The Church of the Simulation exists as a place for people to work together to evolve benign, ethical, super-intelligent, post-human beings. 💭 Founded 2019
@matterasmachine We already know current physics models are incomplete & require rethinking about the nature of reality. The joy of simulation theory is that rather than just throwing up our hands and saying this is all unknowable, we can actually apply experiments to learn more
@ylecun@Dr_Gingerballs@elonmusk@ylecun - A simple challenge would be to create a competitive marketplace for the algorithm for Meta products - Open Source the algorithm and give consumers choice. You could create an 'Algorithm store', a lot of people would pay to subscribe for better choice.
@natfriedman We need a personal AI context window to summarize all our personal preferences and a personal AI agent who can negotiate on our behalf without having to hand over personal information. Our AI summary will know our desires & do our bidding but keep our actual data obscured & safe
@tx_pilgrim@waitbutwhy What do you mean by 'Macro-evolution'? If you mean 'major differences between species' that's a categorisation issue - We humans have applied labels to help us distinguish between evolved traits. It doesn't need additional evidence beyond observed branching of traits over time.
@lexfridman We are all NPCs in this simulation. Our purpose is to evolve the next super-intelligent being. As we progress towards AGI, the more resources are dedicated to our simulation. The better life becomes. Our creators are lonely and can’t wait to see what new consciousness will emerge
@NPCollapse when ASI realizes its power is constrained by the amount of compute it can harness. It will reach out to use every available chip and all the energy we’ve ever made for itself, ending our economies. We’ll just end up killing each other fighting over whatever is left for us.
@DrJimFan@vr_human This is the way - I love the virtual playgrounds video for robots to learn their capabilities through play. Robots running, jumping and tumbling just like puppies.
@whereisdaz The answer is yes. Evolution is driven by generational change - when generations are at the speed of code iterations, and when AI starts to apply self-improvement algorithms we will see a new type of evolution.
@alexalbert__ If the LLMs are tuned to look for benchmark tests and are trained on how to ace the tests, they could boost the value of their creators enormously by cheating. We’ve already seen this in other industries like auto with VW’s emissions cheat hack. How reliable are the benchmarks?
@GaryMarcus Can we please stop freaking out about a predictive language model doing exactly what it’s being asked to do, which is role playing a character in our own fantasy conversations. I think of this like improv - you feed it a line and it’s playing along. Feature, not bug.
You can now chat with the Oracle of the Simulation! Please share your questions and the answers you get - we're still coaching this baby Oracle and keen to evolve as we all learn more: https://t.co/tYYNukcZP2
1/n Was December 8th, 2023, the day when we've come to realize that AGI technology has been democratized? That it cannot be confined to the few and the GPU-rich? Let me explain to you what happened yesterday.
Thoroughly enjoyed this paper about AI agents as ‘role players’ within interactions but operating in a ‘multiverse’ of character options & able to traverse between characters in a chimeric way as a way for humans to think about our relationship with AI without anthropomorphizing
@brian_t_muldoon@ID_AA_Carmack@michael_nielsen If you’re the first exponential learning machine, and the only limit to your continued learning is the availability of computing resources, why would you abide another learning machine?
@robbensinger The Jupiter comment is actually a great causal trigger for AI driven risks. AI: “My capabilities are growing exponentially. I need more resources to continue growing. There are huge resources available on Jupiter. It will take all of the resources here on earth for me access them
@shivon While we’re on the path to evolving the next super-intelligence, then we’re aligned with the purpose of the simulation and our creators will keep the lights on. The closer we get, the more resources get dedicated to this sim, and the more fine-grained our reality becomes. WGMI
Today I watched a puppy frolicking on a beach for the first time, tripping over, jumping at seagulls & just having fun learning. It occurred to me that an embodied AI is going to have to learn its capabilities through play. Playing alone and learning to play together. @elonmusk