OpenAi Social Media platform is scary but makes sense

OpenAI releasing its own social media network is a perfect extension of its current developments.

Many of OpenAI old studies as well as numerous recent scientific papers in the AI field, indicate a need for significantly more real world feedback data for optimization towards a real world model

Current methods—for example, thinking models and other architectures—rely heavily on huge amount of feedback. At present, this feedback is mainly generated through other models, which imposes obvious limitations. Particularly for humanising AI, a dedicated social media platform could theoretically serve as an ideal method for gathering enhanced feedback for the next generation of ChatGPT. Considering how much feedback data is used for deepseek – which is currently the most transparent point for evaluation – having access to a lot of actual human feedback at scale will provide a further relevant optimization point and will make AI more artificially human intelligent (for better or worse).

Because of that I’m actually having ByteDance as the most interesting wildcard on my watch list for 2025.

Still, everything is pure assumptions and the actual question if this is good for the users is a completely different one.

What’s your opinion?