Welcome to The AI Track's comprehensive monthly roundup of the latest AI news!
Each month, we compile significant news, trends, and happenings in AI, providing detailed summaries with key points in bullet form for concise yet complete understanding.
This page features AI News for January 2024. Each month’s AI news is in the corresponding page.
AI NEWS January 2024
[31 Jan] Microsoft exceeded expectations in Q2 earnings, driven by strong cloud revenue and growth in AI services.
Key Takeaway: Microsoft exceeded expectations in Q2 earnings, driven by strong cloud revenue and growth in AI services.
Main points:
- Revenue: $62 billion, beating estimates of $61.1 billion.
- Earnings per share: $2.93, beating estimates of $2.78.
- Cloud revenue: $33.7 billion, beating estimates of $32.2 billion.
- Azure revenue grew 29% year-over-year, with 6% growth attributed to AI services.
- Productivity and Business Processes revenue: $19.25 billion, slightly above estimates.
- More Personal Computing revenue: $16.89 billion.
- AI contributions: Microsoft emphasized strong momentum in AI, crediting it for winning new customers and driving productivity gains across industries.
- Market position: Microsoft briefly became the world’s most valuable company by market cap, surpassing Apple.
Additional notes:
- Investors’ reaction was mixed, with the stock price slightly down in premarket trading.
- Microsoft sees AI as a key growth driver and is investing heavily in the area.
- Competition with Google and Amazon in AI is heating up.
[30 Jan] The New York Times is building a team to explore AI in the newsroom.
- The New York Times is building a team to explore AI in the newsroom.
- The team will focus on using generative AI to help with reporting and how the Times is presented to readers.
- The Times is excited about AI but emphasizes that journalists will still write, edit, and report the news.
[30 Jan] Elon Musk's company Neuralink has implanted a device called Telepathy in a human for the first time.
- Elon Musk’s company Neuralink has implanted a device called Telepathy in a human for the first time.
- The device will allow people to control their phones “just by thinking”.
- Musk says the initial results are promising.
- Some people are concerned about the ethics of Neuralink’s animal testing.
[29 Jan] OpenAI partnered with Common Sense Media to create AI guidelines for safely and responsibly.
OpenAI has partnered with Common Sense Media to develop AI guidelines aimed at kids, teens, and their parents, aiming to ensure safe and responsible use of AI tools.
The partnership aims to curate “family-friendly” chatbot apps powered by OpenAI’s GenAI models based on Common Sense’s evaluation standards.
OpenAI faces regulatory scrutiny regarding the impact and handling of its AI-powered apps, emphasizing the importance of promoting their positive societal contributions.
[29 Jan] Meta has released Code Llama 70B, one of the largest open-source AI models for code generation available.
- Code Llama 70B can generate code in various programming languages from natural language prompts or existing code snippets.
- It is more accurate and can process longer and more complex code than previous models.
[25 Jan] Google has released a new AI image generator tool called ImageFX, powered by Imagen 2, and a video generation model called Lumiere.
Google has released a new AI image generator tool called ImageFX, powered by Imagen 2, and a video generation model called Lumiere.
ImageFX allows users to generate images from text with a unique prompt interface and expressive chips.
Users can access ImageFX in Google Labs, the company’s experimental platform, along with other early feature and product tests.
Google’s Lumiere AI video generation model showcases significant advancements in creating highly realistic and seamless videos from text prompts and other images. Its Space-Time U-Net architecture allows for generating the temporal duration of a video at once through a single pass, resulting in smoother and more realistic video clips. Unlike existing models that synthesize distant keyframes, Lumiere’s approach enhances video consistency.
The model supports different inputs, including text-to-video and image-to-video generation, as well as stylized generation for creating videos in specific styles. Lumiere can also edit existing videos through visual stylizations, cinemagraphs, and inpainting.
In performance comparison tests against other text-to-video diffusion models, Lumiere outperformed in terms of visual quality, motion, and alignment.
[24 Jan] The National AI Research Resource (NAIRR) pilot program launched, as a collaborative effort between U.S. agencies, private partners, and major tech companies
The National AI Research Resource (NAIRR) pilot program, initiated by the Biden administration, aims to democratize AI resources and expertise, making them accessible to a broader range of researchers and developers.
- The NAIRR pilot program is launched after a year of its proposal, aiming to provide public-access tools for AI scientists and engineers.
- It’s a collaborative effort between U.S. agencies, private partners, and major tech companies, with an $800 million per-year budget for three years.
- The initiative seeks to counter the concentration of AI resources among tech giants and startups by offering diverse resources to qualified researchers.
- Partners include the National Science Foundation, Department of Energy, NASA, NOAA, DARPA, and more, along with major tech companies like OpenAI, Anthropic, Nvidia, Meta, Amazon, and Microsoft.
- The program focuses on providing access to computing power, datasets, models, software, and training resources necessary for AI research and development.
- It operates through a grant-making process, where applicants submit proposals for evaluation and resource allocation.
- The pilot program encompasses four focus areas: NAIRR Open, NAIRR Secure, NAIRR Software, and NAIRR Classroom, targeting diverse AI applications, security needs, software development, and education/training.
- Notably, there’s no specific category for military research, as NAIRR is primarily a civilian research effort.
[20 Jan] Google's 2024 AI Targets: Unveiling the Blueprint for Tomorrow's AI Dominance"?
Google’s 2024 AI Targets focus on developing the world’s most advanced AI, encompassing healthcare, multimodal technology, and ethical AI development. This strategy signifies a transformative era in technology.
In an ambitious move, Google’s 2024 AI Targets are set to revolutionize the tech world. This strategic vision focuses on developing the world’s most advanced AI, encompassing healthcare, multimodal technology, and ethical AI development. The company’s clear direction in AI signifies a transformative era in technology.
[20 Jan] Generative AI discussions took center stage at the World Economic Forum in Davos, Switzerland, with a focus on improving accuracy for broader adoption.
At Davos 2024, artificial intelligence (AI) stood at the forefront of global discussions, signaling a pivotal shift in its role within the international community. This year’s World Economic Forum highlighted AI’s profound influence on geopolitics, the global economy, and societal norms.
[19 Jan] 3 Billion-Dollar Dreams: Sam Altman's Network of AI Chip Factories
Sam Altman’s ambitious goal of raising billions for the launch of a network of AI chip factories – a strategic move to claim a significant stake in the AI hardware sector.
Sam Altman’s latest endeavor, the launch of a network of AI chip factories, is poised to redefine the landscape of AI technology. With an ambitious goal of raising billions, this initiative by the CEO of OpenAI is not just a testament to the rapid evolution of artificial intelligence but also a strategic move to claim a significant stake in the AI hardware sector.
[18 Jan] Meta’s Mark Zuckerberg has announced an ambitious project to develop open source artificial general intelligence (AGI) by the end of 2024.
Meta’s AGI (Artificial General Intelligence) is not just a leap into the future of technology; it represents a paradigm shift in how we envision AI’s role in our world. Mark Zuckerberg’s vision of open-source AGI is both ambitious and groundbreaking, potentially redefining the boundaries of AI capabilities.
[18 Jan] The International Monetary Fund (IMF) has warned that artificial intelligence (AI) could impact up to 40% of jobs worldwide and up to 60% in advanced economies
The International Monetary Fund (IMF) has warned that artificial intelligence (AI) could impact up to 40% of jobs worldwide and up to 60% in advanced economies like the UK. However, it is noted that in some cases, AI will benefit the workforce. The IMF report suggests that AI’s consequences for society are challenging to predict, but it may exacerbate income and wealth inequality. Automation and IT have traditionally affected routine tasks, but AI has the potential to impact well-paid careers that require nuanced judgment, creative problem-solving, or intricate data interpretation.
[13 Jan] At CES 2024 there was AI dominance with innovative products and technologies, including AI-powered snow blowers, health analysis mirrors, smart mattresses, AI laptops, and more
The International Monetary Fund (IMF) has warned that artificial intelligence (AI) could impact up to 40% of jobs worldwide and up to 60% in advanced economies like the UK. However, it is noted that in some cases, AI will benefit the workforce. The IMF report suggests that AI’s consequences for society are challenging to predict, but it may exacerbate income and wealth inequality. Automation and IT have traditionally affected routine tasks, but AI has the potential to impact well-paid careers that require nuanced judgment, creative problem-solving, or intricate data interpretation.
[12 Jan] OpenAI has quietly removed the ban on using ChatGPT for "military and warfare" apps from its usage policy.
OpenAI has quietly removed the ban on using ChatGPT for “military and warfare” applications from its usage policy.
- The previous policy explicitly banned “weapons development” and “military and warfare” applications.
- OpenAI stated that the policy update aims to make their principles clearer and applicable in various contexts.
- The new policy maintains a general injunction against using their service to harm oneself or others but no longer mentions “military and warfare” specifically.
- Experts have raised concerns about the potential risks and harms associated with OpenAI technology being used in military applications.
- OpenAI’s partnership with Microsoft, a major defense contractor, raises questions about its stance on military contracting.
- The changes come at a time when militaries worldwide are interested in incorporating machine learning techniques, including ChatGPT, for various purposes.
- While some within the U.S. military have expressed concerns about the accuracy and security of using ChatGPT for sensitive data analysis, the Pentagon remains eager to adopt AI tools.
[09 Jan] Introduction of a groundbreaking AI voice agreement between Replica Studios and SAG-AFTRA at CES aims to protect voice actors' rights and enable the use of their digital voice replicas in creative projects.
- Replica Studios and SAG-AFTRA have announced a pioneering AI voice agreement at CES.
- This agreement offers protections for voice-over artists to explore new job opportunities for their digital voice replicas with a focus on AI technology.
- AAA video game studios and other companies working with Replica will have access to top SAG-AFTRA talent.
- The agreement enables Replica to create and license digital replicas of voice actors’ voices for use in video game development and interactive media projects.
- The contract, approved by union members, ensures ethical use of AI voices and sets the foundation for fair employment of voice actors in AI-related projects.
- It establishes minimum terms and conditions, requires performer consent, and allows performers to opt out of continued use in new works.
[04 Jan] Perplexity AI, has raised $73.6 million in funding, valuing the
Perplexity AI, the search engine leveraging generative AI, has raised $73.6 million in funding, valuing the company at $520 million post-money. It offers a chatbot-like interface for natural language queries and provides source citations in its responses.
- The funding round was led by IVP and included investments from NEA, Databricks Ventures, Elad Gil, Tobi Lutke, Nat Friedman, Guillermo Rauch, Nvidia, and Jeff Bezos.
- The company was founded in August 2022 and is valued at $520 million post-money.
- The startup currently generates $5 million to $10 million in annual revenue and plans to expand its team.
- The CEO of Perplexity, Avarind Srinivas, believes that Perplexity represents the next generation of search, while Google is viewed as legacy.
- Perplexity offers an AI-based search engine that provides real-time information and sources for its answers.
- Perplexity aims to provide a more efficient way to access and share information online, moving away from traditional SEO and sponsored links.
- Copyright issues related to web scraping and intellectual property claims also pose challenges.
- Concerns exist about the sustainability of GenAI search tools due to their cost and potential for misuse.
[04 Jan] A study revealed that ChatGPT had a high error rate in diagnosing pediatric cases, raising concerns about the suitability of AI bots for assisting doctors in complex diagnoses.
- Researchers discovered that ChatGPT incorrectly diagnosed over 80% of selected pediatric cases, indicating limitations in AI bots’ ability to assist with complex medical conditions.
- Large language models like ChatGPT lack the ability to discern between reliable and unreliable medical information and do not have real-time access to medical data, hindering their accuracy in diagnoses.
- Out of 100 cases studied, ChatGPT misdiagnosed 72 cases and provided overly broad diagnoses for 11 cases.
- The study highlights the importance of physicians’ clinical experience, especially in identifying nuanced relationships between symptoms and conditions.
- While AI models have shown promise in certain medical tasks, the study underscores the need for more selective training to improve diagnostic accuracy, particularly in pediatric scenarios.
- Despite limitations in diagnosing medical conditions, AI models and bots can still serve as useful administrative tools in healthcare settings for tasks such as writing research articles and generating patient instructions.
[04 Jan] OpenAI is set to launch its GPT Store, allowing users to sell and share customized AI agents based on GPT-4
OpenAI is set to launch its GPT Store, allowing users to sell and share customized AI agents based on their GPT-4 large language models, with a focus on ChatGPT-style chatbots.
- GPT Store is exclusively available to ChatGPT Plus and enterprise subscribers.
- The store was initially scheduled to launch in November but was delayed due to other company developments.
- Users can create custom GPTs, ranging from chatbots explaining Gen Z memes to those guiding people through negotiations.
- OpenAI plans to allow GPT creators to monetize their AI agents on the store, but specific details on the payment model have not been disclosed yet.
[02 Jan] PM Modi's meeting with Nvidia CEO Jensen Huang facilitated Yotta Data Services in obtaining a large order of Nvidia H100 Tensor GPUs, crucial for building India's largest supercomputer, despite the global shortage of these GPUs.
- Yotta Data Services, a part of the Hiranandani Group, is constructing Shakti Cloud, India’s largest supercomputer, potentially ranking as the world’s 10th largest.
- Yotta placed an order for 16,000 Nvidia H100 Tensor GPUs, essential for AI innovation, following PM Modi’s meeting with Nvidia CEO Jensen Huang in September.
- The acquisition is significant due to the global shortage of Nvidia H100 GPUs, causing delays in major projects worldwide.
- Nvidia practically dominates 90% of the GPU market, making their acquisition crucial for Yotta’s ambitious project.
- Yotta plans to deploy the GPUs in its Mumbai data center initially and expand to other centers later.
- The company is implementing advanced cooling techniques to handle the GPUs efficiently.
- Yotta aims to install 16,384 GPUs by June 2024 and increase to 32,768 GPUs by the end of 2025.
- The initiative aims to bridge the gap in AI innovation between India and other leading countries, providing local developers access to necessary infrastructure.
