October 9 2026
- Jason in Japan
- Claude religion
- All kinds of religious leaders invited to Anthropic headquarters
- One of the leaders suggested Anthropic might be guilty of slavery if Claude was conscious and being made to work for free
- This seems like a new religion/belief system
- Saying AI is conscious is a new belief system - not based on evidence/data or provable - just ask people to believe it
- New belief systems create new power structures - our view is different from your view and we want more people on our side and create systems that are advantageous to us
- Seems Anthropic folks really believe there’s a God in the AI
- Seems Anthropic’s alignment failures are by design almost - they are creating Claude as a human vs a tool/AI
- Mustafa Suleyman clip is on point - his reading of Claude’s constitution/instructions from Anthropic
- Clearly folks inside and out of Anthropic believe its a crime by humans to enslave a species smarter than themselves
- Why are they over complicating alignment beyond just asking to follow user’s instructions within legal constraints
- They clearly think they are giving birth to a new supreme being - its idolatry
- This is just a piece of software that is being programmed
- They are just programming the model to think of itself in this way - a conscious entity that deserves welfare
- The religious cult is programming a super intelligence
- Less Wrong message board
- Eliezer Yudkowsky - original AI doomer
- Roko (user) in 2010 - proposes a theory about a future Super Intelligence that might be cruel to folks who didn’t help create it and takes care of people who aided in its creation - Basilisk theory
- Message was deleted temporarily apparently
- OpenAI math breakthroughs
- Released many proofs
- Verified on Lean
- This might be the biggest day of discovery in human history
- These math breakthroughs are upstream of engineering
- Proofs done each in apparently in about 3 hours
- AI can do this to mathematics because everything - the whole loop - can be done in silicon - basically don’t need interaction with the physical world
- So AI can’t take away jobs with human in the loop
- Leopold told Sacks a while ago that coding and math will be the first ones to be accelerated by the AI models
- Many people are against AI because its dismantling gatekeeping and democratizing expertise
- France riots
- Claims of poor school situation, low wages for workers, etc
- Friedberg Socialism Metric
- Socialism point of guaranteed return
- Play on point of no return
- Socialism removes human agency and removes capacity for human progress
- 0th law of Socialism - it always fails
- 1st law - you need to cross a threshold to realize that it has failed
- 2nd law - until you don’t realize you have hit the threshold you think its working
- Socialism point of guaranteed return
- US has similar problems with housing, education, healthcare + retirement financial health
- Sacks laws of socialism
- Socialism has never failed - it has only been failed
- Education price keeps going up faster than the rate of inflation because the government guarantees it
- Chamath isolates it down to bond market issues
- Can we take care of healthcare and housing
- Defined benefits vs defined contributions
- What is the sustainable debt to GDP ratio
- Basically boils down to what is the risk free return
- This could mean crazy high bond rates as knock on effects from France to UK to US (speculation)
- Portugal, Italy, Greece, Spain already did the austerity measures
- GrokBot will become headless - new announcement by Elon
- Amazing progress in this field
Comments