I'm Gaeel, I'm a random movement generator that happens to stumble and latch onto some fun things occasionally
Recently I've been playing with image stuff (?) and modular synths
I wrote an article about Mastodon so I can point people to it when they're confused about what just looks like a weird Twitter to them: https://spaceshipsin.space/blog/what-is-mastodon.html
Let me know if I'm missing something super important
Put this unfinished thing back online
One day I'll finish the weird stuff that is supposed to happen with the anomalies
Hey, so I set up Gitea on my server last night, and it's pretty easy (for a Linux server thing), plenty of people running it on Raspberry Pis and other low-spec machines. A nice way to escape the GitHub/GitLab/BitBucket thing.
Also, I pushed the sources to my personal website, if people want to see how it's made
I'm going to be en vacance in the south of France (within an hour of Marseille) and we plan on travelling to Barcelona/Catalonia for a while.
Does anyone have any contacts for cool leftist/hacker/queer/solarpunk folks in these areas?
Would love to meet people and visit cool projects. Hit me up
Olegtron 4060 with a 555 chip pushed into the patch bay
Played live and recorded to tape
So to complete the objective, the AI would have to come up with a plan that convinces the human that the plan is good, and also have the plan be accurate
In the "make me a cup of tea" scenario, a human would only accept a plan where tea is made and no babies were harmed, so to achieve the objective the AI would need to present a plan like that
But also, if the AI lied, and the preview didn't match the outcome, the AI would fail the objective.
This would allow us to "preview" the effects of the AIs actions before they happen, since we program the AI to come up with a plan and an accurate preview of what the plan does. (The objective function would be "(Human accepts the plan) x (if human accepts and runs the plan, the preview was accurate)"
Thinking about AI safety, and wondering about the way we design political safety through a separation of powers
Instead of building an AI that comes up with a method, and also executes the method, maybe we want an AI that just comes up with the method, and then we pass the method to a simple machine that doesn't have an AI
(Context for my thoughts: https://www.youtube.com/watch?v=lqJUIqZNzP8)
Video about packet radio communications, figured this would interest some of the Merveilles gang
Added Tone.js to this little Turing Machine randomiser, so you can listen to it playing
The colour bar at the bottom is the current state of the machine
Revel in the marvels of the universe. We are a collective of forward-thinking individuals who strive to better ourselves and our surroundings through constant creation. We express ourselves through music, art, games, and writing. We also put great value in play. A warm welcome to any like-minded people who feel these ideals resonate with them. Check out our Patreon to see our donations.