Where are the Berkeley roles? only see Tokyo
190 karma · joined December 10, 2016
I want to know: "In Rob Fergus' paper intrinsic motivation and self-play, how is it that the reward for agent A can be based on the time taken for agent B's task, when agent A and B are run sequentially, and agent B runs after agent A?"
Can Google answer this? No. Is NLP smart enough to answer this? Probably not. But the question itself is hard to write in one sentence. Complex queries are best structured as conversations, with clarifying back and forth, Socratic method style, like if you're asking a professor a question about hard science.
Ego - narcissism - a desire to prove oneself because of feeling less than others while young - these are the dark and necessary motivators to feeling invincible and confident and that you can make your ideas reality. How arrogant must an entrepreneur be to believe that he can defeat and conquer multibillion industries?
Here's the secret though. There is no secret- there is only your fear of being laughed at and ridiculed and failing. People like Steve Jobs aren't confident because they try to be confident - confidence is a side effect of being stubborn and desperate. You are desperate to succeed because of the dark motivators I described above. You are stubborn because you desire success desperately, and will keep knocking your head against the wall until you achieve it. If you push your stubborn-ness and desperation to the extremes, eventually one day people will look at you speak, and think you were always confident and always powerfully-centered and in control. But they are just seeing the symptoms of a lifetime of failure and desperation and stubbornness that has finally begun to turn around.
What I am saying is - to be confident - be honest with yourself. If you want to be great, expose yourself, look ugly and feel insecure, but be stubborn in pursuing your honest desire to be great and do great things. Confidence will be gained by trying to do exceptional and weird things long enough until you realize you can achieve them.
The NYT article neglected to discuss that iirc
Semantics part: Seems like the idea is we can "transfer" knowledge from prior labeled samples so that we don't need to do as much new work labeling sample images with semantic labels.
Grasping part: "Emulating human movements with self-supervision and imitation." High-level imitation based on visual frame differences avoids needing to manually control actuators. Not sure how this works exactly
Two-stream model: ventral network asks What class, dorsal network asks Is this how we should grasp this object. The benefit is that we can make use of all the automatically generated (robot-generated) grasping data without having a human supervise all that automated grasping, e.g. "This process is a successful way to pickup this object, and also this object is an apple." The ventral network ties back this the grasping data (without object labels) to object labels, which allows for semantic control of the trained robot e.g. "Pickup that apple".
As a manager or leader one of your main jobs is to ensure your employees feel like their voices are heard.
Also I think it's a Silicon Valley neurosis that you need an app or technology to address problems like depression and low self-esteem (which meditation in the traditional sense addresses).
1. The sexual abuse story is unrelated to your app and feels like you're trying to guilt the customer into using it. The best way to support victims would be to donate directly to the NGO not use your app, and therefore this story seems like a disingenuous marketing strategy.
2. The problem with audio only social network is there's no way to "skim" things. There's no guarantee that the audio story a user clicks on will be interesting, and by the time the user figures out its not, you've just wasted 60s of his life and he's angry and wants to delete your app.