1,329 karma · joined October 9, 2014
Github: @pdelfino
Personal website in English: https://www.pedrodelfino.com
Personal website in Portuguese: https://www.pdelfino.com.br
For instance, try picking up an outdoor sport which is new to you, e.g. kite-surfing, snowboarding or climbing. It will fulfill the three requirements.
"E boa sorte nessa aventura, meu amigo!"
Moreover, I was expecting to see "Paradigms of Artificial Intelligence Programming", from Peter Norvig. I have not fully read it, but it seems to be an awesome book. Not just to learn Lisp but a good book in general, since the code is beautifully designed. Despite the title, AI (GOFAI, actually) is just the domain, the lisp techniques and code approach can be applied to other areas.
This Reddit's post presents another list that might be an interesting complement to the one we are discussing here: https://www.reddit.com/r/Common_Lisp/comments/ddcoar/selecti...
"Take these two examples of desired universal laws of software engineering:
FU: every system, big and small, should have provisions that allow it to be FIXED and UPDATED either from within or outside the system proper;
CK: every system, big and small, should have provisions that allow it to acquire and make use of CONTEXTUAL KNOWLEDGEto decide whether it is functioning properly or not.
The combination of the two laws gives us the FUCK property: A system that obeys both the FU and CK laws can verify the conditions of its operating environment and decide whether its actions are coherent or not with the current context of that environment.
Furthermore, if it decides they are not, the system can fix itself, ask for an external fix or be fixed, forcibly, by some sort of external agent.Think cells and cellular systems. They are obviously a FUCK system"
This book is tough. Really tough. However, you are really going to do "deliberate practice" while having a go on the assignments.
It bothers me that the author does not provide the answer to most of the questions. This is specially bad in this field of Mathematics (Probability). As an example, in calculus you can always plot the curve, the derivative and see if the result makes sense, in probability theory it is hard to have a simple and safe sanity check.
You can watch all classes from Harvard here for free: https://www.youtube.com/watch?v=KbB0FjPg0mw&list=PL2SOU6wwxB...
If you want, there is also a MOOC in EdX with the same material: https://www.edx.org/course/introduction-to-probability-0
Sheldon Ross' book is also a good one. It was mentioned here. I specially like the section on theoritical problems, only with exercises on proofs. Every chapter has one. However, Joe Blitzstein problems are more challenging and will train you more on intuition.
Mindset
Peak
Part 1 - https://www.youtube.com/watch?v=T647CGsuOVU
Part 2 - https://www.youtube.com/watch?v=2HrSG0fdxLY
Part 3 - https://www.youtube.com/watch?v=N9QOLrfcKNc
Part 4 - https://www.youtube.com/watch?v=DThAoT3q2V4
Part 5 - https://www.youtube.com/watch?v=65wYmy8Pf-Y
Part 6 - https://www.youtube.com/watch?v=z5IG_6_zPDo
Part 7 - https://www.youtube.com/watch?v=YHvR8siIiD0
Part 8 - https://www.youtube.com/watch?v=iecUL8_OxrU
Part 9 - https://www.youtube.com/watch?v=dLn5H69lS0w
Part 10 - https://www.youtube.com/watch?v=pNp8Qf20-sA
Part 11 - https://www.youtube.com/watch?v=0hiWbdc8QEk
Eli Bendersky http://eli.thegreenplace.net/
FreeCodeCamp/Quincy Larson on Medium