Chris and Paul discuss the current financial landscape, focusing on the optimism surrounding the Trump administration and its potential impact on the economy. They explore the differences between the current economic environment and that of 2016, emphasizing the need for investors to remain diligent and not complacent. The conversation also touches on the potential risks and opportunities in the market, including the role of AI, the state of commodities like copper and oil, and the implications of government policies on inflation and economic growth.
Trump Optimism and Market Conditions
Paul attributes much of the current market optimism to the Trump administration, drawing parallels to the post-2016 election period. However, he cautions that the economic environment is different now, with higher housing prices, interest rates, and inflation. He advises investors to remain vigilant and not become complacent, as the market conditions are not the same as they were in 2016.
Market Participation and Risks
Paul highlights a concerning trend where major indices rise, but the average stock does not participate, indicating potential underlying market weaknesses. He references historical instances where similar patterns preceded significant market declines, urging investors to be cautious and pay attention to market signals.
AI and Market Bubbles
Chris and Paul discuss the potential bubble in the AI sector, using Supermicro Computer Inc. (SMCI) as an example. They note the company’s accounting irregularities and its significant role in NVIDIA’s production, suggesting that a burst in this bubble could have broader market implications.
Key Data
SMCI accounts for 25% of NVIDIA’s production orders.
United Health Group’s stock has surged significantly since the introduction of Obamacare.
Predictions
Paul predicts that the current market optimism may not last, and reality will set in, potentially leading to market corrections.
Chris anticipates more inflation in the coming years, driven by government spending and potential economic stimuli.
Implications
Investors should be prepared for potential market corrections and remain vigilant in their investment strategies.
Inflation could continue to impact retirees and individuals on fixed incomes, necessitating careful financial planning.
Recommendations
Investors should work with financial advisors to ensure their portfolios are well-managed and aligned with their long-term goals.
Individuals should educate themselves on financial markets and seek alternative sources of information beyond mainstream media.
Consider diversifying investments to include assets like gold and silver as a hedge against inflation.
On the general subject of AI, like everything to do with IT it is hyped to high heaven in order to generate ‘buzz’ and boost revenue.
That said, it is a very useful development even if it isn’t actually anything to do with real sentience. I think it has, and will, lead to a lot of job losses as AI can produce useful ‘intellectual content’ and replace ‘thinking work’ that would previously require a human to do.
AFAIK Supermicro’s issues seem to relate to some pretty poor business practice which developed over the boom years rather than pointing to weakness in AI generally.
A lot of the tech companies think that they operate in ‘the Wild West’ with organisations pushing the limits of what they can get away with and often somehow thinking that laws and regulations don’t apply to them. It certainly worked well for the big companies that built de-facto monopolies over the last couple of decades in IT and just paid off the authorities/regulators afterwards.
That’s really weird stock chart for 2024. Supermicro sells motherboards and other server equipment but no AI cards.
SMCI has gone totally under my radar.
How is Chris opinion on investing in “low tech”, ie basic gsm phones, phone networks, internet devices instead of “high tech” like AI, top of pyramid in tech sphere… meaning incase economic crash, shortages, society simply cuts luxury spending, they go back to basics that are needed (in longer term, I dont see pressure for real lifestyle change yet).
Other question is how wasteful it is to fomo way that investors (possibly gov included) are like sheep herd driven by singular sheepdog pouring trillions to one thing, now it has been AI, while everything else in economy is left to rot underfunded. Until new trend comes along.
Civilized, robust, multiplural society should be able to do better. Afterall what use are values if money is dumb.
I saw this timely article just before I saw the headline of Chris’s video:
For years I have been warning that “scaling” — eeking out improvements in AI by adding more data and more compute, without making fundamental architectural changes — would not continue forever. In my most notorious article, in March of 2022, I argued that “deep learning is hitting a wall”. Central to the argument was that pure scaling would not solve hallucinations or abstraction; I concluded that “there are serious holes in the scaling argument.”
Basically, people are noticing that OpenAI’s recent release of their ChatGPT is only an incremental improvement from their previous version.
This leads to contrarian AI experts warning that improvements on AI has reached a plateau.
Even for a layperson like me, I can see that Large Language Models (LLM) is not the way to achieve Artificial General Intelligence (AGI).
The elephant in the room is that LLMs can’t do deductive reasoning and inferencing. For this, you need other approaches to AI such as Symbolic AI.
So, why are companies like OpenAI and Microsoft throwing vast amounts of money building data centres to hoover up large amounts of human-generated data and consuming vast amounts of electricity for processing power to train LLM? There’s even talk about using nuclear power to generate vast amounts of electricity to train LLM.
If throwing all these vast amount of resources is going to result in diminishing returns, then it means that innovations in this current trajectory of AI is going to reach a plateau.
In that case, wouldn’t that mean there’s going to be a lot of mal-investments into AI? And with that, the popping of the AI financial bubble as a lot of AI money are going to waste?
So, is the news about SMCI and LLM plateauing coincidental?
Do you see anything to worry about in the interim period if the “deep state” does a “scorched earth” retreat focused on the economy in order to hand a mess over to Trump?
Yeah this idea that “more training” will make LLMs produce the optimal results in an instant is a pipe dream. The obvious solution requires that they “think” longer (inference), which means slower processing times > more tokens > more energy consumption > higher costs for the “best” performance and results.
They’ve all been able to see this for some time, but that doesn’t at all infer that those first results aren’t already more than good enough in most cases. Claude-3.5 sonnet is pretty mind blowing to me, and it’ll only get more polished with time.
Of course, none of this precludes the massive data center and energy boom for big tech+big gov
What percentage of SMCI is mother boards. What percentage of people who buy mother boards use built in video. What share of the separate high end video card market does Super Micro have.
What I see is that the average person has little need for a laptop or desktop when not at work. Most rely almost exclusively on their cell phones for Internet related tasks.
Those who do bother with a desktop tend to do so for activities that demand ( at least in their minds ) a full strength video option.
A lot of folks use tablets when not using their phones and a large percentage of these use Chinese cell phone chipsets, which offer alot of bang for the buck, if all you are doing is interfacing with webpages, email and office suite documents.
They’re in the business of providing the hardware that competes with Dell / HP as a whitelabel OEM. Also known for being kind of the king in enterprise AMD systems. This sort of puts things into context:
Exactly why “compute” and data storage has transitioned away from end user systems to cloud and SaaS services that are driven by enterprise-class servers (increasing their demand).
Side note, that little hump right before the plummet is so gross.
And no one is asking me to delete all my selfies and cat videos. I’m supposed to take shorter showers and turn down/up my thermostat to save a smidge of electricity. If man made climate change was really a thing they would realize all those selfies stored on the cloud with redundancy are chewing up a lot of electricity.
I was at the amusement park and got to chatting with the girl whose job it was to prevent any ride photo with obscene gestures from being shown on the overhead screens. So I asked, seems they have all the ride photos saved all the way back to the beginning of the year. And those are the ones she could click a button and access.
There’s even talk about using nuclear power to generate vast amounts of electricity to train LLM.
