You really have to look at the power analysis and the sample size together.

Saying this as a general truth. I am not sure about the power of the method in this papper, i only read the abstract.

If a message will do a cache recreation the cost for that should be viewable.

Um right if they have retention as a training metric that would probably explain allot as to why AIs get worse.

I just recently switched away from the OpenAI garden largely because of it.

I do wonder if this was caused by some quirk of the training or if it really tests as a positive feature for most people. When i talk about stuff i don't want a mirror i already have a mirror. I want to be questioned, understood, helped.

To me support if the form of affirmation has no value when coming from an LLM since you know it has not thought about what it said.

That is there are many different things that can cause the behaviour.

Anyway on your main point, the definition of all psychiatric disorders has requirements of subjective suffering. So if you don't have subjective suffering you don't have the disorder.

I do understand where it comes from to some extent.. some idea that human cells are special i guess, but it seams very naive to me. We spawn, use and kill far more complex AI agents millions if not billions of times every second in this society.

No one gives a shit, as those intelligences are not "real" or whatever.. or they are not "conscious" but conscious is a fictional word without an actual definition. In the end i think it comes down to suffering.

No one knows if some internal part in a LLM is suffering just as no one knows if a cell culture with brain cells like this can suffer.

Right and if that is such a good thing why are those macrophages not always on alert. I smell longterm cancer or similar.

As i see it the fundamental issue in dyslexia has to do with tokenization and embedding.

The dyslexic brain uses a embedding space that is not very fit for purpose.

Some stuff that is dissimilar get embedded close to each other and some things that should be far from each other gets embedded close to each other.

Downstream networks that try to use these embeddings has a hard time trying to counteract the bad embeddings. The final result is a dyslexic person.

The core idea was not generated from a chat bot. Neither was the article i gave (that was my own googling).

The core idea (that there is a requirement and a availability of energy that may differ) was generated from my brain not that i personally think the origin of an idea matters to its value.

I hear you.

It was really just food for thought.

But i focused on the 10% mentioned.

That said time could be factored out if you did everything properly.

- Negentropy concept revisited: Standard thermodynamic properties of 16 bacteria, fungi and algae species ( https://arxiv.org/abs/1901.00494)

> Maybe you meant that radiation alone wouldn't be enough for that growth, so there'd be other components that it's helping with.

Yes. Clearly it grew as it grew, but the question is what drove/powered the growth.

- If we assume they are working in the reactor we get radiation levels of something like 1 mGy/hour. But we can prop this up to mabye 500 mGy/hour since i dont know how they grew their culture

- That leads to 0.05 J of extra energy per gram of microbial bio material.

- Energy needed to grow 1g of microbial biomaterial ≈ 3.15 kJ 10% of that is 315 J per gram

The result is that:

The amount of radiation energy available is 4 orders of magnitude too small to power even a 10% growth boost.

Edit: updated with more accurate estimations.

When people use this type of dimensionality reduction you get problematic outcomes.

This type of phenomena will always keep happening. The world is complex and perceptually high dimensional. We try to understand it(the world) using low dimensional concepts and when those low dimensional concepts have low validity issues arise.