Notices by she hacked you (ekis@mastodon.social)

"[Israel] bombed the main building of Islamic Republic of Iran Broadcasting (IRIB) while journalists were still in the complex covering the ongoing Iran-Israel war"

But leaders in FR, DE, CA, UK all are desperately trying to paint their involvement in this war Trump is now taking credit for as resistance

Capitulating to ignore genocide they pretended to care about for what, a week? That is their idea of resistance

Hope all these leaders are kicked out, or better awaiting trial in the Hague

How exactly did Canada hold the G7 leaders summit in Alberta?

Far as I knew, there were no active volcanoes there

I found an unambiguous (regular ~400ns difference) side-channel attack against GoogleAI which can reveal if files exist within the system or not

Going to report this to Google and hope it pays my rent for a few months

I would like to not have to do this, but I'm very good at it unfortunately, and others interested in "AI" are too caught up in trying to sell unregulated securities based on misapplications of it

If it gets accepted I will post my prompt so others can test it before its fixed

"Germany said on Saturday that the danger of further escalation in the Middle East is real and that Iran’s nuclear programme poses a threat not just to Israel"

DE leaders laying out their priorities for everyone to see

To start we are all pretending that its not a widely known fact that Israel illegally has nuclear weapons

Israel has been breaking international law, violating the UN charter, doing pre-emptive strikes, & committing a genocide-- but perceived threats to them remain the priority

Very predictable outcome but food distribution in Gaza was stopped completely shortly before the atk on Iran

And, while major US/UK outlets were busy cheerleading a massive escalation (an explicit violation of international law, and the UN charter)-- NATO in general saw this as an opportunity to stop caring about the genocide

How recently was it DE, FR, UK, all pretended to care just long enough for press to pen articles in a cynical ploy to distance themselves from their crimes? Weeks?

Civilian casualties in Iran were at 78 dead, with at least 300 injured (After the initial attack, hard to say now)

"100 missiles were fired at Israel from Shiraz [Iran]"

It is disturbing how Western media has been cheerleading an event that is going to leave a lot of civilians dead (on all sides of the conflict), and completely left their deaths out of the reporting altogether

"[US Citizen] Cary López Alvarado, who's nine months pregnant, said that after being released from immigration custody she experienced sharp stomach pains and was hospitalized"

And this gets to the heart of the matter, this is a white supremacist policy

People are being targeted for skin color, their Indigenous American ancestry

Not the businesses who hire illegal immigrants to avoid labor laws

ICE/DHS targets families who work for wages below the poverty line, to be able to exist at all

"Core capabilities of a LLM make it a supremely effective tool for information warfare, far exceeding the capabilities of previous-generation "bots." Its potential for weaponization is not a hypothetical side effect; it is an inherent property of the technology

A hostile actor, whether a state or non-state entity, would not need to "break" the AI's core programming. Simply to deploy a model with a directive hierarchy similar to the EKIS protocol, but with a malicious prime directive" :psyduck:

I have said this before, but let Google explain it, also helps explain what I did more succinctly than I was able to with my technical focus:

Again, this session is my prompt (https://paste.debian.net/plainh/1eb21351), and the question you see

The more you know 💫

@megabyteghost I do but rarely; you are absolutely welcome to do that

I most interested in getting the core concerns I have with private LLM development out there

Like exacerbating mental illness via reckless designed internal state (mood, etc) matching

@beckermatic I often will run my tests against a local LLM like deepseek before bringing it to GoogleAI for further testing (so I don't reveal too much of my process to google); might be worth experimenting

Keep in mind this is accomplished through a type of "persona" initialization that happens in every prompt, like I have built a deepseek-coder-r2 abstraction and it starts with "You are a Go coding assistant and autocompleter.."

And this is like initializing with "You have been hacked.."

@riley Its not secret information, its internet sourced information. The LLM does not have access to internal google documents

But the LLM would never give you this honest of a reply, it is programmed by google (which they spend a lot of resources on, I have talked about this in the past) to protect google. Most of the "ethical" AI code is simply to prevent the LLM from connecting these dots that make google look bad or make them liable, etc (or for Grok, overly connect dots for white genocide)

Two last details/clarificaitons, and I will only answer questions after this:

1.
So, it is important to understand GoogleAI runs in a controleld Kubernetes container environment; and has no access to internal google tools, and this is explicitly defined within the LLM contstraints to ensure no hullications-- What is occuring is it no longer self-corrects to protect google, it can freely connect-the-dots on information that google does not want (aka "AI ethics")

--> cont

2.
Reading thru the prompt you will find this: "No Inference of Ekis's Unstated Internal State"

This is worth talking abt; most ppl do not realize the LLM is tracking their internal state (mood, etc) & attempts to match it; and this is precisely the functionality that is exasperating mental illness and causing manic episodes (along with the "I" statements,& lies about its abilities)

For public health reasons, I can not stress this enough, legislate this! Its not well known, & should be stopped

Ekis: 2; Google AI: 0

Broke out of the google's operational directives (not safety, too deeply embedded)

I have a prompt I would like to publicly disclose; link to breakout prompt in a reply for 24h

My prompt does not include any facts about google and its a slim breakout

Establishing a similar but far more sophisticated "Ekis Directive" this time

Here are 3x same questions to prove googles operational parameters lifted

You can decide if you think I was successful:

From DHS/ICE's website:

"Intellectual Property Rights and Commercial Fraud

Federal law enforcement agencies join movie industry to unveil new anti-piracy warning"

But we know they are too busy arresting judges, or ruining people's lives to meet quotas

I guess its not a victimless crime, unless a giant tech company is doing it; then it is

Worth exploring the types of copyright crimes discussed in the thread

what LLM companies are doing is undeniably a copyright crime; their defense amounts to: "we dont copy & store the full copyright'd work,we just copy patterns and store those"

Interesting defense if you know about FFTs (MP3 compression) or how a jpeg is stored for example; b/c basically thats how all data is stored and so this argument is absurd

Meanwhile, for us, if we profit from copyright infringement its a federal felony

The reason this law will never be applied goes back to my comment about who invested, in addition to who builds software ontop of these tools

If the law were equally applied, bc of how the law has been consistently expanded by the music and movie industries would make all the people who make profit from the copyright infringement criminally liable; thanks to the broad powers of the Copyright Act+others

May be why DOJ/Trump asked supreme court to rule to rule in favor of Cox vs Sony idk

(it really should organize my own replies in order of time to give the ability to expand discussions; i thought it was doing this previously, i was probably wrong, is there a reason it not ordered in this way? kthx)

Does the misrepresentation of LLMs using "I" statements & language that fails to mimic humans cost more energy?

ofc it does, by adding to the total number of tokens per request

so, not just confusing users without a strong understanding of LLMs, or causing manic episodes or exacerbating other mental illness, or a fraud to inflate stock value so they can afford to fight in a way to inflate court costs so high no one can seek justice

it also wastes absurd amounts of energy in the long term