Wiki Contributions


Don't know much about Punk Rock, but there is Transhumanist Metal.

See Meshuggah - Future Breed Machine, or Rational Gaze again by Meshuggah.

  1. It is a different net for each game. That is why they compare with DQN, not Agent57.
  2. To train an Atari agent for 100k steps, it only needs 4 GPUs to train 7 hours.
  3. The entire architecture is described in the Appendix A.1 Models and Hyper-parameters.
  4. Yes.
  5. This algorithm is more sample-efficient than humans, so it learned a specific game faster than a human could. This is definitely a huge breakthrough.

I myself did some of Andrew Ng's courses, and I understand where you're coming from. Although this was several years ago, but I do remember Octave!

I saw your guide: I just want to say: Real Analysis is overkill in my opinion if your goal is to simply become an AI researcher. Also, I personally like Karpathy's advice (which seems like it should radically alter your guide):

How to become expert at thing: 
1 iteratively take on concrete projects and accomplish them depth wise, learning “on demand” (ie don’t learn bottom up breadth wise) 
2 teach/summarize everything you learn in your own words 
3 only compare yourself to younger you, never to others

Also, just as a side note, even mathematicians do this in research if not for basic things like real analysis:

Most big number theory results are apparently 50-100 page papers where deeply understanding them is ~as hard as a semester-long course. Because of this, ~nobody has time to understand all the results they use—instead they "black-box" many of them without deeply understanding.


That would be interesting to try, I sent you a private message with my contact details. 

Just one thing I want to point out is that the problem is finding the right buddy/collaborator rather than any buddy/collaborator. It seems like there are a lot of platforms like Discord or IRL friends etc. that would do the trick if you just wanted any random buddy.

For example: I did try this with one IRL friend, and that quickly failed. It worked for a day or two but then it just felt like I was spamming him with updates, and the whole interaction felt like it was not adding value to him. Basically the problem was that he was not equally invested in any similar project, so I wasn't getting any updates from him.

It is interesting to note the limitations of the system. From the paper:

6.4.5| Failed Hand-authored Tasks

Gap tasks Similar to the task in Figure 21, in this task there is an unreachable object which the agent is tasked with being near. The object is unreachable due to the existence of a chasm between the agent and object, with no escape route (once agent falls in the chasm, it is stuck). This task requires the agent to build a ramp to navigate over to reach the object. It is worth noting that during training no such inescapable regions exist. Our agents fall into the chasm, and as a result get trapped. It suggests that agents assume that they cannot get trapped. 

Multiple ramp-building tasks Whilst some tasks do show successful ramp building (Figure 21), some hand-authored tasks require multiple ramps to be built to navigate up multiple floors which are inaccessible. In these tasks the agent fails. 

Following task One hand-authored task is designed such that the co-player’s goal is to be near the agent, whilst the agent’s goal is to place the opponent on a specific floor. This is very similar to the test tasks that are impossible even for a human, however in this task the co-player policy acts in a way which follows the agent’s player. The agent fails to lead the co-player to the target floor, lacking the theory-of-mind to manipulate the co-player’s movements. Since an agent does not perceive the goal of the co-player, the only way to succeed in this task would be to experiment with the co-player’s behaviour, which our agent does not do.

GPT-1: *sentiment neuron*

Skeptics : Cute

GPT-2: *writes poems* 

Skeptics: Meh 


GPT-3: *writes code for a simple but functioning app* 

Skeptics: Gimmick. 


GPT-4: *proves simple but novel math theorems* 

Skeptics: Interesting but not useful. 


GPT-5: *creates GPT-6* 

Skeptics: Wait! What? 


GPT-6: *FOOM* 

Skeptics: *dead*

-- XiXiDu (I added a reaction to GPT1).