Welcome to the "Towards AGI" Newsletter

Author

Shen Pandi
April 30, 2024

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. Each edition brings you the most compelling news and insights from the forefront of General 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 OpenAI, the organization behind some of the most advanced AI models like GPT and DALL-E. Equally, we'll give you an inside look at the initiatives and strategies from ClosedAI, a hypothetical entity committed to pushing the boundaries of closed-source AI development.

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!

Pappaya Cloud Services offers a comprehensive range of cloud solutions including Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS), along with specialized GPU and Virtual CPU (vCPU) resources for high-performance computing needs. Our offerings are enhanced through strategic partnerships with industry leaders like Lenovo, AMD, and Arista, ensuring cutting-edge technology integration for optimal performance and reliability. These collaborations enable us to provide scalable and flexible cloud environments tailored for applications in AI, machine learning, and more. Our complete cloud ecosystem supports seamless integration and management, making it ideal for organizations looking to leverage robust and innovative cloud capabilities. For more information or to get in touch, visit www.pappayacloud.com or contact [email protected].

Lenovo Unveils AI-Centric Infrastructure for Hybrid AI Innovation

Lenovo Group has launched a new suite of AI-centric infrastructure systems and solutions, enhancing Hybrid AI innovation from edge to cloud. These systems, optimized for GPU-rich and thermally efficient operations, are tailored for compute-intensive workloads in sectors like financial services and healthcare, handling large data sets requiring high I/O bandwidth. The offering includes Lenovo TruScale, providing flexible, as-a-service AI workload integration, and is supported by Lenovo Professional Services for easier navigation of the AI landscape. Mr. Jagadeesh Venugopal, COO of Pappaya Cloud, a strategic partner of Lenovo and AMD, praised the compute potential of the SR685a V3 server and its cost-effective AI capabilities, highlighting Pappaya Cloud's goal to capture 2% of the global cloud market with plans to deploy over 5000 servers by 2025 to support AI-driven cloud services.

How GenAI Boosts Efficiency in Handling Customer Feedback

Many businesses are exploring generative artificial intelligence (GenAI) for boosting internal productivity and enhancing customer interactions, yet few have fully integrated it into their operations. Challenges such as upskilling employees, modifying workflows, and integrating new technologies mean that many are stuck in a cycle of ongoing experimentation.

For companies struggling to effectively deploy GenAI, "voice of the customer" applications offer a promising solution. These applications, which involve understanding and responding to customer inputs from various channels, tend to be simpler to implement than those aimed at boosting employee productivity because they require less change in staff behavior. Additionally, the financial benefits of improved customer satisfaction are typically more tangible and measurable.

Historically, many organizations have faced difficulties in systematically capturing, analyzing, and responding to customer feedback due to the large volume and unstructured nature of the data. Generative AI can significantly enhance this process by automating the capturing, analyzing, and responding to customer feedback, making the process more efficient and less resource-intensive.

Here’s how GenAI can revolutionize customer interaction:

  1. Data Capture: GenAI can transcribe customer voice calls into text and is already equipped to handle text from emails and social media. This transcription process is faster and more cost-effective than manual transcription.

  2. Data Summarization: GenAI excels at summarizing customer feedback, often matching the accuracy of human performance but with greater speed and lower cost.

  3. Topic Classification: With predefined categories, GenAI can effectively categorize customer feedback, identify common issues, and suggest actionable solutions.

  4. Sentiment Analysis: GenAI has improved capabilities in sentiment analysis, detecting subtle nuances such as sarcasm or nostalgia, and can perform accurately across multiple languages.

Real-world applications of GenAI are already demonstrating significant benefits:

  • At DBS Bank, a GenAI-powered virtual assistant helps call center agents by transcribing and summarizing calls, and suggesting actions, reducing call handling times by approximately 20%.

  • A public utility has used GenAI to analyze interactions between customers and billing agents, assessing politeness, empathy, and regulatory compliance, and even predicting customer payment probabilities.

Ultimately, GenAI not only helps in understanding customer feedback but also enables companies to craft personalized responses and decide the best internal routes for addressing customer issues. This streamlined approach can lead to more effective and efficient customer service strategies.

Apple and OpenAI Reengage in Talks for Next-Gen iPhone Features

Apple has reinitiated discussions with OpenAI about incorporating the startup's technology into upcoming iPhone features, sources close to the situation revealed. The talks, which are ongoing and private, are exploring how OpenAI's capabilities could be integrated into iOS 18, Apple's next operating system for the iPhone. Previously, interactions between the two companies had slowed, but they have now resumed. Additionally, Apple continues to negotiate with Google about possibly using its Gemini chatbot.

While no final decisions have been made and there's no certainty of an agreement, Apple might choose to partner with either OpenAI, Google, or another provider entirely. Representatives from Apple, OpenAI, and Google have not commented on these discussions. Apple is also developing its own large language model and seeks partnerships to enhance a chatbot feature similar to OpenAI's ChatGPT.

These developments were first mentioned by Bloomberg in March, which also noted discussions with AI startup Anthropic. These advancements come as Apple prepares for its Worldwide Developers Conference in about six weeks, where it plans to unveil new AI software and services. Apple aims to promote these features as being more seamlessly integrated and privacy-focused compared to competitors. Last year, CEO Tim Cook mentioned using ChatGPT and acknowledged several issues that still need resolution.

Conduent Shares Surge Following GenAI Collaboration with Microsoft

Conduent's stock (NASDAQ:CNDT) surged over 7% at Monday's market opening following the announcement of a new partnership with Microsoft (NASDAQ:MSFT) to incorporate generative AI into their business process solutions via Microsoft Azure OpenAI Service.

The collaboration is set to improve quality, enhance productivity, and shorten cycle times for Conduent’s worldwide clientele. The initiative has already launched three pilot programs focusing on areas like healthcare claims management, customer service enhancement, and fraud detection.

Conduent President and CEO, Cliff Skelton, expressed enthusiasm about leveraging their legacy in technology-driven business performance improvement to create advanced business solutions powered by generative AI with Microsoft.

The pilot projects utilize generative AI for several applications, including the intelligent extraction of data from healthcare claims to expedite processing using Azure AI Document Intelligence; enhancing fraud detection in payment systems with Azure Data Factory; and improving response times in customer services with Azure AI Language and Speech Services, as detailed by Conduent.

Financial Strains Hit AI Startups Amid High Development Costs

Since mid-March, the economic strain has significantly impacted several leading artificial intelligence startups. Inflection AI, after raising $1.5 billion but generating minimal revenue, has shut down its original operations. Stability AI has reduced its workforce and replaced its CEO, while Anthropic is urgently trying to bridge a significant financial gap between its modest earnings and its massive expenses, estimated at around $1.8 billion. The burgeoning AI industry, despite its potential, comes with substantial costs, creating financial challenges for tech firms betting heavily on this technology. This is particularly true for a select group of startups that have invested billions into generative AI, the tech behind chatbots like ChatGPT. They are realizing the formidable financial challenge of competing with tech behemoths like Google, Microsoft, and Meta.

Ali Ghodsi, CEO of Databricks, a data warehouse and analysis firm that collaborates with AI startups, has expressed skepticism about the business viability of these ventures, regardless of their technological innovations. Over the past three years, investors have invested $330 billion into approximately 26,000 AI and machine learning startups, as reported by PitchBook. This investment surpasses the $200 billion allocated to 20,350 AI companies from 2018 to 2020. This surge in investment highlights the escalating costs that newer AI firms face, contrasting with the early successes of OpenAI. Backed by $13 billion from Microsoft, OpenAI has leveraged the popularity of its ChatGPT system to establish a revenue-generating business model, including a $20 monthly subscription for its premium chatbot service and enabling other businesses to utilize its underlying technology. 

The expenses associated with developing AI technologies have surprised many in the tech industry, as these costs dwarf those of previous tech innovations like the iPhone, which relied on existing components and required significantly less investment. The advanced, expensive, and scarce chips needed for generative AI, coupled with the high operational costs of each AI query, underscore the financial hurdles facing this sector.

Life2vec AI Death Predictor Issues Warning Against Fraudulent Imitations

US and Danish researchers have developed an AI tool called Life2vec that can predict the likelihood of an individual's death with about 78% accuracy. The tool, which was reported on by the New York Post, bases its predictions on a series of questions about a person's wealth, occupation, place of residence, and health history. Although Life2vec was released in December, it has not been made publicly available yet.

The developers of Life2vec have also issued a warning about numerous fraudulent versions of their tool that have appeared online. These imitations claim to offer similar predictive services but are primarily designed to steal personal and financial information from users. Some of these fraudulent sites even falsely promise to predict users' financial futures in addition to their lifespan.

The legitimacy of Life2vec is backed by extensive research conducted from 2008 to 2020 involving over 6 million Danish participants of varying ages and conditions. This research identified individuals who were likely to live beyond January 1, 2016, and the AI was able to accurately predict who would die before 2020 with more than 75% accuracy. The creators of Life2vec have distanced themselves from the fake websites, emphasizing their lack of association with any of them.

Generative AI Disrupts Entry-Level Coding Jobs in Campus Placements

The impact of automation on job opportunities has become increasingly apparent during the 2023-24 campus placement season at engineering colleges, including prominent institutions like the Indian Institutes of Technology (IITs) and leading private tech schools. Recruitment for traditional software roles has declined, while manufacturing companies have taken the lead in hiring across many campuses.

Placement officers at these institutions attribute the shift in hiring patterns to the widespread adoption of generative Artificial Intelligence (AI), compounded by a global economic downturn in the tech sector over the past year.

At IITs, particularly the older, more established ones, a noticeable number of students from the 2024 graduating class have missed out on campus placements, which are set to conclude by June 30, before the start of the new academic session.

The role of AI in replacing entry-level coding jobs has become more pronounced. "AI isn’t a new concept, but its application has significantly advanced in recent years. This academic year, post-pandemic, has seen a stark manifestation of its impact on campus recruitment," explained Prof SK Shrivastava, placement-in-charge at IIT-BHU. He noted that while recruitment for pure coding positions has decreased—where previously four students might have been hired, now only two might be—this may eventually stabilize as company operations expand.

Prof Shrivastava also highlighted that manufacturing firms are leading this year's hiring, often requiring software skills and selecting students for software-related roles as well. He added that startups, including those achieving unicorn status with international clients, are actively recruiting from the IIT talent pool.

The placement schedule has been unusually extended this year due to the general elections, with some companies, particularly PSUs, postponing their campus visits until June.

Officials from both first and second generation IITs have observed that recruiters are now seeking more specialized software roles compared to the more general positions filled in previous years, with recruitment numbers for these positions significantly reduced.

Prof V Ramgopal Rao, group vice-chancellor at Birla Institute of Technology and Science (BITS), Pilani, and former director of IIT-Delhi, remarked that this is the first year the influence of generative AI on placements has been clearly visible. "The dip in placements for this year's graduating batch is due to generative AI technologies like ChatGPT eliminating entry-level coding jobs, as well as major IT firms scaling back hires due to the economic slowdown. AI has definitely reshaped how the industry recruits fresh talent," said Prof Rao, noting that a significant portion of employees in many large IT companies are underutilized, a hangover from pandemic-driven over-hiring.

OpenAI and Financial Times Strike Deal for Content Licensing

Since 2022, under the leadership of CEO Sam Altman, OpenAI has been at the forefront of artificial intelligence technology, though it has faced challenges along the way. Despite support from tech giant Microsoft, OpenAI has been actively seeking reliable and accurately verified data sources to train its AI models more effectively, aiming to eliminate errors and reduce "hallucinations" in generated content. Addressing this need, OpenAI has recently entered into a licensing agreement with the prestigious Financial Times, adding to its roster of media partnerships which already includes the Associated Press, Axel Springer, Le Monde, and Prisa Media. This demonstrates OpenAI's commitment to incorporating a diverse range of languages in its training data.

As reported by Reuters, the collaboration with Financial Times will enhance the ChatGPT chatbot's ability to respond to queries and provide summaries that link back to the newspaper. The partnership will integrate FT's archived content into the chatbot and will focus on developing new AI features specifically for FT's audience. The financial details of the deal have not been disclosed.

ChatGPT, developed by OpenAI, is a cutting-edge language model that employs deep learning to produce text that is remarkably human-like. It is capable of conducting conversations, answering questions, and adapting its text output across various styles and subjects. With its extensive applications in customer support, content creation, language translation, and education, ChatGPT's robust natural language processing abilities make it a versatile tool in numerous industries.

ChatGPT Gains Enhanced Memory for More Personalized User Interactions

OpenAI recently launched an enhancement called Memory for its ChatGPT system, which was introduced in February. This new feature allows ChatGPT to store information from interactions, such as queries and customized modifications, for future reference. Initially available to a limited audience, Memory is now accessible to ChatGPT Plus subscribers outside Europe and Korea. The Memory function enhances the user experience by personalizing responses in two ways: it allows users to direct ChatGPT to remember specific details for more relevant future interactions, and it gathers insights from similar conversations to improve its understanding and response quality. This upgrade moves ChatGPT towards its goal of being a more adept AI assistant that can adapt to user preferences more effectively.

OpenAI also recently updated how Memory works, as detailed in a blog post. Now, ChatGPT informs users whenever it updates its memory, providing more transparency and giving users control over what is remembered. This allows users to delete any information they no longer want ChatGPT to retain. The practical benefits of Memory are evident in various applications. For example, users who prefer structured meeting notes can have ChatGPT remember to format their summaries with headlines, bullet points, and concluding action items, making the AI a valuable tool for meeting organization.

Entrepreneurs have also used Memory to enhance their social media strategies. For instance, a coffee shop owner can inform ChatGPT about their business, enabling the AI to help craft targeted promotional content for events like new store openings.

Additionally, Memory has proven useful for personalizing special occasions. Parents can tell ChatGPT about their child’s interests, such as a love for jellyfish, to get customized suggestions for themes, like jellyfish-themed birthday cards.

Austrian Group NOYB Files Complaint Against OpenAI Over ChatGPT Errors

On Monday, a privacy advocacy organization based in Vienna, known as NOYB ("None of Your Business"), announced plans to file a complaint against ChatGPT in Austria. The group criticized the AI tool for producing incorrect answers and accused its creator, OpenAI, of being unable to rectify these inaccuracies. NOYB highlighted that ChatGPT often provides false information and OpenAI has admitted its inability to correct such errors or clarify the origins of its data and what it retains about individuals. This, according to NOYB, breaches EU law which demands the accuracy of personal data.

Maartje de Graaf, a data-protection lawyer with NOYB, emphasized that systems lacking accuracy and transparency should not be used to process information about individuals, asserting that technology must comply with legal standards.

The organization pointed out specific instances where ChatGPT inaccurately reported the birth date of NOYB's founder, Max Schrems, and failed to correct this information upon his request, which OpenAI claimed was not feasible. Moreover, NOYB criticized OpenAI for not adequately addressing Schrems’s requests for access to his personal data, a further contravention of EU regulations.

NOYB, established in 2018 as a vocal critic of major tech companies, has urged Austria's data protection authority to investigate and potentially fine OpenAI to ensure compliance with EU laws. This action follows the global excitement around ChatGPT's debut in November 2022, which has since seen legal challenges in several countries, including a temporary suspension in Italy and ongoing investigations in France.

Join the movement - If you wish to contribute thought leadership, then please fill the form below. One of our team members will be back in touch with you shortly.

Form link

Keep reading