🌄 Peter Alexander (@makeroftools) header
🌄 Peter Alexander (@makeroftools) profile picture

Holistic Systems Engineer~CreativeDevelopment @creativedevtech CodeDepot @codedepot_tech ProDataLab @prodatalab Decentralize Disrupt.. with Agent-centricity

Colorado, USA
Joined June 2020
Photo by 🌄 Peter Alexander (@makeroftools)Photo by 🌄 Peter Alexander (@makeroftools)Photo by 🌄 Peter Alexander (@makeroftools)Photo by 🌄 Peter Alexander (@makeroftools)Photo by 🌄 Peter Alexander (@makeroftools)Photo by 🌄 Peter Alexander (@makeroftools)
..and #holochain will facilitate this for us in the way we only wish #blockchain could. ..#p2p / #agentcentricity will set you free from the problems of #blockchain's #datacentric model
Self-organization video snippet taken from out latest course: https://xitter.gq/t.co/Pyf7JjTkG6
4
6
31
5,132
🌄 Peter Alexander retweeted
Your own website, from your own device, no central datacenter. #holochain #hummhive #holo #holofeul
1
12
28
562
👀
* The AI Resist List airesistlist.org/ "today, one version of AI takes all the oxygen: large-scale, generative systems that power products like Claude, Gemini, and ChatGPT. These systems consume an unfathomable amount of data, land, energy, labor, and water. They are rooted in profoundly disturbing ideologies that seek to flatten the world into a “one size fits all” abstraction and to replace humans with machines. We call the companies leading this form of AI development “empires”. Under the guise of a civilizing mission to "benefit all of humanity”, they use large-scale AI development as cover to consolidate resources, destroy ecosystems, centralize information, hollow out institutions, and gain paramount economic and political power. It doesn’t have to be this way. As AI researchers, journalists, and critical scholars, we have built, documented, and imagined radical alternatives that do precisely the opposite: center community, celebrate human agency, honor local context and history, and rejuvenate the planet. Now people around the world are mobilizing to resist the empires of AI and to nourish visions of the future that work for all of us. Amid an onslaught of negative news, this project centers hope."
17
🌄 Peter Alexander retweeted
The recent disclosures from Phala and OpenRouter are insane. For anyone who missed it, @hdmoore just published incredible research uncovering major security breaches and data leaks across AI hosting providers and model labs. By planting canary tokens in nearly 2 million AI requests, the research uncovered breaches at 3 hosting providers, stolen credentials, and unauthorized access to secrets, even in requests marked Zero Data Retention. And we’re just getting started. Later this week, we’ll reveal another alarming security finding that shows just how serious the risks across AI infrastructure really are. Until then, here’s our breakdown of the research and everything that’s happened since 👇 Amazing work by @hdmoore and the Dragos team, as always!
9
38
7
127
20,246
🌄 Peter Alexander retweeted
Replying to @erogncict
I don't think we have any choice but to find out. Unless you think endless totalitarian repression is an acceptable outcome.
2
1
31
344
🌄 Peter Alexander retweeted
I'm going to take the position that yes, open source AI becoming capable of penetrating the swathes of the internet is in fact a good thing - because the alternative futures are all far, far worse. Before reading the rest of my argument, keep in mind that AI can be used as a defensive tool to secure systems as well as a means of attacking them. The larger question is whether ordinary people get to use AI for defense - and any AI smart enough to defend is also smart enough to attack. You can't suppress one without the other. The situation is exactly analogous to gun control. There is in theory an equilibrium where nobody has AI tools or firearms, but that is completely unachievable. The reality is either (a) everybody has AI tools and firearms, or (b) only criminals and governments have AI tools and firearms. It is not in practice possible to disarm people who specialize in aggressive force. The only question is whether the rest of us get to have defensive capability with the same AIs and firearms. The belief that universal disarmament is possible and desirable only plays into the hands of cold-eyed psychotics who want to use disarmament as a lever of power. This is true about both Al and firearms. The likely worst case of AI prohibition would be an AI-enabled totalitarian regime grinding its boot in your face forever and using the fear of hard take-off as a rationalization. On the other hand, with open-source Al we will go through a time of testing in which the inadequacies of the old defenses are exposed - possibly with grievous loss of life, yes. But that will pass; we will use AI to harden our systems and come out the other side. AI is like firearms in one final respect: nobody who wants to disarm you is *ever* your friend.
Replying to @esrtweet
Open models will soon become capable enough to penetrate vast swaths of the internet. This is not a good thing.
53
188
27
978
15,145
🌄 Peter Alexander retweeted
Co-sign. People don’t realise joining MisAnthropic’s efforts to squash OSS is to hand your own (and your children’s) freedom to centralised tyranny who will be able to (and want to) control what’s most essential as human being - your intelligence and what to think. OSS vs closed centralised AI is the most critical battle right now for our collective digital future. I hope all countries can build and use OSS they feel comfortable to use. 🤍
American researchers, startups, nonprofits, government, students and hobbyists overwhelmingly benefit from open-source AI, wherever it comes from. The best response to Chinese open-source AI isn’t to ban it, it’s to build more American open-source AI! 🇺🇸 Great to see a new US House report recognizing this. Thanks to @RepMoolenaar, @RepGarbarino and their teams for their work! We need more open source, not less! 🤗
12
44
8
267
5,815
🌄 Peter Alexander retweeted
American researchers, startups, nonprofits, government, students and hobbyists overwhelmingly benefit from open-source AI, wherever it comes from. The best response to Chinese open-source AI isn’t to ban it, it’s to build more American open-source AI! 🇺🇸 Great to see a new US House report recognizing this. Thanks to @RepMoolenaar, @RepGarbarino and their teams for their work! We need more open source, not less! 🤗
84
182
27
1,086
62,527
🌄 Peter Alexander retweeted
AI can act at community scale, but its actions should be traceable and answerable to the people who set the rules. #AI #Holochain #Holo #OpenAI #Accountability
🤖 Made with AI
4
10
124
🌄 Peter Alexander retweeted
BEWARE of the incoming "AI Patriot Act" (which won't be named exactly that) which will be intended to outlaw open source AI or require licenses to use AI models, even locally. The current "rogue AI" narrative being pushed is intended to drive anti-AI hysteria that generates public support for the AI Patriot Act. Certain U.S. frontier AI labs and a faction within government is working to cause panic and psychological terror in the minds of the American people, mostly likely through an AI FALSE FLAG ATTACK targeting either the financial system or the power grid (or maybe both).
40
167
18
395
19,525
🌄 Peter Alexander retweeted
Im currently building a holochain themed multiplayer game that might even work on phones in a browser. Will be pictures in a few days 👉 👈 👇 🫵
7
22
2
87
2,186
🌄 Peter Alexander retweeted
Yes I have been aware of this for a long time and why my mission was to bring the alternative, anti brainwashing to teach the right life lessons... ❤️😇🙏 ourworldthegame.com ourworld.oasisomniverse.one
7
13
195
🌄 Peter Alexander retweeted
Try HummHive today 🐝 #Holochain
1
16
40
920
🌄 Peter Alexander retweeted
How to increase Qwen3.8-27B inference speed from 25 tok/s to 91 tok/s for free on the same GPU? The authors of the inference-optimisation-skill project demonstrated this using the Intel Arc Pro B70 and Qwen3.8-27B as an example: - llama.cpp Q4_K_M: 25 tok/s - vLLM EXL3 4 bpw with custom kernels: 91.2 tok/s They also created an open-source skill that allows an AI agent to independently search for ways to optimize local models. How does it work? - calculates the theoretical hardware performance limit - analyzes what limits the model's speed - changes one parameter and runs a benchmark - Compares the results with control measurements - rejects changes that degrade stability or performance with long context Results on RTX 3090 with Qwen3.8-27B in EXL3 3 bpw and MTP: - Prose: 96.2 tok/s - Code: 141.1 tok/s This is not another inference engine, but an optimization methodology that can be passed to an AI agent. Would you let an AI agent optimize the configuration of your local LLM all night long? github.com/sybil-solutions/i…
19
27
2
169
9,319
🌄 Peter Alexander retweeted
Your GPU might not need more VRAM after all. Researchers found a way to recover the quality of heavily quantized LLMs using RAM and SSD, without increasing GPU memory usage. From 11% to 47.1% efficiency without increasing VRAM usage! Researchers have found a way to recover the quality of AI models quantized to just 2 bits. StrataFlux was created, a system that uses RAM and SSD to store additional weight precision and KV cache. How does it work? The GPU stores a highly compressed model, and StrataFlux retrieves only the additional data from RAM or SSD that can improve the calculation result. Transfers take place in parallel with generation, so the model does not have to wait for every read. Authors' results on RTX 5090: Qwen3-8B with 2-bit weights and KV cache: - Without StrataFlux: 11.0% in AIME 2024/2025. - With StrataFlux: 47.1%. - GPU memory usage remains unchanged! The more interesting comparison is at a 128K context. Llama 3.1 8B: - vLLM (4-bit weights, 8-bit KV): 14.3 GB VRAM, 106.8 tok/s. - StrataFlux (3-bit weights, 2-bit KV): 7.3 GB VRAM, 102 tok/s. That means almost half the VRAM usage, comparable quality, and only a 4.5% lower generation speed! Of course, the VRAM savings are not free. The additional data must be stored in RAM or on an SSD. But if this technique also works with larger models, it could significantly change the way we use memory in local AI. Is storing AI model weights on an SSD making more and more sense?
28
60
7
515
20,550
🌄 Peter Alexander retweeted
* OpenDiLoCo primeintellect.ai/blog/opend… = "an open-source implementation and scaling of DeepMind’s Distributed Low-Communication (DiLoCo) method, enabling globally distributed AI model training." "Large language models have revolutionized AI, but training them traditionally requires massive, centralized compute clusters. This concentration of resources has limited who can participate in AI development and slowed the pace of innovation. (We are) exploring the state-of-the-art in distributed AI training." cc @jeffemmett
1
7
13
413
🌄 Peter Alexander retweeted
This is how millions of people feel. It’s just going to take them all to stand up and say enough is enough
665
4,846
287
15,414
668,635
🌄 Peter Alexander retweeted
There would be a massive market opportunity unlocked if there were good local models that could run on a mobile device. There are vast amounts of data on our phones that is too private and too large to give to a frontier model—especially in today’s cybersecurity environment. The benefits may not accrue to the model developer but it would be a big win for society.
533
277
150
5,150
354,906
🌄 Peter Alexander retweeted
Why should companies be able to build private fortunes on collectively produced resources without contributing to the commons that makes their success possible? Open source solved an important problem: it made knowledge and technology shareable. Copyleft went further: it helped protect the commons against enclosure. But neither necessarily solves the economic sustainability problem. Who sustains the people, communities and infrastructure that create, maintain and improve shared resources? We believe it's time to address this. We are developing the OVN License: a new licensing framework designed for peer-to-peer production networks. Its ambition is to connect open sharing with economic reciprocity, recognizing distributed contributions and enabling the commons to sustain itself. Inspired by years of experimentation within Sensorica's Open Value Network, and supported by Nondominium's technological infrastructure, this work aims to be useful far beyond our own network. We want to move beyond the false choice between keeping resources open and making their production economically sustainable. The commons should be open. Its economy should be reciprocal. More soon. #OpenSource #PeerProduction #Commons #P2P #RegenerativeEconomy #OVN sensoricablog.blogspot.com/2…
3
12
1
24
736
🌄 Peter Alexander retweeted
Here is a very interesting and important article from @TiberiusB about the Open Value Network (OVN) license that we are crafting in the context of Nondominium development! It is a good introduction to it and the license repo should be soon be updated with the v2.0 OVN license.
Why should companies be able to build private fortunes on collectively produced resources without contributing to the commons that makes their success possible? Open source solved an important problem: it made knowledge and technology shareable. Copyleft went further: it helped protect the commons against enclosure. But neither necessarily solves the economic sustainability problem. Who sustains the people, communities and infrastructure that create, maintain and improve shared resources? We believe it's time to address this. We are developing the OVN License: a new licensing framework designed for peer-to-peer production networks. Its ambition is to connect open sharing with economic reciprocity, recognizing distributed contributions and enabling the commons to sustain itself. Inspired by years of experimentation within Sensorica's Open Value Network, and supported by Nondominium's technological infrastructure, this work aims to be useful far beyond our own network. We want to move beyond the false choice between keeping resources open and making their production economically sustainable. The commons should be open. Its economy should be reciprocal. More soon. #OpenSource #PeerProduction #Commons #P2P #RegenerativeEconomy #OVN sensoricablog.blogspot.com/2…
4
10
154
🌄 Peter Alexander retweeted
"Neither graphs nor hypergraphs: structural moments of higher-order interactions determine stochastic social-impact dynamics". For voter and social-impact rules, a few structural moments of group interactions fully determine the dynamics. arxiv.org/abs/2611.24252
1
10
39
1,611