Doom scrolling and the next financial crash!

Cus

The market is wide ranging, I exited my heavy weighting of global equities to other asset classes for 2 reasons. One was that an equity market crash now I have retired would have upset me too much, and thankfully I have enough without having to take that unpleasant risk.

I still play with a small portion of my portfolio in risky assets/individual biotech for emotional fun. It's funny how I kind of want a market crash to validate my prediction/decisions even though it would be overall negative for me..

Bostonerimus1

https://forums.moneysavingexpert.com/discussion/comment/82111509#Comment_82111509

I'm trying to avoid the worst of a crash and protect some capital for the heirs. Being an indexer for the last 40 years has been very profitable, but now I believe that the major indexes are far too biased to tech and have moved towards cash and value stocks. I'm still 67% in stocks, but only 13% of my portfolio is in tech stocks. If I was younger I might just buy the global index as there should be a chance to recover, but as I'm thinking about leaving something to the next generation I'm being a bit more conservative and reducing what I see as risky components in my stock portfolio. It's an example of how old people tend to emphasize capital preservation over growth.

I'm not sure what I'm doing is exactly timing the market, more adjusting my asset allocation as long term bond rates have increased and tech has become 40% of S&P500. I haven't changed my approach to investing…it's the indexes that have changed.😉

PS here is a video by Patrick Boyle that just popped up on my feed and I think he explains things well. FYI I checked and my VVIAX value fund has 10% tech and is not dominated by Apple, Microsoft etc. although Micron is the second largest component.

https://www.youtube.com/watch?v=wTiYaWFP59Q

cloud_dog

Let me preface this by saying I haven't a clue but...

The flip side of the discussion; accepting that a crash will happen and could well be instigated by concerns re tech capital investment returns, is if this type of technology level in world markets could be the new 'norm' (until the next new 'norm')?

masonic

https://forums.moneysavingexpert.com/discussion/comment/82112464#Comment_82112464

It could well be that the market has priced these companies correctly, or even underestimated the revolutionary change that they are set to make.

In those positive cases, some money is being left on the table for someone else. That is the cost of insuring against the bear case.

Bostonerimus1

https://forums.moneysavingexpert.com/discussion/comment/82112542#Comment_82112542

IMO the LLM models have been hijacked by some pretty unscrupulous people and way over sold to investors. The return on investment required to justify all the borrowing is ludicrous, but I was finally convinced that there was a con going on when Sam Altman said that "we are close to creating a genie that can grant any wish". That's trombone and monorail salesman levels of BS. However, I'm an old retired guy and I might just be "yelling at a cloud."

masonic

https://forums.moneysavingexpert.com/discussion/comment/82112663#Comment_82112663

FWIW, I'm already finding these tools useful in ways that validate some of the claims being made. I use them regularly for finding needles of information in large research haystacks, for pair programming, for planning and hypothesis generation in scientific work, and for chewing through repetitive HR admin (performance appraisal write-ups and the like) that would otherwise eat into my day. I've also seen plenty of businesses deploying them in customer service, though that's a use case customers themselves tend to be lukewarm about at best.

So I don't doubt there's real value being created. The harder question is whether today's pricing reflects a sustainable business model, or whether we're still in a heavily cross-subsidised phase where the true cost of running these models isn't being passed on to users. And on complex tasks, you still need an expert checking the output, which limits how much labour these tools can genuinely displace right now.

The main beneficiaries of a productivity tool like this have to be the end-users, or businesses will simply carry on paying humans to do the work instead. If most of the value ends up accruing downstream to the companies deploying LLMs rather than to the companies building and scaling them, that leaves a large question mark over whether the market's confidence in the money being ploughed into scaling the underlying research is actually justified.

So even if you believe some of the claims, there are still questions to be answered on the economics.

Linton

For me the key issue for investors is not whether AI will succeed but rather who, if anyone, will make serious profits from it. History shows that early developers do not last long, perhaps because their skills and motivations are no longer appropriate. How many, if any, of the most highly valued companies of the 2000 .com boom still exist? Did they ever justify those valuations? How many of the current major players are in that list?

The next question is where will the profits be made? As I understand it the Chinese AI developers are benefitting most because their products are much cheaper than the better known US ones thanks partly to electricity being much cheaper than in the US and partly because they have focussed on the efficiency of their software. It was interesting to read that Inherent, a UK AI company set up by former DeepMind developers, are basing their work on Qwen, an AI produced by Alibaba.

EthicsGradient

https://forums.moneysavingexpert.com/discussion/comment/82112721#Comment_82112721

Amazon was part or the dotcom boom; IPO at (price adjusted for splits) $0.12 in May 1997; peaked at $5.65 during Dec 1999 , dropped to $0.28 by Sept 2001, didn't exceed that $5.65 until Oct 2009. Now over $258. The trick, as always, is finding the next Amazon without too many failures alongside it.

Amazon.com, Inc. (AMZN) stock historical prices and data – Yahoo Finance

Dead_keen

IMO the LLM models have been hijacked by some pretty unscrupulous people and way over sold to investors. The return on investment required to justify all the borrowing is ludicrous, but I was finally convinced that there was a con going on when Sam Altman said that "we are close to creating a genie that can grant any wish". That's trombone and monorail salesman levels of BS.

I certainly don't understand all of this sort of stuff. For me, generative AI (like LMs) is absolutely incredible and I think companies will pay huge amounts of money to use it.

In my naive way, I assume that:

(i) inference is very profitable for AI companies, like Anthropic,

(ii) training the next frontier model is costly but can be managed to make it cost way less than the inference profit, and

(iii) many AI companies are going to go bust because they (a) see the massive gross margin on the inference, and (b) then commit to buying so much more compute in 2028 without having the cash in the bank today to pay for it. If AI is as absolutely incredible as I think, they will have the 2028 customers chucking more than enough cash to them that they will be be able to pay for the compute that they have bought and be owning all the gold in the world. But if the genie can only grant three wishes, and not an unlimited amount, they won't get enough cash from customers to pay for the compute they have committed to buy.

Or an analogy for people with low mortgage fixed-rate mortgages about to expire:

1. The people building data centres are the banks.
2. The AI companies are the people who spend 50% of their income on paying their 0.5% pa interest-only mortgage
3. The 0.5% fixed rate period ends in 2028 but, fingers-crossed, they will get a pay rise by then.

Linton

https://forums.moneysavingexpert.com/discussion/comment/82112914#Comment_82112914

Good try, though Amazon at £1Bn was a relatively minor player with nowhere near the highest valuation. but between 2000 and 2002 it dropped by 90% and only survived because Jeff Bezos hadnt spent the money he had raised. It only seriously moved beyond being a book and CD/DVD seller around the time of the crash. So the chances of comeone predicting its current strength would be prertty small.

BT is a better example as it was seen as a tech company supplying .com infrastructure rather than a niche user. In December 1999 BT reached a market cap of £110Bn. It is now £20Bn.

Thanks to Gemini!