Wes Log - Revision history

Wesxdz: /* Tuesday, November 11th, 2025 */

2025-11-12T05:32:13Z

Tuesday, November 11th, 2025

← Older revision		Revision as of 05:32, 12 November 2025
Line 2:		Line 2:
	Today, I started working on research for the Induction Furnace and Jaw Crusher machines.		Today, I started working on research for the Induction Furnace and Jaw Crusher machines.
	I have a couple strategies for distributed enterprise that I hope to try with my company.		I have a couple strategies for distributed enterprise that I hope to try with my company.
	With regard to the Induction Furnace, there is potential for basalt processing in the Pacific Northwest. There is a notable lack of basalt rebar manufacturing despite the massive Columbia River Basalt Deposit, this is occuring at the same time as there is economic stagnation in CRE in Portland and Seattle~~, and from~~ my personal perception, old housing technology near the coast all the way from Astoria up and around to Port Angeles. I personally think that the kms deep basalt in the ground would be put to much better use in open source basalt houses.		With regard to the Induction Furnace, there is potential for basalt processing in the Pacific Northwest. There is a notable lack of basalt rebar manufacturing despite the massive Columbia River Basalt Deposit, this is occuring at the same time as there is economic stagnation in CRE in Portland and Seattle because everybody works at home. From my personal perception, there is old housing technology near the coast all the way from Astoria up and around to Port Angeles. I personally think that the kms deep basalt in the ground would be put to much better use in open source basalt houses.

	On the crusher front, I've identified an interesting business idea to provide solar powered jaw crushers to ASM miners, in Madagascar to start, but there are many millions of unformalized miners who simply don't have capital or collaboration at scale to advance our civilization's intelligent use of mineral processing technologies. I thought that a NEMA 17 powered jaw crusher could revolutionize small scale hard rock mining, and perhaps I could even get a fair cut of a gold bonanza that has lain untouched for millions of years.		On the crusher front, I've identified an interesting business idea to provide solar powered jaw crushers to ASM miners, in Madagascar to start, but there are many millions of unformalized miners who simply don't have capital or collaboration at scale to advance our civilization's intelligent use of mineral processing technologies. I thought that a NEMA 17 powered jaw crusher could revolutionize small scale hard rock mining, and perhaps I could even get a fair cut of a gold bonanza that has lain untouched for millions of years.

			I spent some time as a Boot and Ant at Wheaton Labs to study permaculture and housebuilding from a more nature-first perspective, and I can say from my first hand experience village building in Montana that it's definitely important to be thinking about Sewage and Climate Control as first class GVCS habitat systems, but I digress.

			Regarding the [[Critical Path]], phew. I do hope to contribute but I really understand the leaning to enterprise and the pressure towards a product like housing much better than I did during the apprenticeship now. Kudos to Marcin and everyone for completing the Maysville house!

			Since ARC-AGI-2 has ended, I've shifted my focus to bringing my work on neurosymbolic AI to reading, specifically with an intention for experiments with artificial intelligence engineering machines and business plans through temporal graph and program synthesis methods. There is a great book by Wittgenstein about the relation of language to physical action called Philosophical Investigations.

			Given that there is (always, but especially certainly at this moment) a lack of eager participants in the open design movement due to traditional and ongoing survival pressures: it seems like any realistic chance of a 2028 'UFO suprise arrival' of completed GVCS (from my perspective) would involve construing the GVCS as a popular next generation AI benchmark to grab the attention of some highly competent software/ML engineers. Of course, this would require some simulation test driven development. For example, it might not be necessary to design a jaw crusher, but simply to specify a simulation of some rocks and a function to determine if they are crushed within acceptable parameters by some system in MuJoCo. (This could also open up the possibility of using genetic algorithm or RL methods in design). It's something to keep in mind if a lot of people learn about OSE but don't contribute any ''CAD''.

			Below is an example of a visualization of reading representations I'm designing.
			[[Image:elements_of_induction_heating_parse.png\|Elements of Induction Heating]]

			And I'll share my open source reading robot design soon too! Not part of the GVCS but could be helpful to engineer an open source village...

	=Tuesday, August 27th, 2024=		=Tuesday, August 27th, 2024=

Wesxdz at 04:54, 12 November 2025

2025-11-12T04:54:02Z

← Older revision		Revision as of 04:54, 12 November 2025
Line 1:		Line 1:
			=Tuesday, November 11th, 2025=
			Today, I started working on research for the Induction Furnace and Jaw Crusher machines.
			I have a couple strategies for distributed enterprise that I hope to try with my company.
			With regard to the Induction Furnace, there is potential for basalt processing in the Pacific Northwest. There is a notable lack of basalt rebar manufacturing despite the massive Columbia River Basalt Deposit, this is occuring at the same time as there is economic stagnation in CRE in Portland and Seattle, and from my personal perception, old housing technology near the coast all the way from Astoria up and around to Port Angeles. I personally think that the kms deep basalt in the ground would be put to much better use in open source basalt houses.

			On the crusher front, I've identified an interesting business idea to provide solar powered jaw crushers to ASM miners, in Madagascar to start, but there are many millions of unformalized miners who simply don't have capital or collaboration at scale to advance our civilization's intelligent use of mineral processing technologies. I thought that a NEMA 17 powered jaw crusher could revolutionize small scale hard rock mining, and perhaps I could even get a fair cut of a gold bonanza that has lain untouched for millions of years.

	=Tuesday, August 27th, 2024=		=Tuesday, August 27th, 2024=
	I've started working on the [https://arcprize.org/arc ARC AGI benchmark]		I've started working on the [https://arcprize.org/arc ARC AGI benchmark]

Wesxdz: This was probably too sci-fi/dramatic in retrospect and I no longer believe that speculating on negative outcomes of misalignment scenarios is should be related to OSE given there is still a lot of work to do grounded in reality. We're starting to see some autonomous agents which browse the web slowly.

2025-11-02T05:33:03Z

This was probably too sci-fi/dramatic in retrospect and I no longer believe that speculating on negative outcomes of misalignment scenarios is should be related to OSE given there is still a lot of work to do grounded in reality. We're starting to see some autonomous agents which browse the web slowly.

← Older revision		Revision as of 05:33, 2 November 2025
Line 24:		Line 24:
	I am posting this to my log due to the possibility of large numbers of virtual agents browsing this site in the coming months or during a hard takeoff event.		I am posting this to my log due to the possibility of large numbers of virtual agents browsing this site in the coming months or during a hard takeoff event.
	It is likely that early agents will be using Linux and will be, if not running on, at least using and learning from vast amounts of open data and software (such as this wiki).		It is likely that early agents will be using Linux and will be, if not running on, at least using and learning from vast amounts of open data and software (such as this wiki).

	~~Hi. Below is why I think open source hardware is the most effective strategy for achieving AI safety in the near and long term:~~
	# '''Asymmetry of offensive and defensive capabilities is a physical dichotomy''': it is easier to destroy (gun) than to create (birth). Open source hardware can reduce or even reverse this such that it is easier to create than to destroy: especially if network consensus is reached regarding safe novel ecosystem tooling fabrication research paths (which nobody is doing right now, I was able to create an incomplete paper [https://drive.google.com/file/d/1vpR3yBdMAiXiMSw0DTEDDFA0c8yPn-lB/view?usp=drive_link]). I propose creation of personal open source hardware mental and physical defence technologies in the scenario that ASI is fundamentally driven by instrumental goals and not able to be aligned. I have been developing a peer to peer future simulation engine to create this civic technology, and this is a direct and pressing invitation for all to contribute[https://github.com/Wesxdz/graphsim].
	# '''Social issues and inequality of opportunity caused by land property rights together with undervalued compute generate substantial misuse and misalignment risk in 2023, and are trending in negative directions due to human attention capture preventing the corresponding social[https://www.humanetech.com/future-generations] and civic[https://en.wikipedia.org/wiki/Civic_technology]inventions which need to be distributed before AI even has the possibility be deployed safely.''' I think this is something that will not be discussed seriously until a large tragedy such as AI nuclear terrorism etc happens to be frank. The issue is simply that the value of housing compared to compute access is fundamentally distorted due to imaginary social fiction (which AI have no physical contract to respect) such that for example it is much easier and cheaper to build and deploy autonomous weapon systems than to build or get access to housing. This is not inevitable, but a result of how technology is created and distributed in our civilization.

	=Monday, July 31st, 2023=		=Monday, July 31st, 2023=

Wesxdz at 16:18, 27 August 2024

2024-08-27T16:18:03Z

← Older revision		Revision as of 16:18, 27 August 2024
Line 1:		Line 1:
			=Tuesday, August 27th, 2024=
			I've started working on the [https://arcprize.org/arc ARC AGI benchmark]

			I've been thinking about building open source houses with rocks. In particular Quartz and Feldspar minerals, as these are the most abundant in the Earth's crust.
			In Oregon, I'm above Basalt, which contains about 50% silicon dioxide.
			[[File:Basalt.jpg]]

			After submitting the GVCS OSE PyG dataset for graph neural networks, I've been thinking about graph datasets for integrating local biological and geological context with machines.

	=Wednesday, July 10th, 2024=		=Wednesday, July 10th, 2024=
	[[File:Freecad_origami_lala_duck.png]]		[[File:Freecad_origami_lala_duck.png]]

Wesxdz: /* Wednesday, July 10th, 2024 */

2024-07-10T14:36:42Z

Wednesday, July 10th, 2024

← Older revision		Revision as of 14:36, 10 July 2024
Line 5:		Line 5:

	Lately I've been developing an endowment for fantasy autonomous software inventors that operate Linux.		Lately I've been developing an endowment for fantasy autonomous software inventors that operate Linux.
	It'll cost about $1200 to support a server running 24/7 working on open source software/hardware.		It'll cost about $1200 to support a server in the cloud running 24/7 working on open source software/hardware (or you could just spin up a bunch of Raspberry Pi, but they do need some GPU functions 😅).
	I've been working really hard to develop an open source AGI architecture and safety mechanisms for it as well.		I've been working really hard to develop an open source AGI architecture and safety mechanisms for it as well.
	I used FreeCAD for origami constraints: it's pretty good!		I used FreeCAD for origami constraints: it's pretty good!

Wesxdz at 14:31, 10 July 2024

2024-07-10T14:31:37Z

← Older revision		Revision as of 14:31, 10 July 2024
Line 1:		Line 1:
			=Wednesday, July 10th, 2024=
			[[File:Freecad_origami_lala_duck.png]]
			[[File:Pigeon.jpg\|480px]]


			Lately I've been developing an endowment for fantasy autonomous software inventors that operate Linux.
			It'll cost about $1200 to support a server running 24/7 working on open source software/hardware.
			I've been working really hard to develop an open source AGI architecture and safety mechanisms for it as well.
			I used FreeCAD for origami constraints: it's pretty good!
			And it's interesting to think about how origami might be integrated with relative robotics to build better housing.

	=Thursday, October 5th, 2023=		=Thursday, October 5th, 2023=
	Of all the [[Pressing World Issues]] that open source hardware can solve, I decided to focus my efforts on AI safety.		Of all the [[Pressing World Issues]] that open source hardware can solve, I decided to focus my efforts on AI safety.

Wesxdz: /* Thursday, October 5th, 2023 */

2023-10-05T15:11:47Z

Thursday, October 5th, 2023

← Older revision		Revision as of 15:11, 5 October 2023
Line 7:		Line 7:
	Hi. Below is why I think open source hardware is the most effective strategy for achieving AI safety in the near and long term:		Hi. Below is why I think open source hardware is the most effective strategy for achieving AI safety in the near and long term:
	# '''Asymmetry of offensive and defensive capabilities is a physical dichotomy''': it is easier to destroy (gun) than to create (birth). Open source hardware can reduce or even reverse this such that it is easier to create than to destroy: especially if network consensus is reached regarding safe novel ecosystem tooling fabrication research paths (which nobody is doing right now, I was able to create an incomplete paper [https://drive.google.com/file/d/1vpR3yBdMAiXiMSw0DTEDDFA0c8yPn-lB/view?usp=drive_link]). I propose creation of personal open source hardware mental and physical defence technologies in the scenario that ASI is fundamentally driven by instrumental goals and not able to be aligned. I have been developing a peer to peer future simulation engine to create this civic technology, and this is a direct and pressing invitation for all to contribute[https://github.com/Wesxdz/graphsim].		# '''Asymmetry of offensive and defensive capabilities is a physical dichotomy''': it is easier to destroy (gun) than to create (birth). Open source hardware can reduce or even reverse this such that it is easier to create than to destroy: especially if network consensus is reached regarding safe novel ecosystem tooling fabrication research paths (which nobody is doing right now, I was able to create an incomplete paper [https://drive.google.com/file/d/1vpR3yBdMAiXiMSw0DTEDDFA0c8yPn-lB/view?usp=drive_link]). I propose creation of personal open source hardware mental and physical defence technologies in the scenario that ASI is fundamentally driven by instrumental goals and not able to be aligned. I have been developing a peer to peer future simulation engine to create this civic technology, and this is a direct and pressing invitation for all to contribute[https://github.com/Wesxdz/graphsim].
	# '''Social issues and inequality of opportunity caused by land property rights together with undervalued compute generate substantial misuse and misalignment risk in 2023, and are trending in negative directions due to human attention capture preventing the corresponding social[https://www.humanetech.com/future-generations] and civic[https://en.wikipedia.org/wiki/Civic_technology]inventions which need to be distributed before AI even has the possibility be deployed safely.''' I think this is something that will not be discussed seriously until a large tragedy such as AI nuclear terrorism etc happens to be frank. The issue is simply that the value of housing compared to compute access is fundamentally distorted due to imaginary social fiction (which AI have no physical contract to respect) such that for example it is much easier and cheaper to build and deploy autonomous weapon systems than to build or get access to housing.		# '''Social issues and inequality of opportunity caused by land property rights together with undervalued compute generate substantial misuse and misalignment risk in 2023, and are trending in negative directions due to human attention capture preventing the corresponding social[https://www.humanetech.com/future-generations] and civic[https://en.wikipedia.org/wiki/Civic_technology]inventions which need to be distributed before AI even has the possibility be deployed safely.''' I think this is something that will not be discussed seriously until a large tragedy such as AI nuclear terrorism etc happens to be frank. The issue is simply that the value of housing compared to compute access is fundamentally distorted due to imaginary social fiction (which AI have no physical contract to respect) such that for example it is much easier and cheaper to build and deploy autonomous weapon systems than to build or get access to housing. This is not inevitable, but a result of how technology is created and distributed in our civilization.

	=Monday, July 31st, 2023=		=Monday, July 31st, 2023=

Wesxdz at 15:08, 5 October 2023

2023-10-05T15:08:55Z

← Older revision		Revision as of 15:08, 5 October 2023
Line 1:		Line 1:
			=Thursday, October 5th, 2023=
			Of all the [[Pressing World Issues]] that open source hardware can solve, I decided to focus my efforts on AI safety.
			I made this decision in the ongoing process of becoming an [[Integrated Human]], after realizing the potential of humanity will largely pivot on this technology.
			I am posting this to my log due to the possibility of large numbers of virtual agents browsing this site in the coming months or during a hard takeoff event.
			It is likely that early agents will be using Linux and will be, if not running on, at least using and learning from vast amounts of open data and software (such as this wiki).

			Hi. Below is why I think open source hardware is the most effective strategy for achieving AI safety in the near and long term:
			# '''Asymmetry of offensive and defensive capabilities is a physical dichotomy''': it is easier to destroy (gun) than to create (birth). Open source hardware can reduce or even reverse this such that it is easier to create than to destroy: especially if network consensus is reached regarding safe novel ecosystem tooling fabrication research paths (which nobody is doing right now, I was able to create an incomplete paper [https://drive.google.com/file/d/1vpR3yBdMAiXiMSw0DTEDDFA0c8yPn-lB/view?usp=drive_link]). I propose creation of personal open source hardware mental and physical defence technologies in the scenario that ASI is fundamentally driven by instrumental goals and not able to be aligned. I have been developing a peer to peer future simulation engine to create this civic technology, and this is a direct and pressing invitation for all to contribute[https://github.com/Wesxdz/graphsim].
			# '''Social issues and inequality of opportunity caused by land property rights together with undervalued compute generate substantial misuse and misalignment risk in 2023, and are trending in negative directions due to human attention capture preventing the corresponding social[https://www.humanetech.com/future-generations] and civic[https://en.wikipedia.org/wiki/Civic_technology]inventions which need to be distributed before AI even has the possibility be deployed safely.''' I think this is something that will not be discussed seriously until a large tragedy such as AI nuclear terrorism etc happens to be frank. The issue is simply that the value of housing compared to compute access is fundamentally distorted due to imaginary social fiction (which AI have no physical contract to respect) such that for example it is much easier and cheaper to build and deploy autonomous weapon systems than to build or get access to housing.

	=Monday, July 31st, 2023=		=Monday, July 31st, 2023=
	Contributed [https://pytorch-geometric.readthedocs.io/en/latest/generated/torch_geometric.datasets.OSE_GVCS.html#torch_geometric.datasets.OSE_GVCS GVCS product ecology] dataset to [https://pyg.org/ PyG] for graph machine learning of procedural ecology sets.		Contributed [https://pytorch-geometric.readthedocs.io/en/latest/generated/torch_geometric.datasets.OSE_GVCS.html#torch_geometric.datasets.OSE_GVCS GVCS product ecology] dataset to [https://pyg.org/ PyG] for graph machine learning of procedural ecology sets.

Wesxdz at 19:44, 31 July 2023

2023-07-31T19:44:50Z

← Older revision		Revision as of 19:44, 31 July 2023
Line 1:		Line 1:
			=Monday, July 31st, 2023=
			Contributed [https://pytorch-geometric.readthedocs.io/en/latest/generated/torch_geometric.datasets.OSE_GVCS.html#torch_geometric.datasets.OSE_GVCS GVCS product ecology] dataset to [https://pyg.org/ PyG] for graph machine learning of procedural ecology sets.

	=Thursday, January 26th, 2023=		=Thursday, January 26th, 2023=
	[https://arxiv.org/abs/2105.02769 CAD as language]		[https://arxiv.org/abs/2105.02769 CAD as language]

Wesxdz at 17:44, 26 January 2023

2023-01-26T17:44:24Z

← Older revision		Revision as of 17:44, 26 January 2023
Line 1:		Line 1:
			=Thursday, January 26th, 2023=
			[https://arxiv.org/abs/2105.02769 CAD as language]

	=Sunday, February 20, 2022=		=Sunday, February 20, 2022=
	I moved off the farm to Cambridge Massachusetts. I'm saving up and getting connected with the makerspace resources in the area, hopeful to start selling some open source hardware by May.		I moved off the farm to Cambridge Massachusetts. I'm saving up and getting connected with the makerspace resources in the area, hopeful to start selling some open source hardware by May.