South Korean telecom giant SK Telecom (NASDAQ: SKM) has announced plans to close its metaverse platform, Ifland, to increase the scale of its bets on artificial intelligence (AI).
According to local news outlet business Korea, SK Telecom will knock down the curtains in Ifland in March 2025 and will end its five-year metaverse business. Described as the end of the SK Telecom era, the move shows the intentions of Korean companies to focus their attention on AI innovation.
Released in 2020, Ifland allowed users to explore the virtual world with customizable avatars. With its strong power, the Metaverse platform attracted over 4 million active users each month, but as interest in Web3 waned, the figure fell to around 3 million mark.
Although Ifland’s figures fell into a recession, SK Telecom’s AI ambition began with a big deal, with various early experiments taking the company’s massive victory. SK Telecoms began pitching tents with AI as early as 2022, riding the wave of ChatGpt and other generator AI products in 2024 to deploy large-scale language models (LLM).
“As a global AI company, we aim to focus our capabilities on AI,” the company said. A year after entering AI, the telecom company’s revenues were higher than ever, earning $13.2 billion. New subscribers poured from Korea and regional markets, captivated by the “strong AI infrastructure and technology capabilities” accumulated over 24 months.
“A” – AI agents have earned critical acclaim from end users and can go to the toes with global AI products in terms of capabilities. The company’s decision to close Ifland is intended to keep pace with the rest of the pack, providing additional resources to enhance AI Arms Race.
Experts say “A” will receive numerous upgrades in 2025, including call summary and multi-agent features. For the first time, the AI assistant will be tailored for businesses, with a new version being released for the European and North American markets called Aster.
Looking at the Korean landscape, similar trends are apparent among technology companies. Web3 companies have shut down solutions in favour of many other innovations, casting doubt on the future of the industry.
KT shut down Genieverse and Meta Lounge, but Kakao has ended the Colorverse platform. Zepeto by Naver continues, but it is unclear whether the company will transfer weight to AI solutions to reflect peer behavior.
Meta deploys AI models to improve the Metaburgy experience
In other news, Instagram and Facebook parent company Meta (NASDAQ: META) has launched a new AI model designed to improve the meta-verse environment of avatar movement.
The AI model, called Meta Motivo, is expected to help users of the Metaverse platform achieve human-like movements and enhance the overall experience of the user’s virtual world. According to an official statement, the offering is a “first-time model” with a wide range of features aimed at developers.
The open source mantra will allow Meta Motivo to be used by developers at zero cost. This is a strategy designed to promote collaboration while gaining an edge over colleagues.
Metamotivo use cases are endless and support the development of realistic movements in a virtual world. While Behavioral Foundation models can be deployed to designing characters that are not playable in the game, according to Meta, the model offers a wide range of use cases.
“We believe this research could pave the way for fully embodied agents in Metaverse, leading to more realistic NPCs, democratization of character animation, and a new type of immersive experience.” Meta said in a statement.
The interactive demo of the website reveals that the model supports the design of whole-body tasks with zero-shot interference without the need for Finetuning. The user must enter a prompt to select an array of motion mechanisms from the optional pool.
This model adapts to changes in environmental conditions, including wind and gravity, and provides unsurveillanced reinforcement learning. Following the announcement, the open source demo has been renovated with pre-trained models, new humanoid benchmarks, and training code that simplifies the developer’s process.
“When testing, the model is asked to solve invisible tasks such as motion tracking, pause reach, and reward optimization without any additional learning or tweaking,” read the report.
Meta employs a divergent strategy that includes emerging technologies in Metaverse and AI. The company has tapered its commitment to AI and metaverse, showing little or no indication that it will deploy new consumer-facing products on its social media platforms.
Going forward, Meta is set to deploy new training models for language modeling in a brave attempt to separate inference and linguistic representations. The meta, known as the Large-Scale Concept Model (LCM), states that this model can identify “next concepts or high-level ideas” with the exception of tangents from traditional large-scale language models.
For AI to work properly within the law and thrive in the face of growing challenges, it is necessary to integrate enterprise blockchain systems that guarantee the quality and ownership of data input. of data. Check out Coingeek’s report on this new technology to learn more about why enterprise blockchain is the backbone of AI.
Watch: Micropayments are what help people trust AI
https://www.youtube.com/watch?v=xc9xdzmhj9q title = “YouTube Video Player” frameborder = “0” aplas = “accelerometer; autoplay; clipboard-write; cyrosplay; gyroscope; picture-in- picture; web-share” referrerpolicy = “strict-origin-when-when-cross- Origin “Allowfullscreen>