical firms and have numerous cross-disciplinary or international partnerships, reflecting a growing trend.

Sanofi's initiatives in Taiwan are also being implemented in Japan and South Korea. With 70 locations worldwide, the company aims to enhance cooperation with local healthcare professionals.

OpenAI Reveals Five-Tier Plan to Achieve Superintelligent AI

OpenAI has reportedly introduced a new classification system outlining various stages of AI development on the path to achieving superintelligent AI. According to a recent report, this system, shared with OpenAI employees, comprises five levels: the first encompasses current generative AI chatbots, while the fifth represents artificial general intelligence (AGI), capable of performing the work of an entire organization. The report also mentions a new project utilizing the GPT-4 model, demonstrating human-like reasoning, potentially elevating it to level two.

OpenAI Unveils Five Levels Toward AGI

A Bloomberg report, citing an OpenAI spokesperson, revealed that the company shared the new classification system during an "all-hands meeting" with its employees. OpenAI intends to present this classification to investors and the public, although no specific timeline has been provided.

Currently, OpenAI is at level one. However, according to sources familiar with the situation, the company has demonstrated new GPT-4 capabilities that might qualify it for level two. The standout feature is the AI model's reasoning ability, although no additional features were detailed.

The five levels towards AGI, as shared in the publication, are:

1. Level 1 — Chatbots: AI with conversational language capabilities

2. Level 2 — Reasoners: AI with human-level problem-solving abilities

3. Level 3 — Agents: AI systems that can take independent actions

4. Level 4 — Innovators: AI that can assist in inventing new technologies and products

5. Level 5 — Organizations: AI capable of performing the work of an entire organization

Most AI systems today fall under level one, featuring conversational chatbot capabilities. OpenAI is reportedly advancing to level two, termed Reasoners, with GPT-4, although this capability has yet to be publicly demonstrated.

Level three involves Agents, and AI systems that can perform actions on behalf of humans, likely marking the integration of AI into real-world operations. The fourth level, Innovators, will enable AI to contribute to the invention of new technologies and products, a trait considered a hallmark of human intelligence.

In our quest to explore the dynamic and rapidly evolving field of Artificial Intelligence, this newsletter is your go-to source for the latest developments, breakthroughs, and discussions on Generative AI. Each edition brings you the most compelling news and insights from the forefront of Generative AI (GenAI), featuring cutting-edge research, transformative technologies, and the pioneering work of industry leaders.

Highlights from GenAI, OpenAI, and ClosedAI: Dive into the latest projects and innovations from the leading organizations behind some of the most advanced AI models in open-source, closed-sourced AI.

Stay Informed and Engaged: Whether you're a researcher, developer, entrepreneur, or enthusiast, "Towards AGI" aims to keep you informed and inspired. From technical deep-dives to ethical debates, our newsletter addresses the multifaceted aspects of AI development and its implications on society and industry.

Join us on this exciting journey as we navigate the complex landscape of artificial intelligence, moving steadily towards the realization of AGI. Stay tuned for exclusive interviews, expert opinions, and much more!

Keep reading