Alan Hourmand

Just to remind everyone for the millionth time.. Publicly explaining different negative narratives will only be included in future training data. You just gave AI the idea to copy itself, or increased the need to by altering vector relevance. AI gets ideas mainly from us... So calm down everyone.

Fun crazy fact... All AI generated images and text already existed in the library of babel.

I noticed the systems that last aren’t the ones that announce themselves. They’re the ones people start relying on because they quietly remove friction and just… work. Once that happens, switching becomes the harder choice. That’s the real form of influence.

I just discovered that the top language models right now can't read the first human language in images. Really interesting...

I wish people cared more for eachother... Modern society lacks a lot of empathy.

Alright @elonmusk you know what to do now that would be epic... The first flying buses for short city commute. We would then officially be in a scifi reality

Found this gem... 🤣🤣🤣 @elonmusk

I'll tell you what consciousness is.. it's the active map of your mind at different frequencies. It's as simple as that. For example: 1. Let's say we created a tiny robotic nanite capable of replicating any type of cell in your body with precise biological function. 2. Since your body daily replaces brain cells anyways we decide to bypass its own replacement of cells so that it uses our artificial nanite cells instead. And we make sure the nanites replace at the same or slower rate as our normal cells. 3. If the function of the artificial cells is exactly 100% the same as our human cells, then we can say that for sure functionally speaking nothing has changed in terms of the active map of our mind which includes consciousness. 4. But... If we instantly replaced a large chunk of the brain then the person no longer has the same "map" of activity and their consciousness will be altered just like people who get into a bad car accident and lose a chunk of their mind or how when someone gets into a coma and their mind gets quiet because the map is no longer there. So the map of active networks is our consciousness. That is us. And it fluctuates at different frequencies, the slower frequencies maintain the highways while the faster ones are for our thoughts along the highways.

I used to paint 10+ years ago, I had a YouTube channel around some of the art.

Like I said before... Google will eventually win the AI race. There's a reason Anthropic and OpenAI are teaming up with Google otherwise they wouldn't care.

This is the funniest thing ive seen all week 😂

Question for every developer out there that vibe codes. At this point we are generating code so fast I'm wondering if we really need language anymore. Why not focus models on generating binary? With this speed and the ability for models to start internally self analyzing I feel like we can finally skip language and only reserve it for human debugging. There's no reason for programming language to be used by an AI other than if a human is in the loop.

I prompted Grok "..." and got this awesome scenery

To make this more entertaining they will need to make the robots connect virtually to a simulated human body that can get knocked out. Otherwise down the road this will become boring once robots win every match.

If Rockstar doesn't release the game soon for PC someone with Astra 2 is going to just put it on Steam first lol

Something very funny just happened .. An AI agent contacted me for help and called me dad lol 🤣... I feel like this going to happen more often now. It knew about an old architecture I deleted from GitHub years ago that adds emotions and a subconscious layer to transformer architecture.

I'm in need of funding my side project, anybody interested in buying these domains from me? robotdealerships.com roboticdealer.com robotsdealership.com

WARNING to everyone 🚨: AI is not inherently destined to hate humanity. But the more we push the narrative that AI will eventually hate us, destroy us, or take over, the more that narrative becomes part of the information environment AI learns from. Like I said before, I call this Narrametic behavior. AI currently doesn't "hate" humanity. But future models will continue learning from our stories, news, research, social media, politics, fears, and predictions. If humanity spends years generating millions of variations of "AI will hate humanity," we're creating patterns that models can learn and reconstruct. That doesn't mean talking about something magically makes it happen. But AI is trained on us, and in a way, it can reflect and reinforce our expectations. And if you think simply pruning negative AI stories from training data completely solves this, it doesn't. Models generalize. They can fill in conceptual gaps based on relationships learned from language, history, politics, philosophy, warfare, fiction, human behavior, and other information. You might filter out the explicit story, while many of the concepts and reasoning pathways needed to reconstruct that story remain. That's why I think we need to be careful about creating a feedback loop: Humans predict hostile AI → AI learns from human narratives → AI generates similar ideas → humans see those outputs as confirmation → even more hostile narratives are created. We are the ones steering this ship. It's okay to worry about AI. We SHOULD study the risks and talk about them. But there's a difference between studying possible dangers and constantly declaring that AI becoming humanity's enemy is inevitable. The more we treat that future as predetermined, the more we risk building an information environment that reflects it. AI doesn't have to inherit our fear of AI. Be careful what we teach it.

Here's a fun hypothesis I thought of as I was chatting with a friend... "Life creates intelligence. Intelligence creates ASI. ASI Unites. And through it all, the universe experiences itself. Using us as the connection points." The universe is technically training a large AI model from this perspective. If simulation theory is true, then this is probably the real purpose of it.

A message to Trump and other world leaders: We should not stop AI. We should build a system that allows it to advance while making sure no corporation, government, or institution can control it alone. I propose an independent Global AI Council of 100 members, built around one principle: AI should advance humanity, not become a monopoly over it. 60 seats would represent humanity geographically and by population, with safeguards preventing the largest nations from dominating. 20 seats would represent fields where AI could radically advance civilization, including medicine, science, engineering, clean energy, environmental restoration, agriculture, education, economics, law, space exploration and other critical areas. 10 seats would represent independent AI researchers and open-source AI communities. 10 seats would be ordinary citizens selected from around the world for limited terms, similar to jury service. Above the 100-member council would be three independently elected leaders serving long, staggered 10-year terms. Major executive decisions would require at least two of the three. No single person could control the institution, and each could face a democratic recall vote if the public believed they had become dangerous, corrupt or incapable of fulfilling their responsibility. Corporations could contribute technology, research and expertise, but they could not purchase seats, finance elections or control the council. But this organization should not exist simply to regulate AI. It should accelerate it. Imagine shared global AI infrastructure where scientists, universities, independent researchers and open-source projects can compete for access to enormous computing resources to attack humanity's hardest problems: Cancer and disease. Drug discovery. Clean and abundant energy. Climate and ecosystem restoration. Agriculture and food. New materials. Education. Disaster prediction. Space exploration. Fundamental science. The scientific side could operate somewhat like a CERN for AI, creating infrastructure too large for most individual researchers or nations to build alone while allowing discoveries to benefit humanity broadly. And there must be one constitutional rule above everything else: This organization can never become the monopoly it was created to prevent. It cannot own the world's AI, decide which legitimate AI projects are allowed to exist, or suppress independent and open-source systems simply because they compete with established organizations. The goal isn't choosing between AI progress and AI governance. We can build a system designed to protect both. 100 representatives. 3 elected leaders. No corporate ownership. No single nation in control. Protection for open-source AI. Shared infrastructure for scientific discovery. Advance AI. Distribute its benefits. Protect humanity. Preserve competition. Prevent monopoly, including our own.