Stock Market/Investing/Fin Planning Catch-All

  • Thread starter Thread starter heelslegup
  • Start date Start date
  • Replies: 474
  • Views: 12K
  • Off-Topic 
With Trumps AI Infrastructure announcement, wondering where to make some returns. SFTBY and ORCL I can buy, but I can't buy OpenAI. Makes me wonder anyway, do you think we really want AGI? TRUE AGI? We will need to do analog computations. Period. And simulating analog with digital is by definition impossible and in practice extremely difficult and computationally expensive. We need other innovations need to occur as well. TBH right now our AIs are mostly just the "Chinese room paradox" in action. I think the real play is in nuclear energy, otherwise we run out of the needed power. If I ask, "how do I make Salisbury Steak", the typical ChatGPT types AIs right now use 10x the electric power as the same query in google or some standard search.
 
With Trumps AI Infrastructure announcement, wondering where to make some returns. SFTBY and ORCL I can buy, but I can't buy OpenAI. Makes me wonder anyway, do you think we really want AGI? TRUE AGI? We will need to do analog computations. Period. And simulating analog with digital is by definition impossible and in practice extremely difficult and computationally expensive. We need other innovations need to occur as well. TBH right now our AIs are mostly just the "Chinese room paradox" in action. I think the real play is in nuclear energy, otherwise we run out of the needed power. If I ask, "how do I make Salisbury Steak", the typical ChatGPT types AIs right now use 10x the electric power as the same query in google or some standard search.
not sure that Open AI is going to beat X (grok) or Google (Gemini) in the long run.

Yeah, we are going to be very energy deficient very quickly with these computing advances and absent some type of huge breakthrough somewhere in the energy sector then Nuclear will be the only answer.
 
An AI investing play is likely in power infrastructure - perhaps even uranium. That said for example PDN is already up 10% today so that may be late.
 
VST Vistra and
GEV GE Vernova
CEG Constellation Energy
TPL Texas Pacific Land
TSMU Taiwan Semi 2.0 Bull
CRWL Crowdstrike 2.0 Bull
ARM ARM Holdings
VTV Vertiv
AMAT Applied Mat4erials
PSTG Pure Storage
ARAN Arista Networks
MPWR Monolithic Power


both have perormed well
Lots of ways to go at the various AI plays from Mag 7 to power to data centers to software and fields yet to come
 
Last edited:
Re the Ai announcement, Altman made a post about it on X, Musk gave a snarky reply "The money ain't there" so they are going at it.
IIRC, Softbasnk promisd some large amounts of $$$ during Trump 1 which mostly did not show up. So we shall see. Also some of this buildouit was already underway during Biden Admin with dufferent funding, Altman has been hitting up anyone with 10 figure funds including Middle East oit typles.
 
We will need to do analog computations. Period. And simulating analog with digital is by definition impossible and in practice extremely difficult and computationally expensive.
???? I don't understand this at all. First, simulating analog with digital is by definition possible. I mean, you're using the word simulation -- which implies less than perfect accuracy -- but leaving that aside you can prove that Taylor series, Fourier series, Fourier transforms and other operations can perfectly replicate any function. It can be computationally expensive in some cases, but it's not difficult (FFT has been around for a while and is included in the standard library packages of most programming languages I know). Turing proved that a simple digital machine that moves on a tape can replicate any computation and implement any algorithm.

Second, what do you mean, we will need to do analog computations? Why? In what way will the need for analog computations exceed the current need for them, which is already well handled. I mean, AIs are TRAINED with analog computations performed digitally, and they compute with digital simulations of analog computations.
 
not sure that Open AI is going to beat X (grok) or Google (Gemini) in the long run.

Yeah, we are going to be very energy deficient very quickly with these computing advances and absent some type of huge breakthrough somewhere in the energy sector then Nuclear will be the only answer.
Why would X have an advantage over Open AI? It doesn't have more money. Open AI has a big lead. And X is led by a bozo who has no idea what is going on.
 
. TBH right now our AIs are mostly just the "Chinese room paradox" in action.
They most assuredly are not. They have reasoning capabilities, and they can perform logic. I've tried it out many times. Apparently they can also solve math problems.

It's easy to see them as chinese room machines, and intuitively it seems that they should be based on how they work. But they are more than that.
 
Back
Top