Raw LLM Responses

I thought DeepSeelR1 coming out months after GPT1 would help more people realize that no one is going to have a monopoly on AGI/AI. At least not for very long. This is the year that we are going to get virtual agents that Sci-fi has called AI for decades now. The only reason it didn't happen last year was the speed of improvement has the venture capital guys gun shy. Why spend a billion on an ROI you won't see for a year? And now with DeepSeek you're asking why spend a billion training a model when it can be done for 6 million? We now have the tools to do much of the drudgework of making and training AI. We have AI that is making PHD level insights and can answer PHD level questions 1 in 4 times without trying to ask again. Tens of thousands of the best minds in the world are working on this project with little iterations and open sourcing it. There is plenty of danger in open sourcing it and democratizing it, but that won't make the end result safer. Facebook is as democratic as social media gets and is incredibly dangerous. Even if we could bake in safey in AI that doesn't mean those responsible for breaking the law would stop if it meant making a buck. The problem as always is *capitalism* not the inherent safety of a technology.