4 ms·
Because the algorithms are currently “just” transforms based on probability. There’s no understanding present in current ML models.
by mrcartmeneses 4y ago
Because the algorithms are currently “just” transforms based on probability. There’s no understanding present in current ML models.
- OrangeMonkey 4y agoQuestion - how do you think consciousness develops? See, I don't know. So, giving an algorithm the ability to store facts, determine weights, and make decisions that they can see outcomes of... who is to say thats not how consciousness spontaneously develops. Smarter men than me perhaps. But, again, smarter men than me also think we should nuke the internet :)
- IceMetalPunk 4y agoDefine "understanding" in such a way that predicting what comes next based on experience of how things have happened in the past is not included. In other words, define "understanding" in such a way that humans have it, but these ML models (especially LLMs) don't.