Hearted Youtube comments on george hotz archive (@geohotarchive) channel.

  1. Shoutout to the guy who makes the Twitch chat subtitles
    649
  2. This should get an award. Useful thinking patterns are what we need to teach everyone.
    432
  3. Finally. That’s what we’re waiting for ❤️
    280
  4. Love this format he should do this more often
    270
  5. 10 hours streaming tech. No wonder this guy is successful in what he does, he puts in the hours and is relentless
    260
  6. When I start a Goerge Hotz video, I'm always thinking I'm not going to watch the whole video, but I always end up watching the whole video.
    243
  7. "Babe wake up new geohotz stream dropped"
    218
  8. 1:16 geo listens to stars - your ex lover is dead (mute) 5:30 un-mute 6:05 geo explains what he's going to be doing 6:23 geo talks about last stream 7:30 who does geo respect 10:02 goes through non contrutive proof 13:16 hater in the twitch chat about george fooling us 15:00 geo finds a bug in his budy beaver code 18:21 explains about schems 18:30 geo talks about why his vim setup doesn't matter 19:50 geo talks about the best way to teach/ learn things 23:21 geo starts working 29:59 geo finds write yourslef a scheme in 48 hours wiki book 31:41 what is monad? 35:22 timeout more people - we need standards on the internet 41:15 parsec haskell 42:54 is learning haskell useful? 45:26 insight into how to make sure you are learning 50:33 let's do a float 1:05:19 what geo is going to do about dumbness in twitch chat 1:08:03 single-qute syntactic sugar of scheme 1:08:20 structure and interpretation of computer programming book 1:10:53 on pervious stream geo on web devlopers 1:12:35 ohh no my float's broken 1:13:00 geo fixes the float - we just gotta try 1:15:17 which exercise do we want to do? 1:15:51 what are vectors 1:16:00 this is all too much not doing any of these 1:17:26 geo's story about motivation 1:18:30 evaluator 1:25:26 things are getting exicting watch this bois (impleating the plus) 1:33:00 testing if plus works 1:35:22 keyboard question after long time 1:36:47 learning about symbol handling function 1:40:26 gifting subs 1:41:08 learning about PVM 1:43:35 looking at the diffrent PVM 1:49:16 can you make GPT-3 for videos? 1:54:00 GPT-3 and turing test 2:14:37 adding variables and assignment 2:15:10 disabling channel points twitch 2:20:01 the problem with internet communities 2:28:18 what is a good question and a bad question 2:38:32 vitalik homomorphic encryption - talks about few other stuff after this 2:44:36 george gives us Home work for next stream + test
    200
  9. Thanks for archiving these streams. I play these in the background and that keeps me going in my studies or side projects. Geo's work ethic and optimism is infectious
    193
  10. I swear to god if I have to wait another 3 months
    187
  11. Whoever you are, thank you for making this archive channel for his VODs.
    182
  12. George: "I hate you guys and I hate streaming" Also George: "So, I'm thinking of streaming on the weekend" Never change George, never change.
    173
  13. 00:07 George Hotz talks about engagement with AMD and lack of substance in communication. 03:14 Understanding the dynamics of dealing with big companies like Qualcomm and Nvidia 08:12 Tech industry leaders' videos reflect their personalities. 11:00 Tiny grad not expected to excel in the near future compared to competitors. 16:13 AMD's GPU drivers have issues and require constant mitigations 18:10 Challenges in competing with NVIDIA and AMD's intentions 22:37 Challenges with AMD firmware and driver 25:18 AMD lacks desire for greatness and won't surpass Nvidia in machine learning. 31:38 Analyzing AMDGPU ioctls and GPU memory 35:47 Discussing AMDGPU ioctls and GPU memory in programming 45:15 Exploring AMDGPU ioctls and GPU memory manipulation in Programming 51:58 Discussing AMDGPU ioctls and GPU memory controls 1:11:37 Understanding AMDGPU ioctls and GPU memory in programming. 1:17:13 AMD's culture tolerates mediocrity 1:24:48 Acquiring a VM and setting memory policy for GPU 1:29:29 Troubleshooting GPU memory allocation errors 1:38:27 Exploring GPU memory allocation and mapping in detail 1:41:52 Needed to handle memory allocation for the GPU 1:50:43 Exploring GPU memory manipulation and debugging in programming 1:55:41 Exploring GPU memory allocation and mapping 2:04:29 Understanding GPU memory allocation and device mapping 2:10:14 Casting and mapping GPU memory in Python 2:21:23 Exploring GPU memory allocation and flags 2:24:49 Understanding GPU memory allocation and guard pages. 2:31:45 Exploring mapping GPU memory and TMR 2:34:59 Explanations on GPU memory allocation and kernel interfaces without using AMD userspace 2:39:21 AMDGPU ioctls and GPU memory 2:53:43 Exploring ways to bypass Python execution environment for graph construction in machine learning. 2:58:38 Comparison between Intel Arc and Nvidia GPUs 3:01:31 Competition with Nvidia may lead to price adjustments for gaming cards 3:07:21 Discussion on the challenges of writing compute drivers and the need for better programming languages 3:11:21 Discussion on reverse-engineering AMD's userspace drivers 3:15:51 Explanation of Ops and UOPs in GPU processing 3:18:13 Indexing movement operations in terms of assembly instructions for GPUs. 3:22:30 Discussing cash aware matrix multiply and graph compiler in computing 3:24:31 Discusses progress on various programming concepts and tools 3:29:12 Optimizing GPU memory allocation for improved performance 3:31:26 Explaining the need for specific hardware command cues over graphs in GPU memory 3:36:00 Improving AMD through rigorous testing 3:38:19 Discussion on generating repeatable examples for open source vs. business approach 3:42:41 George Hotz emphasizes on speed and Nvidia over AMD for normal users. 3:44:55 George Hotz criticizes AMD for not open-sourcing and improving their drivers. 3:50:36 Advocating for open-sourcing AMD firmware to avoid exploitation. 3:52:41 Warning against illegal activities and encouraging leaks 4:01:22 Challenges in creating a GPU simulation environment 4:04:35 Discussing chip tape-out cost and fundraising for Tiny 4:09:52 Planning to take out 12nm chip, closer to GPU design 4:12:56 Emphasizing respect for greatness in technology innovation 4:20:20 Discussing plans to own Fab equipment and create own atoms 4:25:14 Exploring FPGA technology and its potential in GPU memory 4:40:43 Discussing GPU options and considerations for software development 4:47:33 Discussion on GPU memory specifications and comparison with other GPUs. 4:56:44 Understanding AMDGPU IOCTLS and GPU memory 5:01:15 Troubleshooting GPU memory allocation and pointer lookup 5:12:18 Discussion on different Q types and AQL packets in AMDGPU driver development 5:15:22 Ensuring minimal GPU resets for stable performance 5:23:21 Identifying and addressing performance issues in AMDGPU ioctls 5:25:56 Exploring kernel optimizations and AMDGPU ioctls for GPU memory 5:32:22 Quality PR with version checking and pre-re improvement 5:35:12 Provided feedback on questionable code quality 5:39:59 George Hotz is discussing AMDGPU ioctls and GPU memory. 5:42:20 Development of new driver 'kfd' 5:51:21 Importance of honest communication in decision-making process 5:54:06 Opt for written communication over phone calls to avoid miscommunication 5:59:19 Effective communication for successful collaboration 6:02:24 AMD's lack of aspiration to greatness 6:06:54 Discussion on hardware bugs in AMD GPU and their impact on schedule 6:10:08 AMD's rushed launch and bad decision-making 6:18:06 Understanding memory mapping and allocation for GPUs 6:23:03 Troubleshooting GPU memory mapping and manipulation 6:30:23 Building custom GPUs and addressing hardware failures 6:33:27 Discussing the potential of AI and limitations of GPU memory usage 6:41:47 Identifying hardware issues with AMD GPU 6:48:57 Troubleshooting GPU memory issues 7:00:58 Challenges with AMDGPU ioctls and GPU memory 7:03:53 AMD does not have the capacity or desire for ML, lacking software development. 7:11:00 George tested a crash on computers and discussed issues with AMD 7:13:28 Fixing AMD GPU crashes with external Raspberry Pi monitoring 7:20:06 Discussion on AMDGPU issues and desire for reliability. 7:22:56 Challenges with AMD's GPU and software cooperation Crafted by Merlin AI.
    170
  14. You're so inspiring in the simplest, most dumb way. You reveal so easily how simple the world is, and it always makes me mad that I knew the answer - I just didn't "see" it.
    162
  15. Thank you for adding the chapters. Appreciate all the hard work you put into these videos.
    147
  16. George vs the mute button, the battle continues
    139
  17. thanks for uploading these streams to YouTube
    128
  18. Love the just chatting sessions - thanks for uploading
    128
  19. "I don't really like talking about myself either. And most of these things here I don't really talk about myself. I do give anecdotes of my life but they're more to illustrate a point. Because anything I say about myself is personal and useless to you. Don't be like me, that's dumb. I try to give you information so you can think." - George Hotz
    120
  20. I don't know how does the channel create so many usefull timestamps in this giant videos, but the timestamps are a big part of why I usually check out these videos, I can just go to the stuff that interests me, its amazing.
    118
  21. Thanks for providing timestamps ❤️
    116
  22. Bruh he just turned a fun project into a serious nn framework wtf
    112
  23. and goerge, yes we care about what these things are about.
    109
  24. I love watching videos while having absolutely zero idea what's going on
    108
  25. glad to see George is back posting! thankful for the archive, keep it coming keep us inspired and pushing this world forward
    108
  26. George dressed like Kanye
    107
  27. NEVER LEAVE US AGAIN GEORGE!
    101
  28. these are the best chapter annotations I have ever seen, it's extremely valuable
    101
  29. Thank you for keeping up the George Hotz Archive! Appreciate it!
    99
  30. The tenstorrent guys in chat trying to work with george while he rambles about bringing back the gold standard - what a time to be alive
    98
  31. "the best way to get personal gain is to make the world better" -Geo love dat shit.
    98
  32. I am seriously considering subscribing on Twitch. We need to support this kind of content!
    97
  33. Whoever maintains this channel is amazing. Almost 9.5 hrs with nicely segmented sections. God bless you, maintainer.
    97
  34. the job on the timestamps is impressive. Incredible good job, its really super useful!
    96
  35. Musk presents robot as somebody in a costume, George presents actual robot. +1 Mr. Hotz
    94
  36. AMD should start paying you money only because you are doing this. There 20-30k of work out there for average engineer with access to internal documentation.
    92
  37. Thanks for the timestamps and upload bro. Incredible work
    87
  38. Respect for this man to put all those chapters. Damn, would that be tedious task.
    81
  39. when I see knowledge of this man, I feel honored to watch this streams..but on the other side I feel useless with my knowledge 😂😂
    78
  40. Its just amazing that someone just takes out the time to actually educate (give nooooob lessons) to nooobs !! thanks George.
    78
  41. Seems more entertaining to watch George stream than to watch a gamer game
    73
  42. Where have you been, man? You have been suspiciously quiet on Twitter
    70
  43. George, software development is a category on twitch.
    69
  44. The fact that you added the paper in description is so good. Also, people who helped recover the stream, much thanks.
    69
  45. Thank you for taking the time to make all the chapters
    68
  46. Happy New Year George! (and everyone else reading this)
    67
  47. Proud to have been the first successful caller (Penguino138)!
    66
  48. We are so back
    66
  49. This stream was SO GOOD. Great rants and thought experiments while very funny.
    65
  50. I like how when he reads documentation he violently selects and deselects the area he's interested in.
    65