I keep recommending people follow @alojoh on twitter. He is the real deal and the reason for my confidence in NVDA. The best $7 a month you will ever spend if you like to invest.
NVDA is still very undervalued at $200 a share.
Honest assertions about the AI bubble
- Thread starter firegiver
- Start date
I keep recommending people follow @alojoh on twitter. He is the real deal and the reason for my confidence in NVDA. The best $7 a month you will ever spend if you like to invest.
NVDA is still very undervalued at $200 a share.
Thanks for the tip. I’m a boglehead though, so there’s nothing anybody can tell me to get me off of my index funds and chill strategy.
There is absolutely nothing wrong with that, and good for you. I wasn't trying to sway you, just discussing our previous topic of AI being overvalued. I still think it's very undervalued, my opinion of course.
Gotcha! Tbh there’s so much that goes into valuations, which is certainly not my specialty. This is one of those topics where we could disagree and both still be right about the value at different points in time.
Nvidia actually makes a product that will continue to be high demand
OpenAI missing revenue and user projections is a big sign that the bubble is popping
They're working with Anthropic, which is actually a useful company with a useful product (Claude)
I use Claude and have been impressed with it’s capabilities.
Claude is my favorite of the LLM’s
Its all I use at work and side projects
Could be a bubble, but I wouldn’t want to miss out it not. Returns could be incredible.
Yeah, this is where the different LLMs differentiate themselves the most imo
As I understand it, the big LLM agents such as ChatGPT, Gemini, and Claude essentially have multiple layers of subprocesses that convert the text prompt you give it to something it was trained on. So depending on how precise and succinct our prompt is, we can get very similar results but one version of the prompt could require a lot more tokens than the ideal succinct version.
Claude seems the least forgiving of the big three I mentioned above in terms of token usage. But it’s still my favorite by far.
I’d love to understand the economics of tokens for these agents. It’s something I haven’t had time and energy to try and understand, I just understand the usage component because I have to actively manage my quotas
One aspect of this I find fascinating is: Even if there isn’t a bubble, if one of the big companies truly achieves AGI first, how does that impact the valuations of the others?
If it indeed happens in our lifetimes it would likely have huge impacts to the value of the other companies. I guess I’m wondering if the overall value of the companies will be conserved, or if one of them getting AGI first causes at least some value in other companies completely disappear?
Is there even an agreed upon definition of what AGI is? I've heard some people say it's already here and others say we have a ways to go.
To your point, i don't think there will be one winner in the race. There will be plenty of cake to go around imho.
That’s a good question and I’m honestly not sure. I’m more of a power user of these tools and less of a comp sci researcher.
I suspect the official definition of AGI is probably very open to interpretation, but the big companies probably have specific targets or metrics.
From what I can tell the foundation models for biology are not doing nearly as well as similar level models for tasks such as coding, writing, Q&A, etc. That’s not shocking, the problems in biology are a lot more complicated and the data isn’t sufficient for these types of models.
But I bring that up only because I’m curious how AGI would be defined and if it’s allowing the models to be essentially high school level for certain complex topics while being an expert in other things such as playing Go
Definitely some areas that are lacking and it still gets stuff wrong.
But i always think about 1) it's a very young technology. In it's toddler years. 2) This is as bad as it will ever be. It's not going to get worse.
Fascinating times ahead.
Attachments
-
IMG_1013.png
The hardware side of AI (nvidia) is doing incredibly well. The software side needs to become profitable after all of the money poured into R&D. The hardware side is helping fund the software side which then generates more hardware sales.
I honestly think this is one of those issues where you don’t know if it’s a bubble or not until it bursts. Not sure if that’s how all bubbles would be defined by an economist, as I recall there’s a serious debate about whether bubbles can exist in the first place.
I think nvidia is an incredible company.
My concern is the network of deals between various leading AI companies, and if investors actually understand where the value is for AI.
pfft no. XAI is operating at a huge loss.... so is ChatGPT. You are conflating actual hardware with AI.
Uber CEO, COO sends stark message on AI spending in 2026
Ninety-five percent of Uber's engineers use AI tools every month. Seventy percent of committed code is now AI-generated. By almost any adoption metric, the rollout has been a success. So why is the COO saying the costs are getting harder to justify? Uber (UBER) COO Andrew Macdonald made a candid ...
finance.yahoo.com
I've overseen 5 teams adopting AI into technology work flows. Being on the ground floor of adoption, its obvious to me that this tech doesn't solve one of the biggest problems. The fact that our economy is rotting from the inside. Business will not survive this ********, they are the actually too slow now.
Uber CEO, COO sends stark message on AI spending in 2026
Ninety-five percent of Uber's engineers use AI tools every month. Seventy percent of committed code is now AI-generated. By almost any adoption metric, the rollout has been a success. So why is the COO saying the costs are getting harder to justify? Uber (UBER) COO Andrew Macdonald made a candid ...
Newsflash: public growth company operates at loss. Amazon operated at a loss for what 20 years? Don’t think that has any relevance to its future success, almost all companies start off operating at a loss. SpaceX lost 12 B last year I think.
Amazon became profitable in 2001, 7 years after its founding 4 years after it went public at 18 dollar per share, 438 million valuation.