How does

> 7) associative learning that goes beyond habituation and sensitisation

square with "evolutionarily habituated survival instincts"?

What definition would you use instead?

Under that definition, a language learning model is sentient.

I mean, there is an entire discussion ongoing whether or not they might be.

But LLMs (today) don't fulfill criterion #7: After training is completed, the model is fixed, so there is no more learning going on. #6 would also be hard to map to an LLM.

That is inaccurate given that the modality for improving a model clearly fulfills criterion #7.

The legal draft did not in any way either specify an AUTONOMOUS process, nor did it even specify any type of localized "learning" in terms of memory aggregation.

A discerning reader might notice that this begins to wade into very challenging philosophical questions.

Ding, ding, ding! You figured it out!

> that all life forms have.

Presumably including us. So are we also not sentient?

If a dog takes a beating, it will remember that, will be affected by it, will suffer for it, long after the physical event has passed. Just like us. The question is whether a crab would remember getting a beating. That is the difference between pseudo-mechanical stimulus-response and learned behavior, and a being that is ‘sentient’.

What's the difference between the two?

It's not my job to answer that question, it's the job of hacks/frauds codifying it into a legal ruling.

Oh no, a slippery slope of compassion! To wit, the long title of the bill:

"An Act to make provision for an Animal Sentience Committee with functions relating to the effect of government policy on the welfare of animals as sentient beings."

Am I the only one who doesn't see a problem?

As long as you make a claim which depends on that answer for support, it absolutely is your job, and whatever axe you mean to grind here doesn't seem so far to be helping do that job.

To reject a proposal as clearly fraudulent, an alternate proposal need not be furnished.

No, but you do need to substantiate the claim of fraud, and thus far you've yet to try. I doubt you will, but perhaps you'll surprise me.

@boronguser2, I vouched for your now-dead sibling comment of 29 minutes ago, because I think you're on to something and I believe your argument should have voice in this discussion. However, even though I sympathize with your position, and even though I don't think your comment should be dead, it's very fighty and condescending. Even if you aren't technically violating any of the rules that I can see, the condescension gets very close to being a personal attack. I think you'd be more successful in voicing your argument if you toned it down a little (or a lot).

Leaving the unjustifiable condescension aside, the argument in the dead comment still founders on precisely the lack of any philosophically rigorous definition of sentience on which it attempts to support itself.

That's the question I called out earlier as being begged: in order to discard a definition of sentience (or of anything!) as insufficient, one must be able to show where it fails. That would admittedly be an impossibly tall order in this case, but my sympathy for the difficulty is curtailed by it being one our interlocutor has chosen to take upon himself, and attempt to bluster his way past anyone else noticing.

Even politely expressed, such an attempt to demand an impossible standard of others while excusing oneself from the same merits exactly the obloquy it's received.

Your impression of how philosophically rigorous conclusions are reached is completely unnecessary to address; it is simply wrong.

You don't seem to understand that a rationalistic approach isn't going to lead you to conclusions that approximate accuracy about the world. See: philosophy that is over 100 years old as a reference, when people realized that this method of inquiry made no sense.

That being said, brushing aside your basic philosophical misunderstanding, it still follows that the burden of this challenge lies on the people who have outrageously claimed that they have uniquely solved it.

I make no such outlandish claims.

Psychological trauma is the difference. Lingering negative impact on wellbeing, things like that.

As silly as begging the question of how sentience is accurately defined, such that this definition can be trivially dismissed as inaccurate?

Perhaps all lifeforms have sentience?

I find this highly doubtful, but it also undeines just how challenging this domain is.

You can disagree with the findings, but why do you think attempting to understand the difference between sentience and evolutionarily habituated survival instincts is "silly"?

Isn't it our job as a sentient, evolved species, to ensure we do no harm to other sentient species?

And the only way we can begin to consider doing that is if we figure out a way to classify them?