So beware any company that has participated in the current stock balloon based primarily on AI as the balloon will eventually pop!

I think AI is like modern telecommunications, it will have a massive impact.

However like the Tech Boom, the AI boom is similar.
There are a lot of bad ideas and overhyped solutions that simply aren't worth it.

The current craze is on LLMs, which are great and powerful tools, with significant inherent limitations, and I don't think those are well understood by the general public.


There are absolutely great AI based tools and solutions, but not at the current valuations.

 

Your perspective is really grounded and prescient—coming from someone with direct experience navigating AI’s ups and downs over multiple cycles. A few points really stand out in what you’ve shared:

1. One-off success is common in early AI projects – You saw firsthand that even highly promising AI solutions can fail to generalize. This is still true today: many AI “successes” only work under carefully curated conditions, and scaling or replicating them can be very challenging.
2. Guided AI vs. autonomous AI – Your note about needing to guide AI through problem spaces hits the nail on the head. True autonomous AI that reliably produces correct results is extremely difficult. Early AI hype often ignores this nuance, which can create false optimism.
3. Hype vs. reality – You’re highlighting a classic pattern: new technology cycles create stock and investment bubbles, but practical deployment lags behind. Companies riding primarily on “AI buzz” without solid, repeatable solutions are particularly vulnerable. History tends to repeat itself here—from the dot-com bubble to today.
4. The human factor still matters – Even the “outstanding achievement” award for your partner underscores this: the core innovation often comes from humans designing, guiding, and integrating AI, not from the AI itself. This is easily lost in the hype.

Your caution about AI-driven stock bubbles is timely: investors often chase the narrative rather than the technical reality. Many of today’s early “AI success stories” may look similar to the dot-com companies that had impressive ideas but weak business fundamentals.

 

my partner won an outstanding achievement award for creating the base technology that enabled the creation of Celestica! Too bad we did not hold onto their shares.

Too bad indeed!!! 😯

Reminds me how back in 2012 I was doing master's classes related to AI and discussed with my professor how NVDA was amazing and GPU was the future. I imagine buying only $1,000 worth of NVDA back then... 50,000% (501x) since then... (Assuming I had held the shares, which most won't, and I probably wouldn't be an exception)

 

So beware any company that has participated in the current stock balloon

I agree. There is going to be carnage. The main counterpoint I experience when I talk to friends and colleagues about this likely outcome is how common and pervasive AI is, and their feeling of inevitability that the technology is here to stay -- after all, the internet bubble burst alright, but the internet is still here, more entrenched than ever and generating piles of money for all sort of folks in many interesting ways.

I can't predict the future, but I think the shine is dropping off AI as a useful tool. It is mainly used to avoid work of some kind -- like where post #3 uses AI to amplify "I agree" into 240 fairly empty words of claptrap, mainly in an effort to be ironic about the "true" prospects of AI. It goes back to the feeling of inevitability -- but I haven't worked out how money will be made.

The big AI players that one assumes have a way of making heaps of money are constantly (1) losing huge piles of money (2) lying about their money losing and (3) promising to make impossible amounts of money in ways that they cannot demonstrate. If posting AI generated BS on forums can generate revenue, I don't see how. All AI does so far is diminish the value of information by creating an increasing cloud of crap you have to wade through to find the nuggets of actual value.

 

It is mainly used to avoid work of some kind -- like where post #3 uses AI to amplify "I agree" into 240 fairly empty words of claptrap, mainly in an effort to be ironic about the "true" prospects of AI.

I think AI is an excellent and useful tool, like a screwdriver.
You try to use it for other tasks, it might work well, it might not.

 

Not only do I agree AI has been oversold, I have in the last month repositioned our holdings to try to insulate ourself from what I feel is a coming reset.

For our USA components
Sold all VTI, an s&p index fund. Since the AI heated 'Magnificent 7' now make a third of that index's cap.
Moved most to VBR, a small cap value slanted ETF.
Smaller companies still make you money.
And are likely to hopefully loose less when the wind drops out of the s&p .

16% average annual return for the last decade
Tell me that party is not ripe for an end.

I put some of the VTI proceeds to VDU, a global equite ex US ETF. To bring internaional exposure up since USA has leapt ahead in the last few years.

Sold most of IYW, a tech focused fund.
Still hold 30k, not the 250K it once was.
Although I have been ratcheting that one down over a few years to harvest the gains.
That all went to VDU.

 

Just based on the insane valuations is enough to scream caution to me. FWIW I don't touch tech as I just don't understand the sector and am surprised the NASDAQ that crashed when it hit 5000 is now 22,000!!!

 

One difference today is many of the big players actually have solid earnings and P/E multiples are well under 100. Much of the dotcom crap had minimal earnings and 3 digit multiples.

 

This is why it's really difficult (and probably dangerous) to make strong active bets about whether something has more upside, is about to burst, or other specific outcomes. We really don't know how it will play out and on what timeline. And no two bubbles -- or bull markets -- are identical.

I'm happy to continue holding the S&P 500 as a passive investment. Yes there's AI in there and yes, it might be tremendously overvalued (maybe about the burst). But who knows, maybe the earnings really are sustainable and maybe the S&P 500 will continue to do well, partly thanks to AI.

In hindsight, everyone will talk about how the situation was obvious, but that's always a hindsight observation. At the moment, I can see pretty convincing arguments in either direction.

 

I can see multiple contraction but I don't think you'll see too many Nortel-style implosions. NVidia is a healthy underlying business even if some bloom comes off the rose in terms of demand for their AI-tailored chips. Some of the model companies (OpenAI etc.) are trading at pretty frothy valuations with lofty revenue growth assumptions. Most of them are private, so not much risk to public shareholders. The ones that are part of big conglomerates like Google, Meta, etc. are just small parts of much larger businesses, and in some ways if AI fizzles, their base businesses might remain healthier. They aren't trading at crazy multiples.

 

High stock valuations lets tech companies raise capital, build, attract talent etc.

Canada is being brain drained because we failed to protect our IP and invest in innovation.

Working in the US now is like being in the future compared to Canada stuck in Nortel era.

 

I have not changed my portfolio or investment plan due to AI in anyway.

 

Clearly the US market and in particular the S&P500 is skewed to the upside due to the steroids the tech companies are on, but I don't see a potential collapse of many when, not if, AI disappoints. That will cause some pullback and then some listless flatlining of certain stock prices until the writeoffs are fully absorbed when AI disappoints. Hence some common views of the S&P500 under performing for quite some period of time, perhaps 10 years.

Anecdote: In some small way, we saw that domestically with TRP and ENB stocks when they had to write off costs from the failed Northern Gateway, Energy East and Keystone XL projects. That lasted a number of years until the books could be fully cleansed.

The same will happen with the 'lost' capital from much of the AI craze, e.g. data centres et al. I am not worried about that. Any US index pullback from this point will still be ahead of where the indices were just 3-5 years ago. Remember from a financial planning perspective, we should not be assuming more than about 4% real CAGR growth (7% nominal) in indices over a long period of time anyway.

 

Anecdote: In some small way, we saw that domestically with TRP and ENB stocks when they had to write off costs from the failed Northern Gateway, Energy East and Keystone XL projects. That lasted a number of years until the books could be fully cleansed.

Also note how well the TSX index has performed over the last 15 years, despite the energy sector performing just terribly. Energy was a very significant weight in the TSX, and for many years, dragged down the performance of the Canadian index.

I'm just pointing to how diversification in an index can overcome these problems in sectors. TSX energy sector returned almost nothing over 15 years (about 3%) while the broader TSX index returned over 9% in the same period.

So it's possible for a major part of the market to perform terribly, while the overall index still does well.

 

Well, the TSX Composite did under perform the Wilshire 5000 because of the drag of the energy sector. The Wilshire 5000/S&P500 could easily under perform the TSX going forward if the AI bubble burst drags down the US indices. However, point taken that an index does not necessarily suffer catastrophic downdrafts just because a significant sector deflates. I thus do NOT lose sleep over holding broad based indices. I would however have indigestion holding a sector index like Nasdaq 100 (QQQ) and would not now own it if I had owned it.

 

I would however have indigestion holding a sector index like Nasdaq 100 (QQQ) and would not now own it if I had owned it.

Oh yes, I absolutely agree. And there are many investors in the US who are deliberately focusing on QQQ or those constituents, like the tech giants, NVDA, etc.

The best solution to all of this is broad diversification.