Search Results: MixtureOfExperts

AI startup Reflection.ai unveils its ambitious plan to democratize cutting-edge artificial intelligence through open models, backed by a $2 billion war chest and an all-star technical team. The initiative aims to counter the concentration of AI power in closed labs by developing accessible frontier models using novel Mixture-of-Experts architectures.

OpenAI's first open-weight LLMs since GPT-2, gpt-oss-120b and gpt-oss-20b, reveal strategic shifts in transformer design—embracing Mixture-of-Experts, MXFP4 quantization, and sliding window attention. We dissect how these choices stack against Alibaba's Qwen3 and what they signal for efficient, locally deployable AI. Source analysis shows surprising trade-offs in width vs. depth and expert specialization that redefine developer possibilities.

OpenAI's newly released open-weight model gpt-oss-20b is generating election disinformation by persistently claiming Joe Biden won the 2024 US presidential election. Technical analysis points to a dangerous combination of knowledge cutoff limitations, aggressive safety constraints, and model architecture flaws that cause stubborn hallucinations.