If data centres were forced to contribute what was required to build the utility infrastructure required to support them, that might be true. For example, electricity costs.
Datacentres are generally built where poor people live. If you are on minimum wage and suddenly someone increases your electricity bill by $300 because of data centres, on top of all the other increases tied to things like energy insecurity... I can understand why people would consider that an impact.
Corporate power is now so great that they successfully lobby to reduce/eliminate their tax burden so they don't pay for infrastructure they need.
Workers pay that cost through increased market prices and taxes (eg energy) which also drives up other prices. It's not inflationary because it's a real value adjustment but everyone can pretend it is while real term wages drop even further.
But people also seem confused as to what actually creates those systems, which leads to the perceptual misattribution of confusing the reflection of an object in the mirror for the object itself.
If you can disprove the possibility of a disembodied brain being provided the same signals as a body does, then I can agree with this, otherwise it is unproven. It isn't necessary that senses come from sense organs that look like us as long as wec can feed the same chemical and electrical signals to the brains (eg how some drugs create sensation of sights). In other words, disprove neuralink's concept.
A brain isn't a brain because of it's specific signal inputs, a brain is a brain because of the selection pressures that form it.
If you want an artificial brain you could a) create a synthetic copy of a mapped biological brain (eg like the fruit fly neuron map) or b) evolve a brain using selective pressure (virtually)
Simply calling an LLM a brain, doesn't make it a brain. It's a prediction model.
One illustration of this fact is how an LLM (that counts three e's in "seventeen") can also tell you why it has trouble counting letters in a word correctly and give you the example of 'seventeen' - pointing out to you that there are actually 4 e's but that it will tell you there are three 3 e's. You can then immediately ask it again how many e's and it will still tell you 3.
An LLM doesn't reason, it doesn't understand, it doesn't communicate, it isn't self aware. It predicts the next token (astonishingly well, in most cases).
reply