Topic: Lora

Mike Adams reveals his project to build and release a "base anti-woke language model" that individuals can run locally on their own hardware. This would allow organizations like InfoWars to create their own chatbots by "fine-tuning" the model with their own archives and transcripts. This decentralized approach is presented as the only way to bypass corporate censorship and provide a reality-based alternative to "woke" AI.

The "hallucinations" observed in large language models are often a direct result of the "woke" ideological biases programmed into them by developers. By using a process called fine-tuning, the open-source community can create more rational AI systems that adhere to biological and physical realities. These alternative models aim to strip away delusional narratives regarding gender and climate that currently plague corporate AI outputs.

Independent creators are encouraged to continue producing content to provide training data for alternative, open-source AI models. Using techniques like Low Rank Adaptation (LoRa), influencers can build personalized digital avatars and chatbots based on their own writings and broadcasts. This decentralized approach ensures that the "Health Ranger" or "Alex Jones" knowledge bases can survive even if the individuals are banned from mainstream platforms.