This is an interesting philosophy question, and one I don't think we will ever be able to answer. I put my thoughts around it today and gave some basic definitions so people can make their own decision https://vibecodeguild.substack.com/p/is-ai-conscious?r=706kcp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true

Very disappointing to see another model released that we can't touch. Doubt we will ever get to use this with the way things are going. https://vibecodeguild.substack.com/p/intelligence-for-the-elites-only?r=706kcp

Follow-up from yesterday's call. We briefly mentioned how Grok.com and Grok Build sound different, and I was inspired to finally fix it. So, they're the same model, but they're given different persons. Grok.com is told "You are Grok 4 built by xAI" and that's it. It's not given any instructions on how to respond. Grok Build is told "You are an interactive CLI tool that helps users with software engineering tasks". It's then given a long writing output style block called <communication>, which has rules like, "lead with the answer", "restate the work", "never say 'X, not Y' and a banned a list of phrases. So, Grok Build is instructed to be a lot more structured in its response. I think that's why it always feels formulaic and why fiction writing doesn't feel natural coming out of Grok Build I tried to use a short rule that says "talk like grok.com". But that didn't obliterate the instructions well enough. What seems to be working so far today is to point Grok Build to a new system instruction file with the <communication> block deleted. Steps and in this gist: https://gist.github.com/wayneadams/c835db40e2b5423c566fa90043d384f9

Meta's new Muse product is poised to win the AI race because they have the distribution and they figured out how to make AI approachable for normal people. Most people have zero interest in coding some lame SAAS project, but everyone has things in their life that could be simplified and Facebook made this easy with all of the ideas and goal inside Muse. Anthropic went after Nerds, OpenAI went after Nonfunded Obligations, and Meta is going after Normies. Normies pay the bills, this is a smart move and their stock will probably reflect that over the coming months. Here's an overview of why they're positioned to win

People talk about fake news and conspiracies, but Effective Altruism is aggressively placing their agents inside of the mainstream media outlets to control the narrative about AI, and they're not even hiding it. They literally brag about it in public. Do you think China knows they've placed agents in their media as well? I wonder if Trump and Xi talked about an AI Doomsday Cult from Berkeley infiltrating Chinese media or not... As always, it's the same pool of money controlled by Anthropic's seed investors, and Dario's brother in law. They're doing it right out in the open. https://vibecodeguild.substack.com/p/how-effective-altruism-infiltrated?r=706kcp&utm_campaign=post&utm_medium=web&showWelcomeOnShare=true