I am developing small-mid size AI models to train them to save on resources and energy, the big LLM a highly expensive and not energy efficenet. Great video, thank you for sharing.
The idea of using AI to filter AI-generated content is fascinating. But how can we ensure that the AI filter itself isn't biased or inaccurate? then there is the issue of 'hallucinations' in AI models, and where do we draw the line between filtering and censoring?
The car will act as a human with Empathy would. It would first try to stop. If midstream it determines it cannot stop, it will veer off as safely as it can determine, perhaps hitting the wall slower, and thus causing no casualties.
The thumbnail didn't jump out at me. I'm watching because after scrolling past a couple of times, I noticed it was you. It just has a look that is slightly off brand, but I don't know why. Just thought I'd share. Anyway, watching now and glad for the video.
The car needs to be taught probability. If the car hits the pedestrian, the pedestrian will be more likely to die compared to if the car veers into the wall, the passenger has a higher probability of surviving that impact provided they wear their seatbelt.
It should be to let the decision to the driver. And if they chose for the pedestrian to die, they ll go to jail. Like what happens anyways. Passemger will hv skin in the game and be more attentive.
I searched for this question last month without an answer and now is in my subscriptions feed
Nice! I hope I was able to answer it sufficiently!
@@Hallden_ somewhat, I would want a more in depth explanation but you oriented me with the most important information so is good
I recommend a paper call "attention is all you need"
❤ I needed this video. Thanks!
Really good video, bro! I missed you on RUclips. Keep up, you're top tier tech content creator, honestly.
car cannot hurt the passenger otherwise no one boughts the car. To sell the car, you need to protect passenger first.
Thx never heard 👍
2:43 it is used the Markov chain technique to make a connection between those URLs based on relevance
I am developing small-mid size AI models to train them to save on resources and energy, the big LLM a highly expensive and not energy efficenet. Great video, thank you for sharing.
this is pretty cool , I wonder what LLM they use for filtering , would be ironic if they used one of the gpt models to build their model 🤣
Thanks for these informations
The idea of using AI to filter AI-generated content is fascinating. But how can we ensure that the AI filter itself isn't biased or inaccurate? then there is the issue of 'hallucinations' in AI models, and where do we draw the line between filtering and censoring?
i think the car must stop to save both, because in the end of the day they are human
This was very cool man , these topics are always quite highly technical and often tough understand but this was very clear explanations 💯
Thank you! That means a lot!
The car will act as a human with Empathy would. It would first try to stop. If midstream it determines it cannot stop, it will veer off as safely as it can determine, perhaps hitting the wall slower, and thus causing no casualties.
Ofc it will try to stop. But The dilemma is it dont have time to stop.
The thumbnail didn't jump out at me. I'm watching because after scrolling past a couple of times, I noticed it was you. It just has a look that is slightly off brand, but I don't know why. Just thought I'd share. Anyway, watching now and glad for the video.
u deserve more than 6k views
make a huge big legal Webcrawler Projekt with Data Science .. lets make some knowledge out of data bro
The car needs to be taught probability. If the car hits the pedestrian, the pedestrian will be more likely to die compared to if the car veers into the wall, the passenger has a higher probability of surviving that impact provided they wear their seatbelt.
It should be to let the decision to the driver. And if they chose for the pedestrian to die, they ll go to jail. Like what happens anyways. Passemger will hv skin in the game and be more attentive.
1st like, first comment