With all this hype talk of Ai and generative ai in general what value is it adding to the companies buying up these nvda chips and trying to out do their competition? Meaning where is the value in these companies implementing ai today? What if these companies trying to implement the use of ai into their company see no return from these investments? Companies are investing billions in technology to build ai applications, what if this is just a total flop? Not saying a total flop to the industry but too companies who think they will exceed in this area just because they believe it's the next new technology boom....everyday there is another company announcing their move into ai, to me that's just showing how trendy these companies are trying to be to extract more money from traders and shareholders, buying up stock in the world of ai hype....
90% of the investors/traders can't make it due to lack of KSA knowledge/skills/ability inability to manage the mind, poor risk management, laziness etc Similarly, some/many companies investing in AI will not make it due to their lousy technology poor management high operating cost etc Look at the US memory chip maker. There were tens of them. Now there is only one.
Five years ago you could go to your cable company website and type in a question and it would try to answer it based on keywords in your query. They gave that a cute girl's name like "Ask Jenni". Today they call it ai - without necessarily having made any changes at all. Years ago we stored things on internet servers, then one day someone decided it would be super sexy to start calling that "the cloud". Nothing changed except the marketing. Those functions (both ai and the cloud) continue to evolve as they would have without the new names.
what’s funny is the blockchain was a the opposite. There they bragged about the tech without and use cases and now use cases are coming up (like Walmart managing their lettuce supply chain)
This is an excellent point! In this great analysis, prof. Galloway argues we might already be in an AI bubble, making exactly your argument. He compares it with the Cisco craze of late 90's: Cisco rallied from ~1994 to 2000 about +2,500%, selling networking hardware to basically everyone because of the internet Then, as several of their customers didn't make it through early 2000's (lack of b2c demand), its stock price plunged almost 90% He argues that similarly NVIDIA is selling H100 GPUs to everybody... but if their customers fail to find viable business models for AI (no one found it yet), its stock will plunge as Cisco did. Funnily, as he says, if you add up all revenues generated by all AI companies in past 12 months, it's a microscopic $20 billion, a tiny value compared to the trillion dollars added to NVIDIA market cap... NVIDIA's customers must accelerate their sales because of AI, otherwise the stock will tank (he argues..)
Yeah its bubble for sure, but could still go up another 100% before it pops. Nvidia and MSFT could end up with 10T combined market cap before its over. Rolling out AI to the point it justifies whatever the peak valuation ends up being could take a couple of decades.
Could be that AI turns out to be mostly efficient in "back office" functions. That would lead to reduced personnel to man those functions, but how much would that add to the "bottom line"? I dunno.... could be all the "AI hype", when it comes to earnings growth and stock prices... is little more than "much ado about little".
It's highly likely that "AI" will be a total flop. Machine learning techniques have been rolling out across the economy for 10-15 years now, but the concept of the all-in-one LLM appears to be a massive misdirection of effort and resources. What will basically happen is that orders for the infrastructure (chips and data centers) will dry up, companies like NVDA will get crucified with shares falling 80% or more as orders implode. The hyperscalers will see a modest hit on revenue, and a multiple contraction proportional to how much of current pricing is based on forecast "AI" demand growth.
Pricing and convenience changed. It used to be hard to access for the average person, then bubbly interfaces came along. Nowadays you can have a fully fledged server with many cores and all services for less than $10, that used to cost a magnitude more. Actual ML stuff is quite different from basic answer bots, not saying the companies necessarily employ that tech.
%% Sure could ; + I occasionally read Snake Oil AI, blog LOL. Good union buster/maybe a bigger trend then we ever dreamed