On Fri, Jun 26, 2026 at 09:04:36PM -0400, funkymail via cypherpunks wrote:
> > > > > https://ar.anyone.tech/QkHIWrCSy3iXoP_fS-2snSKIdPYBqGHNmPDya89lWHE#QE8jBZoNz... > > > > > https://ar.anyone.tech/ExicPTy-13WSydg1y-DHVOPtR8L6IMDMiMc7ZAGt2_Q > > > > > > > > > > > class Go: > > > > > > > https://ar.anyone.tech/ELH4g_MtmIp76iolJ-Dhvt8iIv26YpOnL4lq1HzQr-I#d9HTHkvo2... > > > > https://ar.anyone.tech/M--JZ_7_HwlQrbc0JDWdtAIBYW48WU2zn3itspcrPvE > > > > > > > > > universe -- i implemented a tiny virtual machine! this was a great > > > celebration as i have been unable to imppement virtual machines for the > > > past decade as one of a handful of powerful CS paradigms that had > > > previously been driving me to spasms when trying to redevelop use of.. > > > > > > ..... > > > > > > eek! > > > > > > https://ar.anyone.tech/U9gwPAH5oYzJOiT_GfEBXtc4SCGEvMIMH_KFMuiHoIM#ZclDwlkrm... > > > https://ar.anyone.tech/SAeLncrOrqpW_p2Lzgkoh7s4UC9WgDgjD0pvzdYMFeI > > > > Let us try now again to implement a tiny machine. One of the hardest > > parts of such an endeavor is using the machine, so maybe that can be put > > at the front somehow. It may be messy. > > > > JE = [0, 1, lambda state: state.sta [[[ uhoh > > > > > > oop > > > > https://ar.anyone.tech/Lultc9dsOLhmVzd5ghntGXUaLOM5J8Y6bGFA866taxc#U2vY-vYFm... > > https://ar.anyone.tech/cFo91pja_9_GiySahG-ZhVIXBSZ2Qgd0dkFSK9MkSfY > > so i dont immediately remember well :s but i recently ended up arguing > the -- oop?? > > https://ar.anyone.tech/THAKi1XsNkdfkAfYLQpXyKJsjs3DAw9dxipXLsNXgd8#cVWTtWThQ... > https://ar.anyone.tech/Og6lSRhfifiHCjUwfZt7p_SS0lqSKOdz1zpH9gojBRg
uh! the turing halting problem with chatgpt and claude. focusing on defining a turing complete language that uses tur0ng's proof-by-contradiction assumption of there being a universal halting decider, to then define a computable syntax rule that excludes diagonal proofs and other contr--
https://ar.anyone.tech/hpjdvyfqc89vTMyv68uyj5uARHdrbViAXrJcP0yfsUI#5fl2J-xQk... https://ar.anyone.tech/6u253XhXFHG7Vp_X4RP-HkIEbWM2lOy9PcbjudHr214
recent possible phrasing [claude] was "closure under diagonalization is logically distinct from turing completeness". unsure if that's anything useful to say bu--
https://ar.anyone.tech/fSKDY6kC7ZHe9zgUnKKN91zY65cLT9POgloJX-ViMr8#9LQlU3oDi... https://ar.anyone.tech/opCDnLribrJyqEmsOxfA0yIOBAIFcGxYQILK0JpF4nM
so, midway earlier patterns said to just implement a halting decider :D it's always fun to try to implement things but i spasm out as soon as anything useful-looking gets involved, but i seemed to be able to start it right out!
so as soon as i had a functioning microvm i froze the work and comm--itted in celebration before anything went wrong cause thats a huge avconplishment for me :D
https://ar.anyone.tech/mgo04RidwMDjZkvnGflrAy0UG0pTapgYa9krnythZM0#Qr3-foyTN... https://ar.anyone.tech/AXEXNFcNt3uEFD4NxqnQjboFT3Rw6kRrKDzg0_NP3Ps
sadly this is presently stuck on an offline machine with its celebratory commit message :/ unsure if cutrent behaviors will transport more of it 0r its concept
ummm but! so thats my current thought for the halting decider goal, is to prqctice making microvms until i trust it to be a more stable behavior. cause there would be other blocks to navigate to do that so we gotta ensure things are ok, and vms are fun and possibly useful somehow!
i tried making a second vm today by iterating 'import opcode' in python. i made a super-succinct disassembler and wrote implementations for all the opcodes in a toy function :D but didn't get as far as actually having it run anything like the first one :/
microvms are intetesting to be near doing because they take a little more human-memory to do :}
https://ar.anyone.tech/SZvB3IwfwpwwBvW5qOiyReogs85Hj9XK0n9SeEMiVZE#1kxrM1wPY... https://ar.anyone.tech/mtQIrdgrG8CRz44E2qcIPXuU4FPy9L17ofoyaCYAWYg
how--
https://ar.anyone.tech/pExc9JWnFK82eUdlz0dRuYH7H2wWqFw0db4v7YRqSBc#qBETjFGvi... https://ar.anyone.tech/9BscdAmfuIOJbCrooJ4UUKW8eMXiNJw9mciCMI07_m8
apppp eeeeeeek
so usually here i might copy over the microvm by hand with my thumbs but it is actually not very interesting and theres not much energy around the copying over approach :s :s i kinda hope to get things off of this computer at some point more easily but i dunno :s
https://ar.anyone.tech/6uTpOotFHbbd4hPUHtg4c7WxG_Ep0VTBKhBI5Ohkkio#1kZvnXfgd... https://ar.anyone.tech/mOXZFg4lT8fXmUGO1o44RVszt0TUbsMcwu9Gg71BAAE
so. what is a vm made of ??? youcan make a vm however you want, but basically: - instructions generally with opcodes and arguments and distinct implementations - machine state, an instruction pointer, optionally other registers and anything needed to do useful things together with the instructions - a storage representation for data and functions, some kind of calling convention - a way to put in code and run it if you're making a static halting decider, you might also want: - algebraic/symbolic descriptions of how each instruction mutates machine state https://ar.anyone.tech/F0EotPMIpxT7SdJbfnNZxbtnVMPSGAk24cTBeZlWgjk#9WQ19wzmM... https://ar.anyone.tech/QNUH0T6djqDChG3_hzajsj7Z-ShzjwJm8WqwJ2b3Qss I am not affiliated with https://ar.anyone.tech .