Mustafa Suleyman on Microsoft and OpenAI: ‘We Are Siblings, Sometimes They Squabble’

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.

Whether you're passionate about machine learning, neural networks, or the ethics surrounding GenAI, our platform offers cutting-edge insights, resources, and collaborations on everything AI.

What to expect from Towards AGI: Know Your Inference (KYI): Ensuring your AI-generated insights are accurate, ethical, and fit for purpose Open vs Closed AI: Get expert analysis to help you navigate the open-source vs closed-source AI GenAI Maturity Assessment: Evaluate and audit your AI capabilities Expert Insights & Articles: Stay informed with deep dives into the latest AI advancements But that’s not all!

We are training specialised AI Analyst Agents for CxOs to interact and seek insights and answers for their most pressing questions. No more waiting for a Gartner Analyst appointment for weeks, You’ll be just a prompt away from getting the insights you need to make critical business decisions. Watch out this space!!

Visit us at https://www.towardsagi.ai to be part of the future of AI. Let’s build the next wave of AI innovations together!

TheGen.AI News

Mustafa Suleyman on Microsoft and OpenAI: ‘We Are Siblings, Sometimes They Squabble’

Microsoft is continuing to develop its own large language model, but the company's relationship with OpenAI remains strong, according to AI chief Mustafa Suleyman. In a recent interview on the "On with Kara Swisher" podcast, Suleyman characterized the connection between the two companies as a "good-natured sibling rivalry." "We are siblings," Suleyman explained, adding that while there are occasional disagreements, both companies are largely aligned.

While Microsoft works on its in-house AI model, MAI-1, OpenAI has rolled out GPT-4o and o1-preview. Despite the competition, Suleyman dismissed any notion of tension, describing the rivalry as "healthy and natural."

Microsoft's deep connection with OpenAI began in 2019 with a $1 billion investment in the AI startup. Suleyman described the decision as one for "corporate history" and praised it as a bold move, given that the technology was still in early research stages and far from practical applications at the time.

Since then, Microsoft has invested over $13 billion in OpenAI, fostering close collaboration on research and infrastructure. The partnership has strengthened, even as both companies continue to advance their own AI ambitions.

Microsoft CEO Satya Nadella demonstrated his support for OpenAI’s Sam Altman after his brief ousting last year, offering positions at Microsoft to both Altman and OpenAI president Greg Brockman. Although both returned to OpenAI soon after, Suleyman assumed leadership of Microsoft's AI division in March, alongside Karén Simonyan.

Suleyman also emphasized that Microsoft's partnerships extend beyond OpenAI, revealing that they have significant collaborations with other companies as well, including Amazon, across multiple areas.

PwC India and Meta Join Forces to Develop Enterprise-Grade GenAI Solutions

PwC India has teamed up with Meta to boost the adoption of open-source AI solutions through Meta’s Llama models. The collaboration is geared toward creating innovative applications for businesses and public services, with a strong focus on the capabilities of Generative AI (GenAI).

This partnership coincides with the Indian government’s launch of the IndiaAI Mission, aimed at developing a comprehensive AI ecosystem in the country. The initiative offers an opportunity for businesses to explore GenAI's potential in enhancing operational efficiency and driving digital transformation. Sanjeev Krishan, Chairperson of PwC India, stated that the collaboration is focused on making AI technology more accessible across various sectors and tackling real-world challenges. Together, PwC and Meta aim to create enterprise-grade solutions by combining Meta’s Llama platform with PwC’s industry expertise.

Sandhya Devanathan, Vice President and Head of Meta India, highlighted how GenAI solutions, including Llama, will enhance operational efficiency and customer experiences for Indian businesses. She also stressed the importance of GenAI in supporting data-driven decision-making as the country’s digital economy continues to grow.

The partnership will build on the work of PwC India's GenAI Lab, which develops innovative solutions using Generative AI. PwC India and Meta will offer GenAI capabilities to clients by merging consulting with technology expertise to meet the rising demand for secure and efficient AI solutions.

This collaboration is a significant step in integrating AI technologies across India, serving both businesses and public services. Vivek Belgavi, Partner & Leader - Alliances & Ecosystems at PwC India, expressed excitement about the partnership, noting that it provides clients with new ways to leverage GenAI, unlocking efficiency and creativity. By reimagining business operations, PwC India and Meta are paving the way for transformative growth and operational excellence.

Bain & Co: Banks Falling Behind in GenAI Adoption for Client Engagement

Only two out of 16 financial institutions interviewed by the Hong Kong Monetary Authority (HKMA) are currently using Generative AI (GenAI) to develop personalized marketing plans or tailored client pitches. According to Bain & Company, banks are lagging behind other industries in leveraging GenAI to differentiate their products and services, missing out on significant monetization opportunities.

This is particularly evident in marketing and customer engagement, where industries like retail and consumer goods have made considerable strides, said Andreas Dullweber, Bain’s global head of marketing and customer experience. He highlighted that financial institutions could learn from these industries and capitalize on the productivity gains offered by GenAI to enhance revenue and reduce costs, citing a global study involving financial institutions.

Banks and financial firms have largely applied GenAI to back-office functions like process optimization and fraud management, but they are only now beginning to test its use in customer-facing services. GenAI refers to algorithms capable of generating new content, such as audio, videos, code, images, and text.

A recent HKMA report noted that integrating GenAI could help financial institutions move beyond standardization and achieve greater differentiation, allowing for hyper-personalized approaches that unlock previously untapped monetization opportunities.

Bain's analysis of 20 global financial institutions showed that deploying GenAI could lead to cost reductions of 20-30% in customer service over two to three years and 10-15% in marketing, with potential revenue gains of around 5%. However, significant structural changes will be needed to fully realize these benefits, according to Bain partner Priscilla Dell'Orto. She noted that such transformations would encounter resistance and bring challenges due to the highly regulated nature of the financial industry.

The HKMA issued GenAI guidelines in August, focusing on governance, accountability, fairness, transparency, and data protection. The guidelines followed the launch of a GenAI sandbox that allows banks to experiment with use cases. Dell'Orto emphasized that the collaborative relationship between financial institutions and regulators would support innovation in the sector.

Banks and insurers can harness internal and external data to better target customers with personalized marketing, avoiding generic offers that may alienate clients. For instance, a global bank in Hong Kong uses GenAI to translate marketing materials and generate conversation prompts for relationship managers, while a major local bank employs GenAI to craft personalized messages for clients.

In wealth management, 74% of affluent clients in Hong Kong are comfortable with AI-guided decisions, with 25% feeling "extremely comfortable," according to a recent Capco survey. Dullweber noted that some financial services firms are already realizing the financial benefits of GenAI, particularly in personalized marketing content such as customized adverts, messages, and images.

Pinterest Launches Generative AI Features for Advertisers to Transform Product Images

Pinterest, the popular image inspiration and shopping platform, is the latest tech company to introduce generative AI tools for its advertisers, following in the footsteps of Google, Amazon, and others. During its Pinterest Presents advertiser event on Tuesday, the company revealed new features for its Pinterest Performance+ suite, which now includes both automation and AI capabilities. One key feature enables advertisers to transform plain, white, or flat backgrounds into lifestyle imagery using generative AI.

This new tool, launching on Tuesday, aims to help advertisers enhance their Pinterest Product Pins (ads) and drive more engagement, according to Pinterest. While the concept of using AI to transform product imagery is not exclusive to Pinterest, it is quickly becoming a standard offering for advertisers as it allows for fast image enhancement with simple prompts. Amazon and Google introduced similar features nearly a year ago, and now Pinterest is following suit, allowing advertisers to apply this functionality on its platform.

Walgreens, an early user of Pinterest’s generative AI technology, tested the tool across 50 products in various categories. The drugstore chain was able to present its products in a more Pinterest-appropriate style, rather than showcasing them on plain backgrounds. As a result, Walgreens experienced a 55% increase in clickthrough rates and a 13% reduction in cost-per-click with the AI-generated backgrounds compared to standard white backgrounds.

Additionally, Pinterest claims that combining AI and automation features will help advertisers create campaigns more efficiently, requiring 50% less input. Advertisers can also customize their use of Pinterest Performance+ tools based on their specific needs. In early trials, Pinterest reports a 64% drop in cost per action, a 30% rise in conversion rates, and a 1.8x return on ad spend.

Pinterest is also introducing new Promotions tools, which offer users discounts and will be available to advertisers in markets such as the U.S., U.K., Canada, Australia, France, Germany, Mexico, Brazil, and Japan upon request. Personalized Promotions and Deals ads modules will surface sales and discounts for Pinterest users. Furthermore, Pinterest Performance+ bidding will now allow advertisers to optimize for the highest value, not just clicks or conversion volume. Pinterest announced that global advertisers can now access Performance+ campaigns for Consideration, Conversions, and Catalog Sales objectives.

TheOpensource.AI News

Google Unveils Open-Source AI Framework to Accelerate India's Digital Transformation

Google is making significant strides in advancing India's digital landscape with the introduction of an open-source AI Agent Framework powered by its AI model, Gemini.

Announced at the Google for India 2024 event, the initiative seeks to integrate India's Digital Public Infrastructure (DPI) with cutting-edge AI technology, aiming to make a substantial impact across various sectors.

Introducing the Gemini Agent Framework

This new AI Agent Framework will serve as a unified platform, enabling users to access products, services, and information from multiple providers, regardless of the app or network. Whether it's electronics, agricultural supplies, educational courses, or job opportunities, the framework connects users to the resources they need.

The platform will support multiple Indian languages and voice interactions, ensuring accessibility for all. Initially, it will focus on two key areas: agriculture and skill development.

Farmers will benefit from the platform by receiving essential information such as weather updates, seed details, harvest timings, market prices, and access to credit, helping them boost efficiency and productivity.

The framework will also assist users in finding relevant courses and job opportunities. Google has partnered with India-based AI company Glance to make the agent available to its 235 million Android users across the country, helping equip young people with critical skills for today’s job market. Additionally, Google is prioritizing data security, especially in sensitive sectors like healthcare and finance. Earlier this year, Google Cloud enabled organizations to store their data within India to comply with local regulations.

Later this year, businesses will have the option to store data and run machine learning processes for Google's advanced Gemini 1.5 Flash model, all within India. This offers organizations greater control and security while using Google’s AI technology.

DPI in a Box

India's approach to DPI has become a global model. To share this success, Google is collaborating with the EkStep Foundation to launch "DPI in a Box." This initiative will provide a ready-to-use model based on key elements of India’s DPI, including:

• Open Networks: Enabling seamless collaboration between different platforms.

• Digital Identity: Leveraging India’s digital identity system for secure authentication.

• Digital Public Goods: Sharing best practices and insights from India’s DPI journey.

By offering these tools, Google aims to help other countries develop their own digital infrastructure and experience similar benefits as India

How Open Source AI is Revolutionizing Population Estimates and Climate Solutions?

To aid the ongoing efforts of researchers, governments, nonprofits, and humanitarian organizations globally, Meta’s Data for Good program is making the first set of training data and sample code for constructing its AI-powered population maps available as open-source.

As the world faces the growing threat of climate change, Meta’s AI-driven population maps, along with the data behind them, present valuable opportunities for better disaster preparedness and climate adaptation planning. These tools can improve estimates of global flood exposure and guide investments in sustainable energy and resilient infrastructure.

By open-sourcing these resources, Meta aims to empower researchers to gain new insights and accelerate the development of sustainable solutions worldwide.

The Need for Better Population Maps

Accurate population estimates are common in many developed nations, where governments use sources like census data and tax records to inform decisions on service delivery. However, in some low- and middle-income countries, population data is outdated or incomplete, and remote communities may be excluded from official counts. This can leave these uncounted populations outside the reach of vital services and programs.

To address this issue, Meta began mapping the world’s population using AI and satellite imagery in 2017. Partnering with institutions such as Columbia University’s Center for Earth Science Information Network (CIESIN) and WorldPop at the University of Southampton, Meta has published hundreds of high-resolution population maps and datasets. These maps have been utilized globally by governments and nonprofits for initiatives like COVID-19 interventions and clean water distribution. As demands for natural resources and energy grow, accurate population data also becomes crucial for sustainability efforts.

For instance, the World Bank used Meta’s AI-powered population maps to identify potential COVID-19 hotspots in Kinshasa, Democratic Republic of Congo.

Overview of Meta’s AI-Powered Population Maps

Meta’s Data for Good program creates AI-powered population maps that estimate the number of people living within 30-meter grid tiles globally. These maps use computer vision techniques—similar to those used in photo object recognition for visually impaired individuals—to detect human-made structures in satellite images. The AI model’s results are then combined with population estimates from CIESIN to determine the population in each tile.

Beyond total population counts, these maps also provide demographic breakdowns, including children under five, women of reproductive age, youth, and the elderly.

Scientific evaluations have found Meta’s population estimates to be among the most accurate globally for mapping population distribution across various regions and applications. For example, a 2022 study published in Nature – Scientific Reports compared population density estimates for flood risk mapping in West Africa. Other studies have focused on areas like landslide risk and malaria eradication in countries such as Haiti, Malawi, Nepal, and Rwanda.

Open-Sourcing the AI Training Data

This initial release includes nearly 10 million labels and over 126 gigabytes of satellite imagery, with human annotations indicating the presence of buildings. The labeled satellite images, which date from 2011 to 2020, are valuable for training next-generation computer vision models, like Meta’s Segment Anything, to better identify buildings across various landscapes. Additional data and code for computer vision training will be released in the future.

By open-sourcing these tools, Meta enables population mapping collaborators like CIESIN and WorldPop to continue advancing population estimation accuracy while reducing development costs for research teams. These resources also allow researchers to enhance building detection methods, particularly with more recent satellite imagery. Future data releases from CIESIN and initiatives like GRID3 will further improve spatial resolution and accuracy through collaborations with African nations, supporting sustainable development.

TheClosedsource.AI News

OpenAI Launches ‘Canvas’ for ChatGPT: A New Interface for Writing and Coding Assistance

OpenAI has launched a new workspace interface for its ChatGPT chatbot, called Canvas. This interface appears alongside the main conversation, enabling users to highlight specific sections or instruct ChatGPT to focus on particular tasks, such as writing or coding assistance.

Currently, OpenAI is rolling out Canvas in beta for ChatGPT Plus and Team subscribers, with Enterprise and education users set to gain access next week. Built on the GPT-4o model, Canvas can be manually selected from the model picker during the beta phase, and it will eventually be available to free users once testing is complete. 

Within the Canvas interface, users can highlight sections to specify where they want ChatGPT to concentrate. In this mode, ChatGPT acts as a copy editor or code reviewer, providing feedback and suggestions while considering the broader project context. Users can also request quick actions, which will reflect on the main project. A menu offers shortcuts like “Suggest edits,” “Adjust length,” and “Change reading level” for writing tasks, while coding shortcuts include “Review code,” “Add logs,” “Add comments,” and “Fix bugs.” Coders can also request ChatGPT to convert code into other languages, including JavaScript, TypeScript, Python, Java, C++, and PHP, directly from the Canvas interface.

OpenAI explained that Canvas is designed to serve as a creative partner, allowing for a deeper understanding of context and providing targeted feedback and suggestions. It also features enhanced situational awareness to determine when minor edits are needed and when full rewrites are more appropriate.

Additionally, Canvas can open automatically when ChatGPT detects it could be useful, or it can be manually opened by prompting ChatGPT in the main interface.

Mira Murati to John Schulman: Leadership Shakeup at OpenAI Signals Internal Struggles

The resignation of Mira Murati, OpenAI's Chief Technology Officer, has sent ripples through the tech industry, highlighting deeper issues within the AI powerhouse. Murati, who played a pivotal role in product launches, stepped down amid growing friction between technical safety concerns and commercial demands. Her departure follows a period of intense AI development and internal conflict over product readiness.

Murati was a central figure at OpenAI, overseeing significant launches like GPT-4o. However, her position placed her at the heart of tensions between research teams focused on safety and the company’s commercial division, which was keen to maintain OpenAI’s leadership in the AI race. These conflicts became especially apparent with the releases of GPT-4o and the newer o1 model, which some internal teams believed were launched too quickly to compete with rivals like Google, according to a Fortune report.

The departure of John Schulman, one of OpenAI’s co-founders, further indicates that Murati’s exit is not an isolated case. Along with other senior figures like Bob McGrew, Chief Research Officer, Schulman’s resignation points to a broader leadership shift. The exodus of long-standing researchers suggests deeper tensions about OpenAI’s transformation from a research-focused nonprofit into a high-pressure, profit-driven company. The push to stay ahead of competitors like Google and Meta has led OpenAI to accelerate product releases, but at a cost. For instance, reports suggest that the internal team had only nine days to perform safety checks for GPT-4o, a timeline many employees found insufficient to meet OpenAI’s safety standards. Murati, who was known for delaying launches to allow more testing, was reportedly unable to prevent this rush, further exacerbating the internal divide.

This leadership turnover raises questions about OpenAI’s ability to sustain its rapid growth. As the company moves toward major business milestones, such as raising $7 billion in a new funding round, these internal tensions could impact its long-term stability. Investors may grow cautious as executive departures increase and conflicts between safety and innovation remain unresolved.

Despite reassurances from OpenAI CEO Sam Altman, who stated that Murati, McGrew, and Zoph left "independently and amicably," concerns about the company’s culture persist. Many within the organization believe its shift from a nonprofit to a for-profit entity has created a divide between its original ethical AI mission and the current emphasis on staying competitive in a fast-moving market. 

At the same time, OpenAI faces another lawsuit from Elon Musk. In August, Musk filed a suit against Altman and OpenAI president Greg Brockman, accusing them of breach of contract and fraud. This is the second time Musk has filed such a lawsuit, after withdrawing the first. The latest claim alleges "betrayal and deceit on a Shakespearean scale," accusing Altman of shifting OpenAI’s goals to prioritize personal financial gain over its original mission.

In the end, the resignations of Murati and Schulman signal more than just a leadership change; they reveal deeper internal conflicts. OpenAI’s rapid expansion and push to commercialize AI seem to be straining its internal dynamics, potentially jeopardizing its position as a leader in AI innovation.

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.

Daily News for Curious Minds

Be the smartest person in the room by reading 1440! Dive into 1440, where 4 million Americans find their daily, fact-based news fix. We navigate through 100+ sources to deliver a comprehensive roundup from every corner of the internet – politics, global events, business, and culture, all in a quick, 5-minute newsletter. It's completely free and devoid of bias or political influence, ensuring you get the facts straight. Subscribe to 1440 today.

Sign up now!

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