Nvidia Part III: The Dawn of the AI Era (2022-2023) (Audio)
Vložit
- čas přidán 5. 09. 2023
- It’s a(nother) new era for Nvidia.
We thought we’d closed the Acquired book on Nvidia back in April 2022. The story was all wrapped up: Jensen & crew had set out on an amazing journey to accelerate the world’s computing workloads. Along the way they’d discovered a wondrous opportunity (machine learning powered social media feed recommendations). They forged incredible Power in the CUDA platform, and used it to triumph over seemingly insurmountable adversity - the stock market penalty-box.
But, it turned out that was only the precursor to an even wilder journey. Over the past 18 months Nvidia has weathered one of the steepest stock crashes in history ($500B+ market cap wiped away peak-to-trough!). And, it has of course also experienced an even more fantastical rise - becoming the platform that’s powering the emergence of perhaps a new form of intelligence itself… and in the process becoming a trillion-dollar company.
Today we tell another chapter in the amazing Nvidia saga: the dawn of the AI era. Tune in!
Sponsors:
Thanks to our fantastic partners, any member of the Acquired community can now get:
Scalable, clean and low-cost cloud AI compute from Crusoe, and listen to our recent ACQ2 interview with CEO Chase Lochmiller
bit.ly/acquiredcrusoe
bit.ly/CrusoeACQ2
Your product growth powered by Statsig
bit.ly/statsigacquired
Free access to our episode research on Blinkist, plus our favorite books on Ben & David’s Bookshelf
bit.ly/BlinkistNvidia
bit.ly/BlinkistBookshelf
More Acquired!:
Get email updates with hints on next episode and follow-ups from recent episodes
www.acquired.fm/email
Join the Slack
acquired.fm/slack
Subscribe to ACQ2
pod.link/acquiredlp
Become an LP and support the show. Help us pick episodes, Zoom calls and more
acquired.fm/lp
ACQ hats are back in stock in the ACQ Merch Store!
www.acquired.fm/store
Links:
Asianometry on AI Hardware
• AI’s Hardware Problem
Episode sources
docs.google.com/document/d/1d...
Carve Outs:
Alias
www.imdb.com/title/tt0285333/
Moana
movies.disney.com/moana
Note: Acquired hosts and guests may hold assets discussed in this episode. This podcast is not investment advice, and is intended for informational and entertainment purposes only. You should do your own research and make your own independent decisions when considering any financial transactions. - Věda a technologie
This is seminal work. You guys should be paid for this and I would gladly pay for such quality work.
The old Internet adage applies. If you're not paying for some good on the Internet, you're the product.
@@darylallen2485yes, these guys are definitely getting paid.
Nvidia’s advantage against AI competitors: “10K person years have gone into CUDA [since 2006].” 😮
damn I had never listen to 3 hours long podcast in my entire life.these guys made me listen all of this. crazy work
this is so insightful, thanks for putting this together. -from a new subscriber 😊
🎯 Key Takeaways for quick navigation:
00:28 🎧 Episode Overview
01:42 📉 Economic Conditions and Nvidia
02:21 📈 AI’s Transformative Role
07:42 🌐 Nvidia’s Current State and Growth
09:15 🧠 Evolution of Neural Networks
11:28 🖥️ GPUs and Applications
13:52 🤖 Influence of AlexNet Team
17:53 🎯 Profits from Google Brain
18:47 🌐 AI Monopoly Concerns
20:10 🤖 OpenAI's Vision
24:02 💡 Crusoe and Data Centers
30:59 📜 The Transformer Paper
Made with HARPA AI
Ai gave up after half an hour
I’ve been waiting for this episode 👏
Brilliant episode
Great guide to understanding NVidia in the context of today's computing world.
There are many more details about Nvidia datacenter products. The cabling is different as it’s all optic. The power requirements are bigger. The cooling is different. Inflection is building a multi-billion datacenter from the ground up based on Nvidia specs.
Superbly done. Thank you guys.
Wonderful program. I learned a lot of valuable, trustworthy data/information. Thanks for all your work in putting it out. :-)
36:30 - dota bots may have used screenshots in some capacity, but my understanding based on an open ai podcast, is they choose dota for their bot project due to it having an api under the hood.
Interesting how Nvidia’s model is now basically “doing things that other people can’t do”. Even though they started by doing things ~90 other companies could do
GREAT INTERFACING, AUTOMATED, ITS REALLY USEFUL ANSWERING PATTERN FINACIALS
Informative show..thank you!
My overriding takeaway? Me buy nvda. Me hold nvda long, longtime...while ignoring the transitory din of perma-bearish doomsayers & fickle analysts.
amazing analysis 👍
Amazing discussion, throughly enjoyed it 😊 The entire discussion has been a gripping story to know the past and future where Nvdia is heading 😇 Sub'd on all platform 🎉 I hope to see more of Ark investment companies like Ui Path, Teledoc, Roku, Crispr, intellia etc.., in the future like NVDIA 😊
Do they ever do video anymore, or is it always just audio?
We will when there's in-person recordings with guests!
@@AcquiredFM we miss your faces.... Also, cameras will help with audience growth on yt. (I know, I know, free advice - worth exactly what you pay for it.)
videos will be great@@MrBrentlilley @acquiredFM
THE BEST! Learned at TON... LONG $NVDA
Why did Nvida wait so long to get into the AI space? None of their earnings calls use the term AI then suddenly in 2022 that's all they talk about.
Lol, did you listen to the podcast? They developed CUDA in 2006. CUDA was used in the machine learning "big bang" in 2011. What part of that is an example of "waiting to get into ai"?
@@darylallen2485 should have been released in 1999
@@kennyadvocat do you know of any example where a gpu was used for anything other than gaming in 1999?
@@darylallen2485 nobody could cause nvida didn't build any. They were gatekeeping the tech. 😢 They didn't even talk about ai on earnings calls till 2022
AMD doesn’t have anything like Nvidia’s NVLink. They tried and failed. So you can’t stack well multiple AMD GPUs efficiently for training.
NVLink was so good Nvidia had to remove it from their consumer cards because they were eating their data center products from below.
Bear case: a lot of Nvidia’s clients go out of business because of AI making them obsolete.
Great observation, but clients going out of business means the value of the task still exists. If the task is now being done with ai, Nvidia will still be positioned to provide that value with their hardware. If anything, Nvidia is just cutting out the middle man.
bear case : there will always be an alternative, or otherwise antitrust from DoJ or EU
Nvidia is so awesome and Jensen is so nice. I could not expect that Jensen would join this channel one month later.
As much as it's interesting, you guise lost me at the marketing. You gotta make money however you can, fine, great, all well and good I'm sure you have tentacles all over the right places. It colors the optics with a tint and there's no plausible deniability. At least you're being demonstrably honest. Going into a subject such as LLMs and the future of this long simmering paradigm with tinted glass is as fundamental an antithesis as the word Open is to Open AI. Once the path to wealth is cleared of the murk all other considerations become secondary or quite frankly irrelevant filtering for the wavelengths you want to see. No one is developing AI now in order to genuinely benefit humanity. It's only a matter of market share. I have zero doubt that you're only scratching the surface from a particular perspective which is obviously the point, that is not lost; but I guess you have a target audience well suited to jump on bandwagons after the horses got going and not before which is about as disingenuous as selling second hand kit kat bars. Yes, everyone can do their own research and they should so as to not need anything more than a reminder of the vastness that is the known unknowns when it comes to integrating technology with civilization.
Am I being over-critical? Yes. Is it unfounded with a slant towards caution? No.
It's really easy to convince stupid people of practically anything while only in alpha stage of development. There are so many examples of this, it's dizzying.
When the question of misinformation comes up, this is what I think of. Same approach I take to a journalist who has an outcome in mind not talked about but hinted to in an article. It might not be intended but give it several years. It will always be wrong at some point especially when evolution and mass capital are involved.
First
If you were a hedge fund and had 100% of your fund invested in NVDA, you would be more measured than the sucking up these two guys did. Embarrassing! Let this recording stay as proof how embarrassing this prediction is going to be..
There is such a thing as a balanced perspective and that is completely missing 😂
Save for the well deserved hate for their closed source drivers, the rest of Nvidia is amazing. They excel at developer support. Try developing in any other platform. As per a first person account said on TW:
“I once emailed Jensen cold last year and he got a driver bug fixed in 12 hours. 30 years after its founding, this guy still acts like a hardcore founder.
Jensen is the best tech CEO in the game rn.”
U are talking all about ai....but this vid does even had a video to show😂😂😂😂...