Skip to main content
Skip table of contents

Mecletter March | ‘This’ is more important than the best LLM, the AI ​​ecosystem is changing 💫

Date of creation : 2024-03-26

Hello, dear readers of the Meckletter!
I greet you in the last week of March, where the spring energy is palpable. 🙇

After the February Meckletter was released, many of you congratulated and supported the revamp, and riding that momentum, today I have brought valuable insights from two IT experts. We will examine the trends of generative AI that are changing and need to change, and also introduce real-world implementation cases in the industry.

Feel the tech power that fills up as you scroll down with the Meckletter today.
🔔The Meckletter comes to you on the last Wednesday of every month🔔

More important than the best LLM: ‘this’, the AI ecosystem is changing

Beyond AI to AGI 🤖

Many experts say that the ultimate destination of digital transformation is Artificial General Intelligence (AGI). They argue that generative AI, which is already surprising us, is part of the journey toward that destination. AGI is an AI that possesses human-level intelligence and can think and learn independently. Unlike existing AI, which is trained by humans through vast datasets, AGI learns ‘on its own’. You might think of ‘Jarvis’ from the Iron Man movie or ‘Samantha’ from HER.

Big tech companies like Google, Amazon, and Meta are also actively researching AGI. Google is pushing forward with AGI development through ‘DeepMind’, while Amazon is doing so through ‘Amazon AGI’. Earlier this year, Meta's CEO Mark Zuckerberg declared that ‘Meta's long-term vision is to build AGI’ and announced plans to secure a total of 600,000 AI chips within the year. With this demand, Samsung has also recently announced the establishment of a dedicated computing lab for AGI chip development.

While AGI may seem like a distant future, the advancement of multimodal technology and the semiconductor chip development supporting it are accelerating predictions about its arrival. On March 19, at NVIDIA's annual developer conference ‘GTC 24’, CEO Jensen Huang stated that ‘AGI will emerge within 5 years’. The AI chip ‘Blackwell’, which reportedly contains 208 billion transistors and 10 trillion parameters, was also unveiled. OpenAI's GPT-3 has 175 billion parameters, Google's Gemini has 1 trillion, and the human brain has about 100 trillion, so Jensen Huang's bold claim does not seem exaggerated.

Since the release of text-based Chat GPT at the end of 2022 and the emergence of multimodal AI ‘Sora’ earlier this year, there has been a dramatic advancement in how AI interacts with humans in just one year. And now, the countdown toward AGI with human-level intelligence has begun. Amidst the rapid changes spreading across industries, there are not many predictable factors, but it is clear that if you are lagging behind in this trend, catching up later will be quite challenging.

Today, we will have Group Leader Kim Han-soo from Megazone Cloud's ADC (AI & Data Analytics Center) join us to explore what companies should focus on in the rapidly changing AI market, along with successful implementation cases from Megazone Cloud.

 

Best LLM vs. Optimal LLM ⚖

[Read the full article](https://www.megazone.com/meclletter_mar_content01/#:~:text=함께 알아보려고 합니다.-,최고의 LLM vs. 최적의 LLM,-(제목의 의미를 눈치채셨나요)

When Media Meets AI

The Online Transformation of Everything Created, Broadcasted, and Viewed 🎥

Broadcasting and online media have historically had different perspectives on each other. It is true that the quality and standardized architecture of digital broadcasting, which has established itself as high-quality media, created significant barriers. However, the axis of media has shifted online since COVID-19. The form of broadcasting based on the internet has become a major service in terms of delivering high-quality media content and providing viewing environments.

In this context, broadcasting technology has also moved to an online base, and the rapid online transformation is taking place not only in content consumption but also in production environments. Now, with the advancement of cloud broadcasting technology, the phrase 'anytime, anywhere' has become a new trend not only in content consumption but also in the broadcasting technology sector of production.

Broadcasting Technology Automated by Cloud and AI ⚙️

AI is a strong service in the cloud. The computational power of AI is maximized on cloud platforms with strong computing power and scalability. Recently, the possibilities of utilizing such AI in broadcasting technology are also increasing. It has become possible for artificial intelligence to edit based on the analysis of an anchor's face by remembering it and providing quick searches.

Tasks that previously required a person to manually edit and produce while watching the screen are now transforming into 'automation' using AI. Megazone Cloud is at the forefront of providing these platform-based AI services and applying them across various industries, including the broadcasting media sector.

 

방송 제작 환경을 개선하는 AI 👩‍💻

전문 이어서 읽기

🌎Go Global🌎

Are you interested in expanding overseas?
Let's also read about the trends in the global IT market!

Recently, a prediction was announced that the AI market size in Vietnam will surpass 1 billion dollars by 2026. We have delved into whether this prediction can become a reality and the substance of the rosy outlook for the Vietnamese AI market.

 

On April 26, in celebration of World Intellectual Property Day, we will examine how our country and overseas are responding to AI creations.

 

When a company establishes overseas branches or subsidiaries for global expansion, multiple AWS Accounts will be created. At this time, companies want to manage several AWS accounts in conjunction with their organizational structure. We will inform you how to efficiently manage these resources.

 

👉For inquiries about global expansion, contact us here!

JavaScript errors detected

Please note, these errors can depend on your browser setup.

If this problem persists, please contact our support.