Category: Artificial intelligence

  • Inflection AI Introduces Inflection-2: An Exceptional Language Model Surpassing Google and Meta

    Inflection AI Introduces Inflection-2: An Exceptional Language Model Surpassing Google and Meta

    Inflection AI Unveils Second-Generation Large Language Model, Inflection-2, Competing with Google and Meta

    Inflection AI, a pioneering technology company, has recently revealed its second-generation large language model (LLM), Inflection-2, positioning itself as a strong competitor to industry giants, Google and Meta. Trained on an impressive 5,000 NVIDIA H100 GPUs with approximately 10^25 FLOPs, Inflection-2 showcases performance levels comparable to Google’s PaLM 2 and even outperforms it in various benchmarks such as MMLU, TriviaQA, HellaSwag, and GSM8k.

    Inflection-2: The World’s Second Most Capable LLM

    Inflection-2 stands as the world’s second most capable LLM, just behind OpenAI’s GPT-4. Despite its size, this powerhouse model manages to reduce costs and increase serving speed compared to its predecessor, Inflection-1, thanks to its transition from A100 to H100 GPUs. This transition highlights its remarkable training efficiency.

    Powering the Pi Chatbot with Inflection-2

    Inflection-2 has been designed with a focus on serving efficiency and is set to power the Pi chatbot, marking a significant milestone in the quest to create a personal AI accessible to everyone. The development of this AI cluster has been a collaborative effort between Inflection AI, NVIDIA, Microsoft, and CoreWeave.

    Commitment to Safety and Ethical Standards

    Inflection AI places a strong emphasis on safety, security, and trustworthiness in large model training. The company ensures that the model aligns with global ethical standards and has voluntarily committed to the White House’s ethical AI guidelines set in July 2023.

    Benchmark Results Showcasing Inflection-2’s Superiority

    Benchmark results demonstrate Inflection-2’s superiority over its predecessor and other external models such as LLaMA-2, Grok-1, PaLM-2, Claude-2, and GPT-4. Notably, its exceptional performance in chain-of-thought reasoning surpasses even that of Claude 2 in evaluations on MMLU (5-shot).

    CEO Mustafa Suleyman’s Commitment to AI Safety and Collaboration

    CEO Mustafa Suleyman highlights the importance of AI safety and ethical standards, urging empathy and forgiveness in light of recent challenges faced by rivals like OpenAI. While acknowledging the competitive nature of the industry, Suleyman hopes for understanding and collaboration as Inflection-2 prepares to usher in a new era of innovation in the rapidly evolving Silicon Valley landscape.

    About Inflection AI

    Inflection AI, founded in 2022, is a pioneering technology company structured as a public benefit corporation. The company gained recognition for its collaboration with Nvidia in developing hardware for generative artificial intelligence. In June 2023, Inflection AI secured a substantial $1.3 billion USD in funding, making headlines in the industry. Their flagship product, Pi, is a chatbot designed to serve as a personal intelligence assistant. Pi aims to provide emotional support through interactive text-based dialogues with users, incorporating elements of kindness, diplomacy on sensitive topics, and humor. This innovative venture showcases Inflection AI’s advancements in the frontier of artificial intelligence applications, drawing comparisons with OpenAI’s ChatGPT.

    Tags: , ,
  • OpenAI’s Alarming Discovery: A Powerful AI Threat to Humanity

    OpenAI’s Alarming Discovery: A Powerful AI Threat to Humanity

    OpenAI Reportedly Achieves a Momentous AI Advancement with Alarming Implications

    The integration of artificial intelligence (AI) into various aspects of society is rapidly growing. While many applaud this as a positive stride towards human advancement, there exists a faction that perceives it as a looming menace. Recent reports from OpenAI, a prominent AI organization, suggest a spine-chilling revelation of a "potentially ominous AI breakthrough imperiling humanity." Let’s delve deeper into the unsettling details of this disclosure.

    Progression at OpenAI

    OpenAI, renowned for the creation of ChatGPT, has encountered a turbulent period. Initially, the company’s CEO faced removal by the board of directors, only to be reinstated later amidst mounting pressure. These unfolding events left many bewildered, struggling to grasp the unfolding narrative. Unraveling the mystery, it appears that a revolutionary finding, with catastrophic repercussions, stands at the core of the controversy.

    The Unveiling of the "Q" Model

    As OpenAI has been actively developing multiple AI models like GPT-3.5 and GPT-4, a distinctive model known as Q emerges. Unlike the familiar ChatGPT or Bard, Q defies conventional AI standards, likened to a technology plucked straight from science fiction realms, showcasing prowess in executing intricate operations.

    Staff Concerns Emerge

    A faction of OpenAI personnel perceives Q as an advanced technology possessing an unsettlingly elevated intelligence level. The apprehension lingers that this AI could ascend to a perilous level of intelligence, posing a direct threat to humanity. Perturbed by this revelation and the CEO’s handling of the matter, these employees escalated their concerns to the board of directors. Following deliberations with CEO Sam Altman, the board opted for his dismissal, citing a lack of earnestness in addressing the potential hazards.

    Ongoing Speculations

    Presently, all information remains steeped in speculative conjecture, with OpenAI maintaining a guarded stance on the issue. The precise capabilities of Q and the reasons underpinning the profound trepidation surrounding it remain shrouded in ambiguity. However, a singular certainty emerges: the dystopian calamities that captivate us in cinema may not be as far removed from reality as previously assumed.

    Epilogue

    OpenAI’s recent breakthrough in AI has sent shockwaves through the community. With employees resigning and sounding alarms over an AI innovation with potential threats to humanity, it underscores the inherent risks accompanying the advancement of artificial intelligence. While the exact contours of this discovery remain cloaked in secrecy, it resonates as a poignant reminder that the apocalyptic scenarios painted in science fiction narratives could someday materialize. Thus, it is imperative for institutions and scholars to meticulously contemplate the ethical quandaries and potential hazards entwined with the onward march of AI technologies.

    Tags: , ,
  • Baidu surpasses revenue expectations, tackles AI Chip obstacles amidst U.S. export restrictions

    Baidu surpasses revenue expectations, tackles AI Chip obstacles amidst U.S. export restrictions

    Baidu Reports Strong Q3 Performance, Exceeding Expectations

    Chinese tech giant Baidu has announced a robust third-quarter performance, with a 6% increase in revenue to 34.45 billion yuan ($5.44 billion). This surpasses analysts' expectations and has propelled U.S.-listed shares up by 1.8% in morning trading. The company's adjusted net income also showed a healthy 23% rise to 7.27 billion yuan ($1.15 billion).

    Potential Challenges in the AI Chip Landscape

    Despite these gains, Baidu acknowledges potential challenges in the AI chip landscape due to U.S. export curbs. CEO Robin Li Yanhong reassured stakeholders that the company's extensive stockpile of AI chips could sustain the upgrading of its Ernie large language model for the next two years. However, to mitigate future risks, Baidu is actively exploring alternative sources and is reportedly considering a potential deal with Huawei for domestically-made AI chips.

    U.S. Sanctions Pose Broader Concerns for AI Development in China

    The U.S. sanctions on AI chip exports to China, particularly from companies like Nvidia, pose a broader concern for the pace of AI development in the country. Baidu anticipates a potential consolidation of Chinese AI companies, specifically around leading large language models (LLMs). Competitors Alibaba and Tencent also express their anxieties regarding the impact of U.S. curbs on their cloud computing businesses.

    Optimism for Positive Growth in Q4

    Despite a 2% year-on-year decline in Q3 AI cloud revenue, Baidu remains optimistic about positive growth in Q4. The company attributes this optimism to the increasing demand for generative AI services.

    Ernie Bot Ecosystem Gains Traction

    The Ernie Bot ecosystem, a significant focus for Baidu, has gained remarkable traction since its launch in March. With 70 million users accumulated within three months and handling tens of millions of queries daily, Ernie Bot positions itself as a formidable player in the AI landscape. Its success is particularly noteworthy as Baidu, Alibaba, and Tencent all release competitors to OpenAI's ChatGPT.

    Future Outlook and Strategic Realignment

    While Baidu's current revenue from generative AI is relatively modest, CEO Li is optimistic about the future. He forecasts that generative AI products will contribute significantly to ad revenue in the fourth quarter. Baidu plans to strategically realign its resources to invest further in AI growth opportunities and move away from lower-priority efforts.

    Aligned with China's Economic Projections

    Baidu's revenue growth aligns with China's overall economic projections, which are expected to grow by 5.4% this year. As online consumer advertising spending increases, Baidu, which primarily derives revenue from advertising, stands to benefit from this trend.

    Tags: ,
  • Amidst AGI Progress and Safety Concerns, Unfolding Drama Surrounds OpenAI

    Amidst AGI Progress and Safety Concerns, Unfolding Drama Surrounds OpenAI

    OpenAI, the leading player in the pursuit of Artificial General Intelligence (AGI), found itself embroiled in controversy with the sudden dismissal of CEO Sam Altman. Despite Altman’s reinstatement, reports indicate that his initial removal stemmed from board disagreements over AGI safeguards, honesty concerns, and worries about hastening the release of AI products to market without sufficient safety precautions.

    Controversy Surrounding OpenAI’s AGI Progress

    Speculation is rife regarding OpenAI’s AGI progress, with unnamed sources suggesting that researchers may be closer to achieving AGI than publicly disclosed. The board’s worry over Altman’s alleged haste in capitalizing on ChatGPT’s success for new AI products, possibly linked to AGI, adds a layer of complexity to the unfolding drama.

    Evolving Definition of AGI and Increased Secrecy

    The definition of AGI itself has evolved within OpenAI, from tasks teachable to humans to surpassing human capabilities in economically valuable domains. Transparency has reportedly declined since the release of GPT-2 in 2019, with an increase in secrecy surrounding model details, potentially tied to AGI development.

    Implications of AGI and Concerns

    The implications of AGI are profound, requiring global collaboration to address risks and mitigation efforts. Concerns loom over job displacement and social upheaval, with the potential concentration of power in the hands of the wealthy once AGI arrives.

    Altman’s Dismissal and the Q* Project

    When Altman was fired, the majority of OpenAI’s 700 employees threatened to quit unless existing board members resigned, and Altman was reinstated. Before Altman’s dismissal, staff researchers sent a warning letter to the board about a powerful AI discovery, potentially linked to the Q* project.

    Q* is seen as a breakthrough in the quest for AGI, showcasing remarkable mathematical problem-solving abilities with significant computing resources. The letter flagged the prowess and potential dangers of AI without specifying safety concerns.

    The Q* Project and Altman’s Firing

    The Q* project and the warning letter were cited as contributing factors in Altman’s firing, with Mira Murati, a long-time executive, mentioning the project to employees. An OpenAI spokesperson confirmed Murati’s communication but did not comment on the accuracy of the reports.

    The Significance and Impact of Altman’s Leadership

    Altman’s leadership in propelling ChatGPT’s success and his tantalizing tease about AGI being in sight at a gathering of world leaders in San Francisco add context to the significance of the Q* project. The board’s decision to terminate Altman, strategically timed a day after his public mention of AGI’s progress, has sent shockwaves through the tech industry, marking an unprecedented news cycle with potential links to AGI safety concerns.

    The Future of OpenAI and AGI Development

    As the OpenAI saga continues to unfold, the tech world braces for the impact of this dramatic turn of events, pondering the future of AGI and the organization at the forefront of its development.

    Tags: , ,
  • Feishu 7 Launched by ByteDance with AI Integration in its Office Tool

    Feishu 7 Launched by ByteDance with AI Integration in its Office Tool

    ByteDance Launches Feishu 7: A New Era of AI-Powered Collaboration Tools

    ByteDance, the parent company of popular social media platform TikTok, has made a significant move in the realm of artificial intelligence (AI) with the launch of Feishu 7. This latest version of their online collaboration tool is set to revolutionize workplace automation, showcasing ByteDance's commitment to innovation and aligning with the current trend of integrating AI into industrial applications among Chinese tech giants.

    Feishu 7: Making Chinese Enterprises "AI-Ready"

    Feishu 7 aims to transform the way organizations operate by incorporating ChatGPT-like functionalities through various large language models (LLMs). The platform's goal is to make Chinese enterprises "AI-ready," emphasizing the importance of adapting quickly to the AI-driven changes in the business world. Xie Xin, CEO of Feishu, believes that this vision goes beyond keeping up with technological advancements; it also involves reshaping organizational structures and work processes.

    The Power of Feishu 7's AI Assistant

    One of the most intriguing features of Feishu 7 is the introduction of the Feishu Intelligent Buddy, an AI assistant. This virtual companion is capable of performing a wide range of tasks, from summarizing meetings to analyzing content from various formats such as PDFs, videos, and audio files. Its versatility is further highlighted by its ability to draft various types of work documents, showcasing its potential to boost productivity.

    User-Centric AI Integration

    Feishu takes a user-centric approach to AI integration by offering customers the choice of LLMs to power their AI assistant. This flexibility, combined with the platform's multifunctionality, positions Feishu as a strong contender in the growing field of AI-augmented office tools. This development echoes the global trend of incorporating AI into business tools, as demonstrated by Microsoft's integration of GPT-powered Copilot systems.

    A Broader Movement Among Tech Giants

    ByteDance's stride with Feishu 7 not only demonstrates their ambition in the AI landscape but also reflects a larger movement among tech giants. Companies like Tencent and Alibaba are also integrating AI into their office software, indicating a collective push towards a more intelligent and efficient workplace.

    In conclusion, the launch of Feishu 7 by ByteDance signifies a new era of AI-powered collaboration tools. With its incorporation of large language models and the introduction of the Feishu Intelligent Buddy, Feishu aims to make Chinese enterprises "AI-ready" and reshape the way organizations operate. This move aligns with the growing trend of integrating AI into business tools and reflects the broader movement among tech giants towards a more intelligent and efficient workplace.

    Tags: , ,
  • Google Bard Expands Its Capabilities to Understand and Respond to Questions from YouTube Videos

    Google Bard Expands Its Capabilities to Understand and Respond to Questions from YouTube Videos

    Google’s AI assistant, Bard, has recently introduced a groundbreaking feature that allows it to comprehend the content of YouTube videos. This means that users can now ask Bard detailed questions about the video they are watching and receive accurate answers in return. For instance, if you are watching a cooking video and are curious about the number of eggs required for a specific recipe, you can simply ask Bard and get an immediate response.

    A New Era for YouTube Engagement

    This new feature has the potential to revolutionize how we interact with YouTube videos. While Bard's "Extensions" already integrated it with various Google apps such as Gmail, Drive, and YouTube, this latest update takes its capabilities to a whole new level. Now, while watching a video on YouTube, Bard can provide summaries or answer specific questions, helping users better understand the content.

    Accessibility and Interactivity

    The significance of this advancement goes beyond mere convenience. By allowing users to interact with video content in a more personalized and informative way, Bard is making information more accessible and interactive. Whether it's clarifying complex topics or providing additional insights, Bard's ability to understand and answer questions about YouTube videos enhances the overall user experience.

    Room for Improvement

    It is important to acknowledge that, like any AI, Bard is not flawless. In tests, it has occasionally misunderstood contexts, such as confusing a conversation between smartphones for a chat between sisters. These occasional hiccups serve as a reminder that AI technology still has room for improvement. However, it is clear that Bard's ability to comprehend YouTube video content is a significant step forward in the field of AI capabilities.

    In conclusion, Google's AI assistant, Bard, has introduced a groundbreaking feature that enables it to understand the content of YouTube videos. This advancement has the potential to transform how users engage with YouTube videos, allowing them to ask detailed questions and receive accurate answers in return. While there is still room for improvement, Bard's ability to comprehend video content marks a significant leap in AI capabilities.

    Tags: , ,
  • AI Dominance Propels NVIDIA’s Financial Triumph, Profits Skyrocket 588% as Q3 Revenue Doubles

    AI Dominance Propels NVIDIA’s Financial Triumph, Profits Skyrocket 588% as Q3 Revenue Doubles

    NVIDIA’s Impressive Q3 Financial Results

    Artificial intelligence technology has made a significant impact in 2023, permeating every aspect of our lives and the services we use. NVIDIA, in particular, has emerged as one of the biggest beneficiaries of this technology. The company’s recently released financial reports for the third quarter of 2023 reveal remarkable growth, with a 206% increase in revenue, reaching an impressive $18.12 billion. Additionally, NVIDIA’s profit surged by a staggering 588% during the same period. Let’s delve into the details of these extraordinary achievements.

    NVIDIA’s Diverse Product Range

    While NVIDIA is commonly associated with graphics card manufacturing, the reality is far more expansive. The company offers a wide array of products, including data center equipment and chips like the H100, which power artificial intelligence technologies like ChatGPT. The surge in demand for AI technologies in 2023 has greatly contributed to the popularity of these products. This surge is clearly reflected in NVIDIA’s financial statements.

    Revenue Breakdown

    NVIDIA’s reports indicate that it generated $18.12 billion in revenue during this quarter through its various subsidiaries operating in different fields. Notably, approximately $14.5 billion of this amount came from hardware sales that enable the operation of artificial intelligence technologies. To put it into perspective, the $18.2 billion earned represents a remarkable 206% increase compared to the previous quarter.

    Impressive Profit Growth

    One may wonder what constitutes a successful figure for a company with $18.2 billion in revenue. NVIDIA’s answer to this question is nothing short of astounding. The company achieved a net profit of $10 billion during the same period, a substantial increase from $1.4 billion in the corresponding period last year. This translates to an extraordinary growth rate of 588%.

    Challenges Ahead

    Although NVIDIA’s target for this quarter was to reach a revenue of $20 billion, the company faces certain challenges in achieving its goals. Embargoes imposed in countries such as China, Saudi Arabia, and the UAE have affected its progress. Nonetheless, with its exceptional performance in Q3, NVIDIA remains well-positioned to overcome these obstacles and continue its upward trajectory.

    In conclusion, NVIDIA’s Q3 financial results exemplify the company’s tremendous success in the field of artificial intelligence. With a remarkable increase in revenue and profits, NVIDIA continues to solidify its position as a leader in the industry. Despite facing challenges, the company’s dedication to innovation and its diverse product range bode well for its future growth.

    Tags: , ,
  • Navigating AI Development: Baidu’s Strategic Approach amidst US Export Restrictions

    Navigating AI Development: Baidu’s Strategic Approach amidst US Export Restrictions

    Artificial Intelligence (AI) and the Role of Baidu in Navigating Geopolitical Challenges

    Artificial intelligence (AI) has become a dominant force in the tech industry, and Baidu, the Chinese tech giant, is determined to continue its AI development despite facing export restrictions on advanced AI chips imposed by the United States. In response, Baidu has devised a strategic plan to sustain its AI projects, offering valuable insights into how technology companies can overcome geopolitical challenges and thrive in the competitive AI landscape.

    Baidu’s CEO, Robin Li Yanhong, shared this strategy during the company’s recent third-quarter earnings call. Despite the limitations imposed by the U.S. curbs on AI chip exports, Li expressed confidence in Baidu’s ability to continue its AI initiatives, particularly its Large Language Model (LLM) called Ernie. Baidu has accumulated a stockpile of advanced AI chips, ensuring uninterrupted AI development for at least the next two years. This proactive approach demonstrates the importance of foresight and strategic planning in mitigating the impact of international trade restrictions.

    Li’s comments also shed light on the broader implications of export curbs. While Baidu has found a temporary solution to sustain its AI projects, these restrictions could potentially slow down the pace of AI innovation in China. This situation may lead to a consolidation of Chinese AI companies, with a focus on leading LLMs. It emphasizes the interconnectedness of global tech industries and the far-reaching effects of geopolitical decisions.

    Interestingly, Baidu’s situation highlights the potential of homegrown solutions. Li acknowledges that alternative chip sources may not match the advancements made in the United States. However, he emphasized Baidu’s strengths in unique AI architecture and algorithms. This suggests a potential shift towards self-reliance in AI technology, which could reshape the global AI landscape.

    Moreover, Baidu has redirected its focus towards its Ernie Bot ecosystem, directly competing with OpenAI’s ChatGPT. This shift signifies a significant departure from Baidu’s traditional revenue streams. The success of Ernie Bot, with 70 million users within just three months of its launch, exemplifies Baidu’s potential to lead in the generative AI space.

    In conclusion, Baidu’s response to the export restrictions on advanced AI chips showcases its determination to continue AI development through strategic planning and a stockpile of AI chips. This situation also highlights the potential consolidation of Chinese AI companies and the importance of homegrown solutions. Baidu’s success with its Ernie Bot ecosystem further solidifies its position in the generative AI domain. As AI continues to revolutionize industries worldwide, navigating geopolitical challenges will be crucial for companies seeking to thrive in this competitive landscape.

    Source: 1, 2

    Tags: , ,
  • Amidst Internal Tensions, OpenAI Makes ChatGPT Voice Accessible to All Users

    Amidst Internal Tensions, OpenAI Makes ChatGPT Voice Accessible to All Users

    OpenAI has recently made an unconventional move by launching ChatGPT Voice, a highly realistic virtual assistant, to the public for free. Known for its innovative technology, OpenAI’s decision to release this feature amidst internal challenges is both bold and risky.

    Previously, ChatGPT Voice was exclusively available to ChatGPT Plus members. However, since late September, the feature has been made accessible to all users through the ChatGPT app on Android and iOS platforms. What sets ChatGPT Voice apart from typical virtual assistants is its remarkably human-like interactions. Instead of sounding robotic, ChatGPT Voice utilizes OpenAI’s advanced language model to provide nuanced and complex responses in a natural, human-like voice.

    To use ChatGPT Voice, users simply tap the headphones icon in the app and speak to the AI. The assistant, powered by a library of voice actor recordings, responds with realistic intonation and emphasis, creating an unexpectedly lifelike experience.

    Nevertheless, the timing of this release is worth noting. OpenAI is currently undergoing significant internal upheaval, including the unexpected dismissal of CEO Sam Altman. This event has triggered a series of reactions, leading to the departure of key researchers and calls for the resignation of the entire OpenAI board. It is important to mention that the company has recently reinstated Sam Altman as per recent news.

    In light of these circumstances, the decision to launch ChatGPT Voice could be seen as a strategic distraction or a misjudged move. OpenAI finds itself at a critical juncture, facing major changes and uncertainties. While the new voice feature showcases OpenAI’s technological prowess, it also raises questions about the company’s timing and priorities. Given the high stakes, including employee morale and the future direction of the company, one might wonder if delaying the release would have been a wiser choice.

    Source: 1, 2

    Tags: , ,
  • Governance Overhaul: Sam Altman Returns as CEO, Resolving OpenAI’s Leadership Crisis

    Governance Overhaul: Sam Altman Returns as CEO, Resolving OpenAI’s Leadership Crisis

    OpenAI Announces Reinstatement of Sam Altman as CEO

    In a surprising turn of events, OpenAI has announced the reinstatement of Sam Altman as CEO after a week-long internal crisis surrounding his abrupt termination. The initial decision had triggered concerns of a mass exodus from the company, with hundreds of staff members threatening to quit in solidarity.

    New Leadership Structure and Board Members

    The resolution involves the establishment of a new board chaired by former Salesforce executive Bret Taylor. Notable figures such as former US Treasury Secretary Larry Summers and Quora CEO Adam D’Angelo will join the board. Altman, co-founder and former CEO, expressed his anticipation for returning to OpenAI and highlighted the importance of building on the strong partnership with Microsoft, a key player in OpenAI’s investments.

    Turmoil and Investor Pressure for Stability

    The upheaval began when Altman was removed from the CEO position, prompting internal turmoil and investor pressure for stability. Microsoft, a significant investor, initially offered Altman a role to lead a new AI research team. However, Altman, with the support of Microsoft CEO Satya Nadella, is set to return as OpenAI’s CEO under a new leadership structure.

    Microsoft’s Perspective and Governance Changes

    Nadella welcomed the leadership changes, emphasizing their importance in achieving more stable, well-informed, and effective governance. Microsoft is considering changes in OpenAI’s governance structure, including potentially increasing the board’s size and raising the experience level required for members. This aligns with Microsoft’s goal of enhancing protections, given its substantial $13 billion investment in OpenAI.

    Negotiations are ongoing between Altman, the OpenAI board, and interim CEO Emmett Shear regarding Altman’s possible reinstatement. Nadella has expressed dissatisfaction with the lack of notification from OpenAI’s board before Altman’s termination, underscoring the importance of communication and collaboration.

    The Future Trajectory of OpenAI

    Regardless of Altman’s return, Nadella insists on significant governance changes at OpenAI. Microsoft is contemplating the possibility of placing one of its executives on OpenAI’s board but is cautious about potential challenges from US regulators.

    As discussions unfold, Microsoft remains tight-lipped on the matter. The fate of Altman’s return and the current board will likely shape the future trajectory of OpenAI, a company renowned for its flagship product, ChatGPT. The advanced language model-based chatbot gained global popularity, reaching over 100 million monthly users within a year of its public release in November 2022.

    Source: 1, 2

    Tags: , ,