AI summary
OpenAI berencana merilis model bahasa terbuka yang baru dengan izin yang lebih permisif. Model terbuka ini akan dirancang untuk berjalan di perangkat keras konsumen kelas atas. Sam Altman mengakui perlunya perubahan dalam strategi sumber terbuka OpenAI untuk tetap relevan di pasar. Toward the end of March, OpenAI announced its intention to release its first 'open' language model since GPT-2 sometime this year. The development of this model is being led by Aidan Clark, OpenAI’s VP of research, and is currently in the very early stages. OpenAI is targeting an early summer release and aims to make the model benchmark-topping among other open reasoning models.OpenAI is exploring a highly permissive license for the model with few usage or commercial restrictions, aiming to avoid criticisms faced by other open models like Llama and Google's Gemma. The company is facing increasing pressure from rivals such as Chinese AI lab DeepSeek, which have adopted an open approach to launching models. Meta's Llama, for instance, has racked up over 1 billion downloads, showcasing the success of open models.OpenAI intends for its open model to run on high-end consumer hardware and possibly allow developers to switch its 'reasoning' on and off. CEO Sam Altman has acknowledged the need for a different open source strategy and has promised thorough safety evaluations for the upcoming model. If the launch is well-received, OpenAI may follow it up with additional models, potentially including smaller ones.
Langkah OpenAI ini merupakan pergeseran strategi yang sangat signifikan ke arah keterbukaan yang selama ini dinanti-nantikan komunitas pengembang AI. Namun, tantangan terbesar tetap pada keseimbangan antara akses terbuka dan keselamatan model yang harus dipastikan agar tidak disalahgunakan atau menimbulkan risiko etis.
