FIRE? 
No. I sold quite a bit whenever it tanked and then recovered its previous high. As a result, i am still working. 
Agree it is a excellent analysis.
“The reason for that is because usually some 5%, 10% of the code represents 99.999% of the runtime.”
That statement is gold, capturing the accent why accelerated computing is needed. This statement is not that true when computers are used for communication (aka data retrieval) which was how we have used computers for the last several decades. This statement is very very true when computers are used for computation, which is what it supposed to do.
What is the ultimate way to scale that frequently run codes? Of course, it is ASIC, which is built to run that piece of codes and that piece of codes only by burning the logics into silicon. That is why all crypto work load are now run by ASIC. That is why ai work load will switch to company like Etched (provided that transformer is still the logic which is very likely) Does it spell the end of NVDA? It may drop for short term, but for long term, NVDA will still prevail because gpgpu is undoubtedly a necessary sweet spot between CPU and ASIC.
“What NVIDIA is really good at, as it turns out, is creating new markets.”
That statement is not entirely true because correlation does not always mean causation. It is not NVDA creating new markets, it is the new previously impossible market finding NVDA.
That will for sure happen over and over again, because after all computers should be used for computation, not communication.
2 Likes
First wave…
Call Centers going away soon.
Data Analytics software going away soon.
Analysts going away soon.
Socially acceptable mobile computing device…
Both Orion glasses and (in the lab) Apple glasses are not the right form factor.
AirPods could potentially become one if Apple knows how.
Note: Btw, I always thought that AirPods and Watch should replace iPhone, and Vision/VP should replace Macs.
NVDA, AVGO, TSM and energy stocks have already rocketed. Any steam left?
Disclosure: Own NVDA stocks. No plan to sell nor add.
An AI-powered search engine.
Not sure it’s entirely due to AI, but tech job market is taking a big hit recently. Maybe remote work also plays a role as companies can more easily hire off shore?
This Berkeley professor said some of his 4.0 students can’t get any offers.
.
In general, over production of degree holders.
You get what you pay for. The reason companies use the US and UK is that some of them require competent customer service for their business model to work and for them to be competitive. No idea how GPT rates quality wise but that’s the more important issue.
Meta announces new AI model that can generate video with sound | Meta | The Guardian.
Mark goes all-in to dominate metaverse aka AI/AR/VR space. From hardware to OS to applications. He wants to rule the future.
At this rate, META is a potential $10T stock.