It's been a while...
I
have begun more formally documenting my AI, ML and generative AI studies. It all
started with Robinhood, bitcoin mining, pancake swap and a different thought
process. I tend to take on technology as a hobby and learn to apply it to business
as I see the tech play out in the real world. My journey began with looking at
Robinhood and choosing stocks based on interest. My first few picks were
Moderna, Nvidia and Royal Caribbean cruise lines. I like to choose volatile assets
to add some excitement to pretty dry subjects. Of course, as things progressed,
I picked up and dropped some AMC multiple times.
From
there I looked deeper into Nvidia’s massive rise due to other shifts in technology
such as crypto mining. You can ask me about wins and losses on pancake swap,
hyper, beefy, aave, pi, safemoon, etc. I ventured out into AI platforms for stock
trading like Delphia and saw less monetary gains and more knowledge gains.
Where
am I going with all of this? There has been a transformation in digital
knowledge in the form of data sharing (whether you like it or not) that has led
to data mining and massive advancements in artificial intelligence. Now we are
looking at a shift from crypto mining rigs eating up available GPU to data mining
rigs.
I priced
out an Exxact deep learning workstation and it costs a lot of money. I configured one
for a quote of $33,000. I figured it would save me money on Amazon EC2 P3
instances powered by NVIDIA V100 Tensor Core GPUs. “Solving these kinds of
problems requires deep learning models that take a long time to train. With 640
Tensor Cores, Tesla V100 GPUs that power Amazon EC2 P3 instances break the 100
teraFLOPS (TFLOPS) barrier for deep learning performance.” – I never knew how
many TFLOPs I would need and it’s not something I can really ask for as a
birthday present.
Why
would I need all of this? Simple – children’s books. Stayed tuned. The next
article comes quick.
Least
common denominator and not advice: Nvidia will continue to explode. Every
threat is an opportunity. Buy more FLOPS since you never knew you needed them. Stay
tuned as I refine my months of notes into something that reads with some semblance
of sanity.
*Disclaimer:
Not written with AI, if it was there would be something really wrong with its trained
data set.
Comments
Post a Comment