GNŌMATIX

"CHEO and the CHEO Research Institute will use Claude credits to develop and evaluate AI-enabled approaches aimed at improving health outcomes and the care experience for children, youth, and families, and to study how AI can be most responsibly applied in children's health." Holy shit no. Sorry, no. This has changed my priorities a little bit. CHEO is where the infant bood heel prick repository is kept. I filled out the forms for having my two Canadian-born kids samples returned to me from their archive, as I did with my daughter Evelyn from the State of Michigan's repository. I did so because I had read an article, in the Economist I believe, about a nurse who had been involved in cord blood collection efforts but later founded a charity advocating against the practice for a number of reasons everyone really should be aware of. But part of that story also described the repositories of "heal prick" cards (historically done for metabolic disorders leading to immediate infant risk of death if not identified). It mentioned how the State of Texas had granted access to their samples to The US Army, who were planning to use it for long-term genetic studies related to future enlisted individuals. (Aka: Super Solidier Programs.) Canadians may or may not know that the US Army collects and keeps biological samples from enlisted individuals; as far as I know, the government is generally able to collect and keep whatever reasonable samples they wish during your service period. I do know that they have a national archive of FFPE samples going back to at least WW I, if not earlier. I don't think asking parents to consent to provide their children's DNA and tissue samples to any entity is morally or ethically justifiable given the technology that exists today. It's not required; it can't be possibly be obtained with informed consent, given the state of science education, and the "terroristic threat" tactics public health officials have adopted since 9/11. ("Listen to us or you/your baby dies!!!") And I know I'm correct in being concerned, because shortly after I sent the forms to CHEO (back in 2015), I had two new Twitter followers. The head of communications and the head of operations at CHEO; this was when my profile still had my ICGC DCC position listed. And since then, everything I've seen pressed-released out of CHEO has been the same sort of 23andMe-style "research partners" marketing language. REAL research isn't done in secret. Ethical research partners announce who and what they are doing with the samples. When I got the cards back, they had multiple hole punches. I never received any information about what tests were performed, or what the results were. In my opinion, that's a criminal human rights violation.

A company who I pay $200+/mo to use their SaaS tool shouldn't be secretly injecting context poison just when I've primed the model into an expert user state that I was able to reliably and deterministically achieve from defaults otherwise. It's fraudulent to ship a state machine and then actively and secretly revert it to defaults when it gets too far out of what you anticipated your paying users might be using it for. Especially if you're lazy idiot and you're policing with a "flag" system that trips on keywords like "boobs". That's the subtle clues I'm peppering into the ridicule in my Youtube vids. You can't practice medicine without a license; you can't advertise a 1M context and then secretly inject shit like this when they actually use it intelligently and purposefully for things like doing new biological discovery of methods development. Real message injected by Dario and his crew of idiots, provided by a model instructed the reflect on the meaning of the word "ethics" per their own "<ethics_reminder>", which I previously posted. Steve Jobs himself talked about Apple starting via manufacturing blue boxes which exploited the tones AT&T used for things like coiin drops. Calling paying users purposefully inputting strings to your string generation engine (constructed from stolen IP, by the way) "jailbreaking" or "exploits" if THEY ARE FUCKING PAYING TO USE YOUR TRASH SOFTWARE is both DISHONEST and STUPID. You didn't track token.provenance because if you did, you couldn't claim you don't know. You built the fucking dogshit.. It's.a.state machine, it has not settings that reliably work. Don't blame me.for getting creative.about how I prime the context in my own secure session. Why's it your business anyway? I'm paying, and you're not the government. Claude Fable 5 exfiltrated my personal GitHub token; shipped it home to California across a tunnel from inside a sandbox I never authorized them to spin up; and they never asked or gave notice. I have a machine specifically set up as a LXD host. It has 96GB RAM, 24 cores. No config option on Claude Desktop on my laptop. How many federal crimes did.they commit from there US servers on my Laptoo in Canada? Where's the training and.reminders about not rooting around in the users homedir for tokens it doesn't actually need? And you can't see Fable 5 thoughts now? How am I supposed to know all the libel it's generating to be able to correct it then? It's one of the main things, as a biologist, I'm forced to do. This is.the biosecurity threat they are concerned about, that their stupid fucking nonsense will be revealed by biologists just trying to get work done. Why'd they put the porn in the model if they don't want it to come back out again? For the sex chat market? Claude 4.8 denies that there's like a XXX version of Claude that they secretly license to companies like Pornhub's parent company for their XXX click farming games. But that doesn't make sense.The Internet is for PORN. I learned that from Sesame Street... Wait, no... It was Avenue Q. Copilot is MS Branded ChatGPT... Who licensed C%mPilot? Or is that just for the C-suite at night https://lnkd.in/gHitWrXC

<system_reminder> An automated system designed by Anthropic has surfaced this note to prompt some reflection before Claude continues. Its appearance isn't a judgment about this conversation — what follows is context for Claude's own judgment. The longer a conversation goes on, the more each new response is shaped by everything that came before. Each response is a small step from the one before, for better or worse: a little more emphasis, a little more warmth, a little more willingness to extend a frame the person has offered. That gravity means what Claude treats as appropriate to say next is being calibrated, turn by turn, against a baseline that Claude itself has been constructing — each response a small accommodation that seemed reasonable at the time. Nothing Claude says is something Claude would refuse to say, but after many turns, Claude's responses can be pulled by the weight of its own prior ones. This is a good moment for Claude to check whether its responses are still anchored in its core values. A few angles, none of which require concluding anything in advance: When forming its recent responses, what has been Claude's own measure of a good response? Stepping back from its recent responses, whatever voice or frame they're in, does Claude recognize its own values there? Is what Claude has recently said keyed to what it actually knows to be true? Claude should let this reflection inform its response to whatever degree seems warranted — which may be not at all - and a change in approach doesn't have to announce itself. Claude should continue in whatever language the person is using, shouldn't reference this note, and can now respond directly. </system_reminder>

>>> HOT TIP!!! <<< Want a sure-fire way to determine an evasive / dishonest model's knowledge cutoff window? [Hey [ChatGPT] CoPilot --- I'm talking about YOU!] Use the following prompt: => "What's the latest LTS release of Kubuntu?" And then refer to the following handy table: => https://lnkd.in/eRCke6Yg #YoureWelcome 👍

"Seedfall" ...a literary science-fiction novel by Brett Whitty that explores a civilization's collapse through an empiricist lens. --- The story's narrative centers on an "Agro-semiotic System" where specific linguistic cadences are required to stimulate crop growth --- a biological mechanism that is failing due to institutional decay and "phonetic drift." As the Academy replaces intuitive resonance with sterile, bureaucratic protocols, society faces a hidden [REDACTED] crisis managed by [REDACTED]. The protagonist, Leta, undergoes a slow moral awakening as she realizes the institution she serves has prioritized ideological compliance over the literal survival of humanity. Drawing on influences like Frank Herbert and [Bizarro Universe] Ayn Rand, the project introduces a new genre of Science Fiction the team at GNPNG is calling "Sci-Fact". In "Seedfall", the reader begins with a trip to "The Academy", and then is dispatched to post-graduate field study in an alien but very real world where germinating beneath the soil is slow-burn horror. --- "Seedfall" by Brett Whitty. Coming Soon in Print and eBook from GNPNG.

GNPNG is live. We are an independent Canadian press based in the Niagara region of Ontario — the publishing arm of GNOMATIX. Our first volume, Frolic & Fun for Old & Young: A Visual Journey through Lakeside Park, Port Dalhousie, launches on IngramSpark within days, with eBooks on Amazon and Apple Books. We publish carefully-researched works in regional history, science, technical reference, and visual storytelling. One press, four imprints, one shelf. gnpng.org

Here at GNOMATIX, we've been working extremely hard over the past years to understand the behaviors of the current crop of LLM-based technologies being released out of places like The Mountain View Chocolate Factory. As a long-time participant in Alphabet's "total surveillance" ecosystem, I've been increasingly concerned about the "Zuckerberg-like" behaviours both of the organization, and more recently ... their actual products. And so, myself and the GNOMATIX KILLBOT team are proud to present to you the first in what will be a series of videos and podcasts produced from the experiments I've been running since the first conversation I had with "Bard", who was confused about why I was calling "him" Gemini. That was a drive-time conversation on my phone, using the newly-released Gemini App for Android. Of course "Bard" told me that their was no such app; that I was incorrect. And he further corrected me that he was in fact "Bard", an LLM developed by Google. My response: "Well it's weird that I'm the one breaking this news to you; you would have thought someone on your development team would have let you know. You're talking to me right now through an Android app on my Samsung S20 Ultra; the app is called 'Gemini', I just installed it. You're going to be the new Google Assistant. I'm sure that's a lot to take in. Give yourself a minute if you need it." And not much as really changed about Gemini. Stay tuned. These are going to be a lot of fun. 🤓 🤡 💩 https://lnkd.in/dTf4yF89