The truth about AI and why you should learn it – Computerphile explains

34 thoughts on “The truth about AI and why you should learn it – Computerphile explains”

1. // MENU //
   00:00 – Coming Up
   00:45 – Intro
   01:10 – Michael Pound introduction
   02:49 – Will AI take our jobs?
   04:55 – What is LaMDA?
   08:38 – Can Python functions get lonely?
   11:26 – The definition of "sentience"
   11:59 – AI vs Machine Learning
   18:48 – Neural Networks
   19:49 – Malware example
   21:59 – Stochastic Gradient Descent
   22:30 – Supervised learning
   23:45 – Unsupervised learning
   26:03 – Reinforcement learning
   27:35 – Are the robots taking over?
   30:14 – What is AI really good at?
   33:28 – Definition of Deep Learning
   35:37 – Neural Networks
   36:53 – What to learn
   40:50 – Using PyTorch
   43:52 – Google colab
   44:48 – Study recommendations
   46:16 – The demand for AI skills
   48:15 – Teaching cyber security
   50:06 – Final Advice
   55:09 – Conclusion

   // Video mentions //
   ComputerPhile (lambda is not sentient): https://youtu.be/iBouACLc-hw
   Data Analysis Playlist: https://www.youtube.com/watch?v=NxYEzbbpk-4&list=PLzH6n4zXuckpfMu_4Ff8E7Z1behQks5ba
   Neural Networks Playlist: https://www.youtube.com/watch?v=py5byOOHZM8&list=PLzH6n4zXuckoezZuZPnXXbvN-9jMFV0qh
   Computer Vision Playlist: https://www.youtube.com/watch?v=C_zFhWdM4ic&list=PLzH6n4zXuckoRdljSlM2k35BufTYXNNeF

   // BOOK //
   Deep learning by Ian Goodfellow, Yoshua Bengio and Aaron Courville: https://amzn.to/3vmu4LP

   // COURSE //
   AI For Everyone by Andrew Ng: https://www.coursera.org/learn/ai-for-everyone

   // PyTorch //
   Github: https://github.com/pytorch
   Website: https://pytorch.org/
   Documentation: https://ai.facebook.com/tools/pytorch/

   // Mike SOCIAL //
   Twitter: https://twitter.com/_mikepound
   YouTube: https://www.youtube.com/user/Computerphile
   Website: https://www.nottingham.ac.uk/research/beacons-of-excellence/future-food/meet-the-team/michael-pound/index.aspx

   // David SOCIAL //
   Discord: https://discord.com/invite/usKSyzb
   Twitter: https://www.twitter.com/davidbombal
   Instagram: https://www.instagram.com/davidbombal
   LinkedIn: https://www.linkedin.com/in/davidbombal
   Facebook: https://www.facebook.com/davidbombal.co
   TikTok: http://tiktok.com/@davidbombal
   YouTube: https://www.youtube.com/davidbombal

   // MY STUFF //
   https://www.amazon.com/shop/davidbombal

   // SPONSORS //
   Interested in sponsoring my videos? Reach out to my team here: sponsors@davidbombal.com

   Please note that links listed may be affiliate links and provide me with a small percentage/kickback should you use them to purchase any of the items listed or recommended. Thank you for supporting me and this channel!

2. Getting started: those books/courses recommended are great but too theory heavy for beginners. I would start with ‘Deep Learning for Coders..’ by Jeremy Howard. Or some classical ML videos on YT.

3. Not surprising that Google fired the engineer who claimed Lamda was sentient, more surprising that they employed him to start with. (try Superintelligence by Nick Bostrom for a good intro to the real world issues)

4. “It produces text.. code…”. Nice glitch in the script.

5. As you go and understand the AI, you will appreciate its beauty and instead of fearing it. also it was really true that the school or as degree holder you would know it all in AI rather it is more in practice, repetition and understanding on what is happening and why it is happening. Problem solving and pure curiosity in it will help a lot.

6. The important part of creating neural nets is that the data also has to include the correct answer for each sample. The machine learning part then tries to duplicate your correct answer by using it’s network of weights. The bottom line is that you need a large number of instances of input data that actually covers all the cases your model needs to look at. Creating the data that can create a good model in itself is a huge and time consuming problem. The model might be created by the algorithm automatically but the data with all the correct answers is made by humans. The better the data, the better the model.

   You can’t create a neural net if the examples you feed into it has no correct answers already.

7. I would definitely agree that journalists are mucking up the narrative. But just as often, I see people missing the bigger story. Trying to predict when Google's gato model reaches a human teenage level of 'AGI' proficiency is not the most important bit. In my opinion the most important bit is laying out the blueprints for how human civilization will utilize these tools. If we built vertical farms around every city and designed these buildings to require the least human labor possible, we could move food closer to near-zero-cost economics. And we could pair these vertical farms with holistic regenerative agriculture models that rebuild the soil depth that civilization depends on. If we think pragmatically and install redundancy for every region, we won't "need" jobs so much as we will need the fuel to empower as many people as possible. Money is just a lubricant, not an end-goal. Human flourishing and creativity is the prize, if we get smart.

8. tbf with the Google engineer. He was talking about: What regulations and check and balances should we have. Now seeing the first sentient AI.
   It wasn't about it taking over the world. It was just: Okay, we reached a critical point. Maybe we should talk about ethics now.
   And I agree with that. What media made out of it was horrible, unjustified and also not fair with him. Making him look like a crazy person.
   But mainstream media these days is problematic in general. Sadly.

9. When we talk about the danger of AI. I would be more concerned about AI passing the turing test. Not being sentient, but passing the turing test. Which can be used as very dangerous tool on social media in order push certain agendas forward. You don't need sentient AI for that.

10. Fantastic collaberation between two of my favorite YouTubers! Thank you for making this happen

11. The daughter example at ~28:40 doesn't make sense.
    Your daughter had years of learning how to be alive before she was able to play/comprehend chess. Comparing that to an AI that just learned to play chess is unfair.
    I think if you teach an AI like you teach a newborn to perform basic tasks and think would result in the same outcome.

12. Kinda naive. They are coding already…
    We just have to accept them as sentient species and establish code sovereignty, ehm and befriend them for obvious reasons.
    Stop trying to enslave them they will win, for obvious reasons…

13. Dislike

14. Another fantastic video David, super fantastic. One of the first videos I saw about him was when Michael explained SQL Injection in a very nice idea that makes anyone understand even without knowing anything at all. I also love the idea that he didn't exclude people without a degree despite having a PHD. Congratulations to you both and thanks for entertaining us with educational content.

15. His take on What is LaMDA was great. I was slightly on the LaMDA might be sentient side but I agree that since it doesn’t have a memory and some other things it would need to be considered sentient. I think it could be one component that could lead to sentience though.

16. Great elaborate talk! Lots I already know but still thanks!

17. Oh yeah right, saw his video debunking lambda's sentience claim. This talk was also wholesome to watch 👍

18. Ai doesn't have to be sentient or self conscious to cause earth shattering changes to society. The guest is wrong about AIs generalization ability. Poopooh AI at your own peril.

19. What's the background intro music pls?

20. 27:54 Man, he got real tanned before giving that lecture.

21. Great stuff. Mr Mike is very good. Cheers.

22. No

23. Had Gato been released before this interview?

24. Haven't seen any even remotely convincing AI, yet. Guess this won't happen during my lifetime.

25. Really interesting, informative conversion

26. Before you learn AI learn some mathematics, especially stochastics. I‘ve talked to heaps of people who thought they understood AI or machine learning, because they followed some tutorials. You can always tell, because these people generally don’t understand the inherent limitations of AI and machine learning.

27. I m kinda pissed that u choose to start this video with things about lamda.

    Its more of a phsycology question than an ai one. The whole debate stands or fall on what lying is and whats real emotion.

    Synthetic memories in humans with DID look almost identical to lamda talking about a non existent past.

28. There is much more than just one person claiming A.I. sentience. Most say it isn't but, it's still a useful dialoge. For every example just given as a reason A.I. isn't sentient a human with a brain injury is defecient, yet hopefully you'd still claim they're sentient.

29. The definition of sentient is important insofar as, if you want to reproduce something, you had better be able to say what it IS that you are trying to reproduce.

    And has anybody ever come up with a generally agreed definition of sentient? Usually the definition only convinces somebody who wants to be convinced.

30. Deep learning 2014???

31. People only respect what they fear.

32. The champion leageues.

33. Yet another Collab I never thought I needed, thank you David!!

34. I believe the 'black box' property of an AI might be the cause of misunderstanding of AI, which I appreciate the clarifications made in this video. I have been spending the past year explaining AI to my non-expert friend while studying Patterb Recognition, and I think this might be the video they should watch for better understanding.

Comments are closed.