Korr Neuro

Explore a new approach to true machine intelligence. KorrTecx - a biomimetic simulation of neural structures and mechanisms, evolving aligned intelligence, multimodal continual learning & curious imagination. Pioneering cutting-edge technology for real-world challenges.

IQ is technically not a measure of raw intelligence. It's a metric of how well an intelligence has adapted & learned our communally/ societally agreed perception/ understanding of human centric, 'actual reality'. If a space faring race landed, they would fail a human IQ test.

@AIandDesign Haha... yeah! 👍

I wish everyone could just get along.

@WilliamShatner Take my money!

🤔 Scotch flavored Gummy Bears.😑

@dangreenheck Quality costs.

Some people ask me why I charge so much for my assets when AI can make a half-decent water shader these days. This example is a good demonstration of where your money goes, and why AI isn't going to make premium-level assets anytime soon. AI is phenomenal at solving focused, technical problems. Better than me for sure. But it's not as good at broad, ambiguous problems. In the video, I'm submerging a test grid of semi-transparent squares underwater using different blending methods. I need to make sure each test case renders properly above water, below water and at the waterline. Then making sure that works when the camera is above water, below water, and at the waterline. That is a total of 72 test cases. Each square needs to composite correctly with the water fog, the water color, the clouds, the sky, and any other objects in the scene. And all of this needs to be done in an elegant way so the end-user doesn't have to do crazy hacks to get it working on their end. It should "just work". You can see that while the squares are underwater, some of them are allowing the clouds to come through due to a bug in the scene depth logic. I'm working on a fix for that right now. Sure, I could write some automation for this. I've actually tried it. But Claude often misclassifies results, comes up with a non-comprehensive solution, or introduces some very subtle regressions. It's cheaper and faster for me to manually test this myself, and I can reframe the problem into a very narrow scope so Claude can go investigate it, rather than spending hours trying to perfect a flaky automated test suite. I've honestly stopped worrying about AI taking my job because it's coming clear that it exceeds my capabilities in many areas of software engineering, it's still missing out on the one thing you can't program: taste.

@mkadnk @Hitchslap1 So true for me. 🤣

@Hitchslap1 Most men never buy their own wallets. They just hold onto them until someone close to them has pitty on their wallet and replaces it.

Men rarely upgrade things. They just wait till they break or are completely destroyed. Men’s wallets illustrate this perfectly.

@ebarenholtz No, language evolved to move ideas (memories) between similar but totally unique minds. It's a communally agreed protocol & although it does hold the embedded essence of constructs, it has no inherent intelligence.

Does reasoning happen in language? This new PNAS paper argues no: the language network stays quiet during logical reasoning, and patients with profound aphasia reason fine. The truth is we've known for a while that many people report little or no inner monologue and seem to reason perfectly well without one. So is language secondary, a mouthpiece bolted onto a deeper 'symbolic' thought system that does the real work? LLMs point to a different possibility: a difference between a word and the modal expression of a word. In a transformer, a word is a token (really, just a vector), an amodal index. The computation runs over these tokens end to end; nothing in it is ever sound or image. Modality appears only if a separate system renders a token outward. Perhaps the same thing is happening in us: a kind of codec that maps amodal word-tokens into a modality. It can be vocalized speech, written text, or the hand gestures of sign language. The fact that language is convertible and learnable in these other forms is a clue. In this view, aphasia destroys the codec, the machinery that turns heard speech into words and words back into speech. Sever it and you lose expressed language, while the generative process upstream keeps running over amodal tokens. The patient reasons; the surface goes dark. The inner monologue is the same codec pointed inward. Switch the rendering off and the words still run, unheard. Note: this does NOT mean reasoning runs in latent space, translated into words only for output. It may still be serial, autoregressive, next-token generation. But the "words" are amodal, sound being one way to render them out. So the dissociation is real and well demonstrated, but it may be the wrong interpretation. It separates externalized language from thought. It says nothing about the format upstream, which may be linguistic through and through. Note: I am not claiming reasoning only happens in language. Non-linguistic animals reason too, in visual and spatial modalities, and so do we. My point is that we likely do reason in (amodal) language, among other ways. But maybe the same split runs through those modalities as well, an amodal generative process upstream and a rendering downstream. That would put a new spin on aphantasia, the mind's eye switched off while the reasoning underneath runs on.

@DustinJuliano Agree, current AI is resource hungry & will never achieve AGI status... AGI requires a paradigm shift in approach.

One of the most important positions I hold on AI, as contrarian as it may be, is that we are vastly overestimating the resources required to train and operate AI in the future, especially AGI.

@WilliamShatner Is that your new Gorn whacker. 🤣

I am going to be doing a live performance of my new live heavy metal spectacle to Douglas Park in Chicago on Sunday August 20, 2026. This is an event you definitely will be telling your grandchildren about! 😉. I hope to see you there. More info at Riotfest.org

@MillerLabMIT Nested harmonics, consciousness is the ∑, the froth on the top.

Cortical processing looks categorical at large-scale, but at the local level, computational strategy favors diversity and flexibility over rigid categories. Rarely categorical, highly separable representations along the cortical hierarchy nature.com/articles/s4158… #neuroscience

@Hitchslap1 But... Never mind. 🤣

Never try to reason with someone more than 30 IQ points below you. They’re working with a completely different toolkit. Think about it.

@andywergedal Succinct list. 👍

Five things I’ve found using AI tools for the last five years 1. AI cannot come up with a new original idea. 2. AI is exceptional at absorbing and being able to manage and process large data sets. 3. Even though AI can organize things, it’ll only organize things the way it thinks not the way you think. 4. Notetaking in meeting minutes should never use a human unless it’s private Confidential!! and you never want any AI system to have what you said or processed electronically. 5. AI is fantastic and analysis and summary and bad at decision-making.

@parmita @VoidStateKate Google search results... LUMA = Currency: It is a monetary subunit in Armenia, worth \(\frac{1}{100}\) of an Armenian dram. Do you see the problem yet? 🤣

@VoidStateKate Any advice?

THIS HAS 120 LIKES AND IT DESERVED 120000 😡

@parmita @VoidStateKate Google search results... MBOP refers to a specialized platform rolled out by Marriott International for its branded properties Marriott Branded Hotel.

@Hitchslap1 It's an ancestor component.

Is there a genetic component to being Japanese or is everyone and anyone born in Japan equally Japanese? Think about it.

@Kekius_Sage When space ends.

When does time itself come to an end?

@ChShersh Yup! me too, my solution is to build a new industry.

I’m overqualified for what I do but underqualified for what I want to do

@andywergedal CEO. 🤣

Every company has one thing everyone knows is broken but no one will name out loud.

@keithfrankish Some. 🤣

People are far more complex than the images of them we have composed from things they have said.

@DJSnM Rudder into wind, full power... VTOL. 🤣

You're on the ground in a small piston plane, engine running, and suddenly there's a 55 knot gust due to a microburst... how do you react? youtube.com/watch?v=b_WmjW…

DMN - default & executive mode networks are not distinct, they're a continuum within the manifold. Vid shows KorrTecx self-sustaining DMN startup through visual stimulation, notice V1 activity rocking forward into frontal, pumped by brain stem creating the base thalamic rhythm.