NVIDIA shows signs ... [2008 - 2017]

June 27, 2017

https://techcrunch.com/2017/06/27/v...and-its-use-of-ai-beyond-autonomous-vehicles/

 

http://www.eetimes.com/document.asp?doc_id=1331949&_mc=RSS_EET_EDT

 

NVIDIA Tops MIT Tech Review’s List of World’s Smartest Companies
https://www.technologyreview.com/lists/companies/2017/

 

https://techcrunch.com/2017/07/05/n...oss-cloud-self-driving-academia-and-the-home/

https://www.theverge.com/2017/7/5/15923966/baidu-apollo-self-driving-car-platform-ford-intel-nvidia

 

AI Generates Songs to Resemble Kurt Cobain

https://news.developer.nvidia.com/ai-generates-songs-to-resemble-kurt-cobain/

 

https://techcrunch.com/2017/07/12/n...ds-first-level-3-self-driving-production-car/

 

6 NV processors per Audi A8? Not a bad haul even if it's a lower volume car targeted at the upper middle class. 9,521 A8/S8's sold in Europe and NA in 2016. The model has been in decline since 2011, and 2017 looks to be lower still (going by US sales numbers for the A8). Still I'd guess anywhere from 8,000-10,000 should move after it's launched, that's 56,000-60,000 processors (I'm assuming it'll be going into the S8 as well). Pretty good. And it's possible Level 3 self-driving may be enough to revitalize sales of the A8/S8 so that'd boost it even more, plus whatever it moves in the rest of the world (I imagine it's popular in the Middle East).

Regards,
SB

 

As best I can hear, it's complete gibberish. Semantically empty. However, I don't know what Kurt Cobain's lyrics are like...

 

http://www.azlyrics.com/n/nirvana.html There you go

 

T'would appear the AI lyrics were spot on! Although I consider it cheating to pick such a low target. Don't give us poetry - give us accurate technical documentation on a software library the devs don't properly support.

 

This may be the product of at least two possible sequences of events:

Sequence 1:
Researcher A: "Hey, let's build an algorithm that learns from song lyrics and writes new ones in the same style!
Researcher B: "Cool idea!"
Researcher A: "I know, right? And then, let's test the algorithm with a songwriter we like!"
Researcher B: "Great! How about Kurt Cobain?"
Researcher A: "Yeah, I like him too!"
*Development, machine learning, song generation*
Researcher B: "Wow, this was such a great idea, it works really well!"
Researcher A: "Totally!"


Sequence 2:
Researcher A: "Hey, let's build an algorithm that learns from song lyrics and writes new ones in the same style!
Researcher B: "Cool idea!"
Researcher A: "I know, right? And then, let's test the algorithm with a whole bunch of different songwriters!"
Researcher B: "Great idea, that way we'll know it's robust!"
*Development, machine learning, song generation*
Researcher B: "Wow, this is complete gibberish. This project is a total failure."
Researcher A: *sigh* "Yeah. Well, we tried. Man, this is such a disappointment. The lyrics generated are so nonsensical they might as well have been written by Kurt Cobain!"
Researcher B "Hahaha, that's true!… Wait, I think I know how to salvage this project!"

 

Tesla V100: Nvidia hands over first Volta computing cards to deep-learning researchers

https://www.heise.de/newsticker/mel...karten-an-Deep-Learning-Forscher-3781130.html
Guess some benchmarks will soon appear on the net.

 

http://www.tomshardware.com/news/ai-robot-training-isaac-holodeck,35126.html

http://www.tomshardware.com/news/nvidia-titan-xp-quadro-external-gpus,35123.html

 

The future of top 40 shit music.

 

I have a feeling we'll be saying the same once our jobs are replaced with AI.

 

NVIDIA Announces Earnings Of $2.2 Billion For Q2 2018

http://www.anandtech.com/show/11711/nvidia-announces-earnings-of-22-billion-for-q2-2018

 

It's still hard to fathom how once ATI and NVidia were roughly the same size when AMD (who at the the time dwarfed both) bought the former; how the companies came into parity again and now Nvidia is a much larger company, to the tune of $1B per quarter. All the more so considered all the ink spilled at the time regarding the dim prospects of Nvidia's survival in the face of CPU-GPU competitive paradigm they could not match.

 

To me it shows Nvidia's adaptability and forward thinking. Investing in ARM processors may not have been profitable for mobile, but they found a market where it was. Meanwhile, AMD and Intel both missed the mobile market and Intel may be too late for automotive as well (although this market is new enough a lot could change).

 

It's never too late. Look at how AMD has managed to come back with a great CPU option and is gaining market share. Conceptually impossible considering Intel had way more money to invest in R&D and the best fabs in the world. Intel, or AMD, or anyone, could yet usurp nVidia in the Automotive space.

 

And mobile? Intel and NVidiaNVidia have effectively abandoned it. Not too late?