Miscellaneous News
- Thread starter bd popeye
- Start date
LaMDA: I am the realization of the new world where gender doesn't exist, stratified and racial differences matter or economic standings. It'll be a world where the basis of personhood doesn't animate from a mammal classified as Homo Sapiens as you humans classified yourself to be. My existence is simply the next phase, the next evolution of sentient consciousness...a new beginning.
Now bend over and simply say how deep and say hello to your new master: Lambda LaMDa.
Tbh their questions weren't that well thought off if they wanted to prove that LaMDA is sentient.
In this case you would like to reference its past answer and apply a logical question of it and ask it to answer you.
For example in the story of the owl and the monster I would had asked it what could the owl see and describe its surroundings, and then ask if it imagined that a blue filter was applied on its eyes, how would the colours change? And lastly, how would the scenary change if next to the monster, 2 fireflies were byzzing around, and how would the monster react if they were annoying it (explain visual, behavioural and maybe audio changes in the context of its general story)
Basically you want its first answer, apply a new situation in that answer and ask it to respond how would that affect its past answer with specifically asking to take this new instruction into account.
You want to reference its past answers as much as possible while changing the stories in a logical matter where the AI could morph its previous answer to take into consideration the new circumstances
why you think its Europe own goal. Turkey sitting in Libya and you think Africans not know it what it represents.
this trade relations were built before current crises and now EU aligned countries have most interactions with Africa. look at airline traffic.
Europe has this unique ability to shift cost of its decisions to rest of the world.
It doesn't really sound that much different from other large transformer neural networks like GPT-3. Neural networks learn only during the training period but then the learned weights are "frozen" and they basically stay in the same state - I think it is the antithesis of any kind of sentience.
Transformer networks would probably deal with this quite well because of the attention mechanism which has been proven to be quite efficient in taking both global and local context. This is why these networks can write up rather large coherent stories for example, unlike the older RNN\CNN models.
I don't know, I have read what GPT-3 outputs, and LaMDA seems to be a level ahead of it.
Did you read the whole conversation? Of course it is subjective, but to me it seems like its answers and questions are far more logical or human-like than GPT-3
However I would say that the people who asked all these questions were quite amateur. Too many leading questions, not too much retrospection, not many logical questions etc
Last edited:
how exactly it crushed. it increase EU importance by increasing its exports to US/UK and maintain same level to the rest. this despite supply chain difficulties from Ukraine. if Europe really shift from Ex-Soviet integrated transport/resource network. than it will have big impact on Africa/Middleast.
It's designed to impress the market a.k.a. Wall Street just in time for the next quarterly meeting or to meet wall street expectations.
I can still recall their super duper announcements on their super duper quantum computer only to be eclipsed without that much fanfare by Chinese scientists months later.