/biz/ - Business & Finance

>>58600741
Because you need to be rugpulled.

>>58600741
I sold at 200. Contemplating suicide everyday. I cant seem to bring myself to pull the trigger. I only have one round of .38.

>>58600741
this is the end times, the average wagie will be replaced

>>58600741
https://www.youtube.com/watch?v=VVxYOQS6ggk

>>58600806
kek

>>58600806
>I only have one round of .38
even if you perfectly hit the brain, instead of killing you it'll just leave you retarded so effectively nothing would change

>>58600741
Because I didn't buy at 900. If I had it would have crashed to 500 by now.

>>58600741
Don't worry it's going to 100 then you can buy 10!

>>58600741
Nvidia's Geforce NOW servers were shitting the bed so hard over the last few days I think they're under siege. There's definitely something fucky going on they're not telling.

>>58600806
just buy crypto

>>58600741
>why is the company printing creating the shovels for the trillion dollar AI gold rush pumping
IQ test

>>58600741

You can screenshot this, if you like. NVIDIA is massively overvalued, it is worth nearly $3 trillion despite the fact it is no where near as influential as apple and microsoft. It will collapse, sooner or later, when the AI bubble bursts. Within the next few years, it will dramatically dump down to $200, before making a rebound to $600. Although it will be eons before it ever gets back to 4 figures. Mark my words.

>>58603664
I don't own any NVDA but I saved your post for when it hits $2000 this year

>>58603560
Kek ai isn’t as useful as you think. It’s like nano tech. You will get a couple cool things but it isn’t revolutionizing anything. I don’t think it’s even as revolutionizing as the lithium ion battery. NVDA is benefiting from people's belief bigger models are better and somehow we need one model to rule them all. We probably passed the main gains in ability from scale, and as we get bigger it is just diminishing returns.
The language ability of much smaller models is very good, and for business apps worried about liability and correct answers (no hallucinations) investing in proprietary data with RAG is probably the answer to most questions.
Faster processing to reduce latency is one thing but then you have remaining network latency. There is a reason OpenAI plugged in the phone on a hardwire for the GPT-4o SKY demo.
With 70% profit margins, NVDA has to attract competition. Market is too rich for it not to happen. The issue is whether or not, GOV allow the big guys to buy up any promising small tech which could be a competitor. That has been to MO of the big software tech. NVDA is trying to copy that playbook.

>>58603685
>I don’t think it’s even as revolutionizing as the lithium ion battery
It's easy to look at literal beta technology and think it's the whole package. Basically no serious AI scientist or developer has gone on record saying that this is it. It's dumb to look at it as a mature technology because it is the literal opposite of mature technology. It might SEEM like yesterday's news because it's been two or three entire years since the first chatGPT model was made available to the public, but you have to realize that it's still at an early stage and under massive development by some of the most brilliant minds of our generation.

>>58603697
>the most brilliant minds
Like chang and raaj?

>>58603697
People will use ai for drone swarms, shitty customer service, optimization of big data sets, but you won’t see anything useful. Companies will profit from it, mostly off people like you that will overvalue their companies because of le ai

>>58603685
>internet isn't as useful as you think, just some autists on college campuses messaging each other
boomer mindset

>with 70% profit margins Nvidia has to attract competition
you think Intel hasn't been trying to get into GPUs for years now? Only people who stand a chance are Chinese government backed competitors but they'll be banned from most Western nations, Nvidia has an effective monopoly for at least a decade and that's why the stock price is pumping

>>58603756
>Comparing the internet to machine learning algorithms
Not gonna make it

>>58600806
you should have 2 just in case you miss once

>>58603756
I don’t think you understand what I’m saying. Take this from someone that uses AI and MLA regularly with their job. It is a useful tool. You can optimize certain systems with it but it isn’t revolutionary like the internet.

>>58603731
Chang and Raj are smarter than you, yes

>>58603811
Brown hands typed this.

>>58603787
>Comparing the internet to machine learning algorithms
people used to compare the internet to fax machines
>>58603806
my take really depends on whether the next generation of models are as big an improvement as GPT-3 to GPT-4. If so, it seems like exponential improvements might continue. But the fact competitors have caught up but not surpassed GPT-4 level performance and OpenAI hasn't released anything impressive in a year has me thinking they have hit a wall with current model architectures and AGI isn't on the table yet. Nvidia will keep printing money regardless

>>58603806
Doesn't matter. NVDA is the only competent GPU manufacturer and has an 80% market share. All roads lead to NVDA.

LLMs basically have no real practical business case, their nature means they will always have a fairly significant error rate, and they can't really be used anywhere that a 5-10% margin of error isn't acceptable (which is basically everywhere)

This whole bubble is literally just retards who don't understand what a LLM is vs AGI, it's going to crash hard and be mad funny.

>>58603853
literally anyone who uses it knows claude opus is superior lol, it's not even close.

>>58603853
also sorry you are such a fucking retard
>AGI isn't on the table yet
AI isn't some spectrum where LLM's get more and more refined and eventually you get AGI.
A LLM is literally just a recursive function with a stupid amount of random number generators, that has a cool feedback mechanism for training, so you get interesting output that is relevant.
It literally has nothing to do with AI/AGI.

https://cspeech.ucd.ie/Fred/docs/WhyHeideggerianAIFailed.pdf
this is a good essay on what would be involved for actual AI/agi

A.I. ain’t Christian. We need to stop developing it or it gonna destroy all of us.

>>58600741
you bought

>>58603853
4o is worse than 4. Pic related. It hallucinates, it has too much data and it doesn’t know what’s right or wrong and it doesn’t actually think.

>>58604127
>no practical business use case
I use it for coding daily, almost anything that a junior dev can handle can be done by AI at this point. Tons of artists, designers, stock photo people, etc. are finished

>literally anyone who uses it knows claude opus is superior lol, it's not even close.
but not dramatically in the way GPT-4 was better than GPT-3, that's what I'm looking for otherwise it seems there are diminishing returns on data size and training costs until they figure something else out

>>58604490
>use this section to make a slide in a presentation to some retards that don’t understand anything but are in a position of power.
That’s all I use gpt for. I use ai and mla to make algorithms that are slightly better than what is existing today, but I don’t see a major use case for it either.

>>58603664
>AI bubble
I don't think you understand just how powerful an AI will be =D

>>58604432
There isn't an AI around that actually thinks, they just display what it's algorithm thinks you want as answer.

AGI is what will actually be thinking, but we don't have one yet

>>58600806
Is it at least a hollow point?

>>58603664
STFU, I spent my dog's heart surgery money on nvidia stock. I need this to pay off.

>>58604518
>but I don’t see a major use case for it either.
that's because their isn't one. Laying off a few pajeet coders/HR roasties/useless middle management to and spreading whatever work they actually did among more competent people sounds nice in theory, but then you realize that's it. These massive companies are pouring billions upon billions building up data centers with NVDA chips chasing a few millions in potential profits. The senior management of these companies likely aren't that dumb, but are chasing short term stock gains with huge spending hype waves because they don't have any actually innovative/disruptive ideas.
The best depiction about whats happening is
>https://www.wheresyoured.at/rotcombubble/
NVDA will be fine for some time as the orders keep coming in, but eventually will drop substantially when investors realize they aren't going to be able to continuously grow revenues by billions every quarter as large companies stop spending.

>>58604544
AGI is like a super conductor

>>58604432
It couldn't tell me when the next total solar eclipse in Europe is either (it gave me the date for the US one). It's just as retarded as Google which is also worse than a decade ago.

>>58604686
Which is also interesting. The invention of the lithium battery came from someone searching for a superconductor. It was an accidental discovery.

>>58602722
Checked
Then he will be .38 special

>>58604268
AI is a misnomer. There is no intelligence and there is no will because both of these are God given. Humans will never create truly artificial intelligence it is more than likely impossible although I don’t claim to know the will of God.

>>58600828
It's ironic because wagies just needed to buy NVDA shares

>Imagine not shorting that thing

>>58605062
>shorting NVDA
Lol

>>58600741
Cuz you are getting shitplays, instead of fundamental ones such as shrl kek