Weekend reading: A sausage-fest of a market
≡ Menu
The Latest Articles
Essential Reading
Better Investing
Property Posts
Get Money Motivated
Monevator is a place for my thoughts on money and investing. Read our disclaimer. Subscribe to get all our free posts via email. Follow us on Twitter, Facebook, or via RSS. You can also send us a message.
Copyright © 2007-2026 Monevator. All rights reserved.
Disclaimer: When investing, your capital is at risk and you may get back less than invested. Past performance doesn’t guarantee future results. All content is for informational purposes only. I make no representations as to the accuracy, completeness, suitability or validity of any information on this site and will not be liable for any errors or omissions or any damages arising from its display or use. Full disclaimer and privacy policy. This site uses cookies and features affiliate links.
Comments on this entry are closed.
I encourage anyone who is thinking about offering themselves for a fireside chat to do so. I have learned a lot from reading them, and even more from being the subject of one.
I was a little nervous, as I was not sure my FIRE journey had anything that was original, interesting or useful, but readers were incredibly supportive (at least in the comments). No one said “you idiot – why didn’t you…”.
I have long used the aphorism – if you want to understand something, try explaining it to someone else. The process of answering the questions forced me to examine and explain my journey. At the very least, it surfaced some of my assumptions and hidden beliefs about family, responsibilities and money, and allowed me to better understand my goals and fears.
Then the comments from other Monevator types were extremely useful. Not only supportive, but making specific points that caused me to check some facts and assumptions, and to deal with some lurking issues that had been in the ‘too uncomfortable to confront’ box. As a result, I have both a better understanding of what I am trying to achieve and a different plan about how I expect to get there.
All in all, well worth the time and the (slightly embarrassing) self-reflection.
I agree with old_eyes it’s a very worthwhile endeavour, and does add richness to readers/contributors to know a little more about each others back story.
I went for double anonymity, but as it turns out it wasn’t necessary
Boltt (aka Fat Fire John)
Ps it’s been 3+ years for some and may be worth a follow up
@Boltt #2
We have had a couple of follow ups I think. I guess there has to be a significant change in outcome or strategy to make a revisit worthwhile. “Still buggering on” is likely to be less engaging than “Aargh! It all went tits up because of unforeseen factor X, and I have had to make urgent course corrections”
I would be happy to be a subject – not sure that I qualify though as I am not quite there yet, although all of the heavy lifting has been done and FIRE according to 4% rule is likely a short distance ahead. To the extent people might find it interesting, I would say the main things of note for me are: being young-ish (early 40s now), raised in a family that does not really think about money or saving, being a tax adviser for a job meaning I am obsessed about tax efficiency (sacrificing very high % of salary) and that is giving me looming issues around how to bridge the gap between potential FIRE and being able to access pensions, actually liking my job and so being unsure whether I even want to pack it in, and finding an investing strategy that seeks to optimise for safe withdrawal rate (bit like Finimus’s article re his long term family portfolio). So, now I look at it, bit boring emotionally but quite in depth in terms of thinking through the mechanics of FIRE for my situation. So maybe not that interesting at all but I have got this far in typing the comment now…
@Mike #4
Sounds like an excellent fireside chat. A different approach and set of issues to others I remember. It is the diversity that makes the series so valuable.
I think you’ve summed it up quite nicely this weekend, TI. It has indeed been a wild ride within the index over the last year or so. The churn has been relentless. Personally, I tackle this by splitting my equity exposure roughly 50:50 between index funds and active antics. I know I should probably just go completely passive and spend my time doing something more productive…but I can’t help myself. Even my passive half is not entirely “passive”, since I have overweighted small cap and EM index exposure. @TI – I am very curious, are you all-in on an active strategy, holding zero index funds?
@Mike You should go for it, it sounds really interesting, I’d particularly like to hear more about this bit: “I am obsessed about tax efficiency (sacrificing very high % of salary) and that is giving me looming issues around how to bridge the gap between potential FIRE and being able to access pensions…”
Pondering the same thing myself. Not that I qualify for the E part of FIRE as 55 is about the earliest possible for me (partly linked to the above issue – my pension access age is 57 but I don’t have enough outside my SIPP for more than a couple of years’ spending).
In other news, there is a (to my mind) terrible article in the FT today about gilt ladders. The author is obsessed with the tax advantages of low coupon gilts and seems very confused about what gilt ladders are for, the trade-offs etc. Doesn’t mention linkers once. I was going to mention @TA’s tour-de-force article in a comment but it’s members-only so not much point. On the other hand @LateGenXer’s tool is getting plenty of love in the comments.
I was invited to FSC in 2023. I could do it but I think it would be relatively dull.
Allocation now: 72% equity tracker, 26% bond tracker, 2% cash
With my investment gains recently far outstripping my former salary I’m trying to get rid of more money. In contrast with former colleagues I’ve spoken to I have come out ahead partly by not being divorced or alcoholic.
I too can highly recommend the process . I enjoyed it immensely and the comments were far more positive than I thought they were going to be.
Given the high valuations mentioned what as a passive investor are we supposed to do about it? Any tweaks feel like market timing in all but name?
I’m still 100% equities age 45 and beginning to wonder when and how (or even if) to take some risk of the table (closing in on £1.2m invested between isas and pensions) . Sadly mostly pensions. In a job I absolutely loathe so at the moment I am going gung ho for the exit as quickly as possible rightly or wrongly
On the topic of AI, I encourage everyone to read the work of Ed Zitron. His analysis of recent results is very revealing: https://www.wheresyoured.at/am-i-meant-to-be-impressed/
@fatbritabroad
Present high valuations do present a quandary … I have tried tactical asset allocation in the past , not sure it really helped that much overall, you win some and lose some, markets move unexpectedly and quickly. It feels good to do something but perhaps not too much ….
I recollect watching CNBC whilst overseas in late 1999, crazy times) I was well aware that valuations were crazy and had positioned my portfolio towards Investment Trusts, in particular Equity Income and the more sensible generalists like Foreign and Colonial.
When the tech fall came it helped a bit, but markets still fell heavily and stayed that way over the next 2 or 3 years.
What I felt was very helpful is that I kept investing every month through thick and thin.
I ignored the high markets and the falls , I invested all I could afford and it worked out very well. The discipline helped a lot , we can see high valuations now but whether it ends with a fall or just muddles along ?
If you are investing for the future, without immediate need for that money for some time then I would carry on.
If you feel you have to do something then allow yourself a % to play with. (I’m still doing that now, on balance I’m not sure it really helps financially….)
Yes that’s the thing alright. I have the stomach for falls ok and will probably carry on working in some capacity just not in my current role so probably less money. So I won’t rely on my portfolio for a while
As you say very difficult to plan the market. Case in point (and very privileged i acknowledge) I’ve finally convinced my father (who doesn’t really have the temperament to invest and is mostly cash) to act as a margin loan for me and lend me a decent sum (about 15% of my none pension assets) for a minimum of 5 years on which I pay him 5% fixed interest per year or pro rata thereof . Anything above that I keep
He finally agreed a few weeks ago (just missing the start of the latest rally) but commented that perhaps I should wait as the Iran war made investing the lump sum risky .
We’re now up 6.5% in 2 weeks . Had he given it me 10 days earlier we’d have been up 16%! Who’d have thought that with a war going on.
@TI re “envying easy life of indexing”: both very true and untrue at once. Changes in flows change prices. When flows are uniformed, as with passive (plain vanilla indexing = give me money I buy at cap weight allocation at any price, withdraw money I sell apportioned to cap weight at any price) then you run into problems beyond a certain market share. This is why just 100% plain vanilla indexing is not the cost and consequences free pancea (to Bessembinder’s return skew) it’s sometimes presented as, see this comment from Paul Kedrosky on the effects of the SpaceX IPO:
“But people aren’t focused on the right things. That much new equity supply hitting in a few months creates a math problem: the money has to come from somewhere. Most of it will come from existing holdings. Passive funds will be forced buyers once these names join the indexes, which will happen much faster than usual, given recent index rule changes. That means mechanical selling pressure on whatever many funds currently own, which is mostly the same large-cap tech stocks everyone else owns.”
Re @Hariseldon, Fatbritabroad #9 and #10: I’m putting my neck out here, and with about £700k each in both ISA & SIPP (and >£100k in GIA) have ‘freed up’ £100k (i.e. ~7%) from equities into cash equivalents (CSH2 MMF like ETF) as a form of (interest paying) ‘insurance’ for a possible near term fall risk. Yes, shock horror market timing ( 😉 ).
If the market doesn’t decline by 15% or more by 1st December 2026 then the thesis is falsified and 80% of the cash gets reinvested into VWRP/VWRL ETF (or Winton Trend Enhanced Global Equity fund, GBP Hedged, for a capital efficient / return stacked option).
If there is a decline of more than 15% then that £80,000 likewise gets reinvested into equities (or into equities and TF with Winton) on first trading day on or after 1st December, but maybe (if just equities) then with a big tilt towards whatever (sector wise) has sold off most.
The remaining £20,000 I’m putting into LatAm because 1) commodities (still underweight compared to my target allocation) and 2) the Southern Cone is cheap (Brazil 10-11 x PE) and under owned (LatAm just 0.8% in some global equity trackers compared to over $7 tn combined LatAm GDP in circa $126 tn Global GDP; i.e. 5% of Global nominal output).
@Fatbritabroad (#9)
Historical modelling of accumulation has suggested that for some countries (e.g., UK and US) holding 100% equities either throughout accumulation or until no more than 5 years before retirement ended up with the highest pot in the vast majority of periods, whereas it was less clear cut for other countries (e.g., Germany).
You could consider ‘derisking with a purpose’ by using bonds to build guaranteed* income streams in retirement. For example, at 45yo, an inflation linked gilt ladder to provide an income equivalent to the state pension (i.e., £12.5k) from retirement at 57yo until 67yo would currently cost about £92k, while a 40 year ladder to provide £12.5k inflation protected income to 97yo would cost about £280k.
* in the absence of UK debt default
> cease to inspire respect in proportion as we know how they are made.
Last time the guts of the hitherto opaque sausage opened up, what was in there? Ah, Mortgage backed securities…
What could it be this time? Private equity loans, any other submerged sharks swimming that we don’t yet know about? At least AI is a good honest ‘it’s all different this time’ mania so it’s sort of above the fold…
Can confirm Alan S post
I was 100% equities in accumulation but went to a very conservative 30/70 asset allocation on retirement 23 years ago at 57 years of age
Currently nearer 38/62
3 index funds only
For me a successful retirement (maintaining a constant 3,3%-3,8% withdrawal rate) has been achieved -so far!
xxd09
@all @boltt — Thanks to all who have put themselves forward, either here or the many more over email/ the contact form. I’ll try to get back to everyone with a reply next week.
@old_eyes — Thanks for the words of encouragement to other potential interviewees. Your comments sum up very well what we’re trying to do with these pieces. I’m glad it wasn’t too harrowing! 😉
@mike — As has already been said, there’s a lot in there in terms of mechanics that could well be interesting, especially combined with your relatively young age. (Congrats!) I’ll hopefully drop you a line next week.
@Mirror Man — I’ve been thinking for the past couple of years now that perhaps I could go 50/50 passive/active, or perhaps do a cruder ‘bucket’ type approach and dedicated 1/2 platforms to active (say the cheapest to trade, such as my zero commission brokers) and run the others passive. At least until this quality drought / AI head-‘vigorous love making’ phase passes. Currently my index allocation is very low, but remember I’ve very active, and it has been getting towards 50% now and then over the past 15 years. Also remember I run a *lot* of holdings typically, with a bit of scalping etc in the mix (don’t do this at home kids) so my general ‘beta’ return can be quite close to tracking a global index fund if I don’t have a particular tilt/posture I’m trying to express. (Which I’m not at the moment, because as I’ve belaboured in posts and here I’m really not sure how this AI thing / the expensive US market will shakedown, and I don’t want to bet overly in one way or the other or, indeed, the other 😉 Hence I could index and have an easy life (relatively, @DH 😉 ) for a few years.)
@c-strong — Please do mention the article in the comments, I think FT readers can find the few pennies required to pay up and access and we have no marketing budget! 😉
I read the same article too and thought it was pretty reasonable, though it’d needn’t have said ‘not in an ISA/SIPP’ (clearly you can do it in and ISA or SIPP, you’re just not getting the tax advantage). It’s clear things have moved on from when these were fringe discussion in 2022. That said I might do something on using gilts to pay lump sums in the future, as that hasn’t been covered much for the past 10 years and I don’t think has come back into people’s thinking in the same way that ladders have yet.
@BBB — Well depending on how much you could wax lyrical, the spending more money challenge could be an interesting focus. By all accounts it’s a huge issue for a lot of lifelong accumulators.
@Fatblokeabroad — Cheers for the words of encouragement re: the FIRE-side chats! Re: Tweaks, I’m sure you saw TA’s member article from last October for passive investors:
https://monevator.com/what-to-do-about-extreme-us-market-valuations/
@ArtArcturus — Agreed, he’s a valuable voice and indeed I’ve linked to him at least half a dozen times in these links. With that said, he was sort of losing me a bit by not changing his stance much even as it’s clear AI use/traffic is booming. I saw a stat recently that something like 50% (I think, this is from memory) of the forward orders for the hyperscalers is from AI use. Clearly that’s a risk factor (to his profitability thesis) but not to his general use thesis, since it’s clear people are using it, for good or ill.
@HariSeldon — Wise words. I’ve positioned myself for falls then they’ve happened and the result has usually been pretty emotionally negligible (“I’m down 15%, oh the market is down 20% big whoop”) or some small cap / speculative growth stock I could bear to trim falls 80% and undoes most of my tactical positioning. Of course this could be because I’m not aggressively going into cash/bonds or whatnot, but then we’re really into market timing where my results are decidedly mixed (I feel I have a bit more of a knack than most, but we’re talking a slightly weighted coin and it could be luck edge, not an edge-edge!) Probably the conventional wisdom is right for this reason. Set your allocations that you can ride out and rebalance but otherwise get on with it. At least you have some rules, rather than a lot of questions and second guessing every day.
@DH — I mean psychologically easier. You might say you have some concerns about market levels or the introduction of SpaceX as a 3% weighting or whatnot as a passive investor, but I am pretty confident having 30-60 positions most of which you consider tradable at any moment in time, for a wider universe of at least 200 potential companies on your watch list, plus a flow of new ideas, plus idiosyncratic plays (temporarily distressed preference shares or whatnot) all marked-to-market second-by-second is an order of magnitude more stressful. 😉
That isn’t a humblebrag by the way. It’s closer to a confession! 😉 Not that it will news to long-term readers, there’s a reason @TA writes the passive articles around here. 🙂
@AlanS — Hard to imagine such a scenario for 100% UK investors after the past 10-15 years (though not so much the past 18 months…)
@ermine — I don’t know about private equity loans, they’ve had a long time to explode now. AI is an issue because there’s a binary element to it (in terms of certain aspects of valuation, not how it plays out). Commercial property was flagged as a big concern a couple of years ago but that seems to have gone away. Perhaps residential property if mortgage rates moved and stayed at/above 6-8% for an extended period for some reason, from a cashflow perspective. A lot of UK homeowners aren’t mortgaged anymore but there’s that still large BTL constituency… Can’t see that happening with government debt where it is though. (i.e. The government would surely choose inflation instead).
@xxd09 — You’re not 100% UK equities though, are you? I always presumed you were a global index fund man. 🙂
@all — Cheers for the comments. Very pleasant behind this paywall, eh? 😉 Have a great Sunday!
Investor-possibly slightly misread Alan S,s comment
I am a U.K. domiciled investor but am a global indexer indeed with U.K. and US equities plus other equities of course
Looking back I think only a US domiciled investor could have done satisfactorily well with US only index fund-would a U.K. domiciled investor managed a satisfactory outcome with a U.K. only index fund-I always assumed not but I stand to be corrected
xxd09
xxd09
@TI (#18), @xxd09 (#19)
Just to (hopefully) clarify my earlier comment:
For a US investor holding US equities and bonds, the highest accumulated portfolio value was found most often (~80-90% of 40 year accumulation periods – it depends on bond duration) when holding 100% equities.
For a UK investor holding UK equities and bonds, the highest accumulated portfolio value was found most often (90-97%) when holding 100% equities.
So, the results for those two countries suggest that holding a large fraction of equities in accumulation was the best solution. However, for Germany this was less clear cut since only 56% of historical accumulation periods for which 100% equities were held ended up with the largest portfolio. In other words, the evidence for 100% equities in accumulation is not as quite as unambiguous as some might have it.
I have not modelled whether the accumulated portfolio was higher with 100% UK equities or 100% world, although I note that since 1970, UK equities have outperformed MSCI world, but not more recently.
Thing is, US large cap equity markets trend up. Looking at the proportion of trading days for the S&P 500 since 1957, the S&P 90 from 1928-57 and the DJIA (or US indices covered Cowles Commission securities price daily data) from 1871-1928; the proportion of time the market spends above its 200 day Simple Moving Average remained remarkably consistent, oscillating between 65% and 72%. The proportion of trading days spent relative to a 52 week high was 77% to 88% for within 20% and, for the modern S&P 500 era (1957 – 2026), the index has spent approximately 66% of all trading days within 10% of its trailing 52 week closing high.
That’s pretty damn good, IMHO.
Remember, since 1871 there’s been frequent “Panics” (1873, 1893, 1907), two massive global conflagrations (including the use of nuclear weapons) over 1914-45, with an epic Depression in the middle, nasty stagflation from 1973-79, numerous speculative bubbles, crashes, recessions, lesser wars and tribulations, including 2000-02, 2007-09 and 2020.
Yet the US main index has still spent the vast majority of its life in a long term uptrend.
Maybe this is just the flabby reassurance of a late stage bull market participant here, but even as I hold 7% cash for now, I have to admit that the case for the US is strong and also that today’s American indices just aren’t comparable to those pre 2000.
US GAAP accounting profits, including treatment of Stock Based Compensation and IP assets, is waaaay more conservative nowadays (on an apples to apples basis, today’s S&P 500 PE of ~29x would be ~23x using the immediate pre 2000 accounting standards).
Right until the ongoing AI Capex boom since 2023 (which, you never know, might even actually pay off, albeit that it’s a very high bar indeed with some $3 tn up to $8 tn of data center related expenditures projected out to 2028 to 2030); the FCF rich and Capex + Opex light model of the FAANG, and then if the Mag 7, put all previous market darlings like the Nifty 50 or Dot.com boomers firmly in the shade.
And that’s before even considering the effects of passive flow on valuations going forward.
So, even though I’m holding some readies for optionality in the event of a crash, if there’s actually one, then I think that the only rational course for the US big tech names would be to then BTFD.
@DH #13
> Yes, shock horror market timing
Attaboy. It gets a terrible press and everybody tells you it can’t be done. TI himself seems to have managed in the GFC and I benefited from that insight, and my own a couple of times later, because I was a johnny-come-lately with, on reflection, not enough to bridge the gap to normal retirement. Passive will get you there at typical average equity returns if you have 30 years of accumulation. I didn’t have that long. It could have been all luck, of course, a sample of three data points out of a fifteen-year investing career is anecdote, not data, but it made me much more than my DB pension capital from 30 years of working.
I am much more conservatively placed now, because I don’t need to hit it out of the park, but you go DH. I laud your conviction in embracing the heresy, passive is not the only fruit. And at a shade over 10% of your working capital, even if you lose out the putative benefits of the AI/Space X boom to the passive crew’s stellar performance tracking the index and you were to crash and burn, you can survive to fight another day, possibly chastened that the theorising doesn’t work in practice. I used to consider that I overthought things at times but I surrender that crown to you with alacrity 😉
I got better results doing much less overthinking, but yes, it’s the big swings that seemed to count. As WB said,
You go DH. And the very best of British luck to you across the dark times that lie ahead. Washtubs, not teaspoons…
@ermine @delta_hedge — Well while ‘don’t try marked timing’ is clearly close to a house view around here, obviously as a super active investor I recuse myself from time to time. So in principal no argument.
In practice, it’s really *really* hard. For every one you get right (ish) (see my Moguls posts about UK stocks finally being a buy two years ago) you will get at least some wrong (see our posts about finally admitting the US market looked expensive from a year ago, which was after literally a decade of comments here and elsewhere saying it was ‘obviously’ expensive, and yet even leaving our verdict that late it’s still been flying up!)
Or similarly on the subject of AI, bubble fears and whatnot. Indeed @DH has run the gamut of opinions on exactly that! Which to be clear is not a criticism. If you haven’t been whipsawed by the advances in AI and by AI-related equities then you haven’t been paying attention… But it is another clue as to how hard market timing is in practice.
Hence I’d always suggest incrementalism and tilts to any market timers, if anyone does go down that path. I did okay in the GFC it’s true, and I didn’t disgrace myself in the Covid crash either. But I honestly can’t be sure whether I would have done better just to sit through it all even as an active investor (i.e. do my stock picking and tactical stuff, but don’t have *any* view on market expensiveness or otherwise). And even if I did/didn’t, we only live an n of one etc. The GFC could have been the Great Depression and 90% falls and then where would we have been. 🙂
I think ‘market timing’ when there’s a crash is just putting more money in isn’t it ? I suppose TI you may well have also increased your % equity exposure as well?
I did the same in covid (sadly the crash was short lived ) and again when there was that prolonged less severe downturn in…was it 2022?
As you say ermine bathtubs not teaspoons. It’s taken Me 3 such crashes to feel comfortable mind you but I did it anyway!
@Fatbritabroad — Well that rather depends on why you have money to invest in the bear market, doesn’t it? 🙂 (i.e. Prior actions…)
@investor or whether you even do! (I dont I keep everything bar my emergency fund invested )
There’s no question that repeated non systemically executed / non rules based market timing is a bad idea. You might get it right a few times on either the out or back into the market limb of the round trip, but, by God, it is, as @TI rightly says, very hard. It would be rare to get both limbs right and as unusual as hens’ teeth to do so on consecutive instances.
In my inner eye I can already see myself making a mistake in holding a little cash (hence my second thoughts, of sorts, in the face of Mr Market’s melt up momentum of the moment, per my comment #21 above; given the overwhelming odds (as outlined in @Alan S’ insightful comment #20 above) in favour of a 100% equity allocation during accumulation, at least in countries not loosing a very major (existentially challenging) war, like Germany (or getting occupied in the same).
But I have to weigh that against the regret of having no dry powder *if* there should be a meaningful nearish term pull back (or worse).
So, whilst it sticks in my gut to temporarily hold (just a little bit of) a long term loser in the form of cash, I’m still giving it a go until the effects of Musk’s SpaceX IPO madness and/or the slow burn Strait of Hormuz situation plays out.
On LLMs, I’m continually blown away with the performance of paid models (GPT 5.5 Thinking) and simultaneously shocked how far the ever improving frontier still is from genuine AI, let alone AGI. It’s like a receeding horizon where, even as we accelerate ever faster towards it, it just moves away from us with even greater haste. Maybe it will be the new nuclear fusion. Always improving and just 30 years in the future, but perhaps it always will be 30 years in the future…Still blooming useful though even if it ain’t ‘real’ AI per se.
Yes please #4 Mike.
Some of the best info has come from IFAs and tax advisors on here. Including the ‘I have no idea either’ type of info!
Re: gilt ladders and the FT Money article. Yes please, an blog post here would be much appreciated. Gilt ladder for school fees is very different to gilt ladder for decumulation or alternative to bonds/equity diversifier. I want to lock up cash and have maximised Premium Bonds and don’t want too much in fixed savings accounts. Gilt ladder seems a good option but can’t get head round it.
@all — Verdad makes the case quantitatively that the chip stuff and much of the associated AI story is a bubble, however real the underlying paradigm shift:
https://mailchi.mp/verdadcap/priced-for-perfection?e=d672c968ec
One caveat I’d note is that while I’ve been banging on for 18 months that AI infrastructure ‘investment’ is a funny sort of investment when the chips are landfill (/ethical recycling fodder) within 3-5 years, this does at least underwrite future demand for the chip manufacturers in a way that perhaps some of the previous chip cycles haven’t, or at least not the same degree.
Still, I can’t chase these things here (though I’m pleased to have a fairly longstanding ASML holding to fondle and mumble “my precious” over… 😉 )
@The Hare – https://monevator.com/index-linked-gilt-ladder/
Related: https://monevator.com/duration-matching/
@TI #29: surely the single biggest issue is not the Capex and depreciation cycle, nor the physical infrastructure bottlenecks and/or the (lack of) model maker moats; but rather whether there’s a near future durable and massive enough enterprise market for pay by token usage LLMs? (BTW, that’s my own bit of antithetical parallelism there, not ML’s 😉 )
Everybody knows that the numbers right now don’t remotely add up.
The question is whether they can be made to do so before the cash burn at the sharp end (by OAI, Anthropic and xAI) utterly incinerates all involved, risking the Remaining Performance Obligations booked into the accounts of MSTF, AMZ, ORCL and GOOG etc going unmet.
Ed (yesterday) thinks not:
https://www.wheresyoured.at/ai-is-too-expensive/
I’m thinking that the free LLM ‘product’ is so clearly enshitified already that, for any half serious use, it’s a case of having to pony up £20 to £300 pcm for ‘Pro’, ‘Max’ and Enterprise level access to frontier models, or just don’t bother with LLMs at all.
Today I asked free Claude (4.6 Flash), Gemini (3.5 Flash Extended), Grok Fast, Chat GPT (4.5 I think, they’re not transparent about what you get to use on the free tier) and Perplexity (ditto) this factual query, investing related prompt, and got wholly incompatible answers from each of them:
“Estimate the average annualised dividend inclusive CAGR from January 1995 to date for an investment strategy executed as follows: 1). On the first trading day of the first calendar year of the strategy buy the current largest market capitalisation stock in the S&P 500; 2). Hold for 12 months; 3). Then, on the first trading day of the second calendar year of the strategy, check if the identity of the largest market capitalisation stock in the index has changed. If not, then hold for a further 12 months. If it has changed, then immediately sell the existing held stock and buy the new largest market capitalisation stock, and then hold it for 12 months. 4). Repeat process each year thereafter to January 2026.”
For reference, the strategy dividend inclusive period average annualised CAGR given in the answers for January 1995 to date ranged from 8.7% (Gemini) to 14.7% (Claude). Claude’s answer was the most detailed, sourced and convincing; but that doesn’t necessarily correlate with it being (and certainly, by and of itself, does not cause it to become) right.
@Delta Hedge — Indeed. When I see results like that (or other articles / my own observations) I do wonder whether we’re being gaslit by an entire industry. I mean Demis was talking about the foothills of the singularity at the Google summit today. Yet these tools, while undoubtedly useful especially for coding, are still delivering the sort of precision-incompatible results you outline which would seem to make that singularity entirely unlikely on the back of this tech, unless we redefine Kurzweil’s original vision to be a self-perpetuating slop machine.
@TI #32: “the Stone Age didn’t end because we ran out of stones”* so how will the ‘Slop Age’ (2022-?) end?
A new paradigm, with Marcus’ neuro symbolic hybrid replacing both deterministic CPUs and logic gate ‘classical’ compute of the pre neural net era and the stochastic slop of the LLMs?
Some breakthrough in ‘grokking’ where LLMs suddenly generalise and stop being fake AI (narrow ML)?
Or maybe something with quantum computing?
Or does it all go wrong, badly wrong? And, after all, it very well might end badly here.
Several metronomes seem to be ticking down in synchronisation:
– Loss/erosion of narrative momentum around AGI/ASI/the Singularity. Like Communism after Hungary and Khrushchev’s (not so) Secret Speech in 1956, people just stop believing. When credulity fails, then securities’ prices fall. Noone pays a premium for a busted storyline.
– Increasing bizarre circular financing and risky borrowing. Capex is not all being financed purely out of FCF anymore.
– Whilst there will be a work around, and this isn’t 1973/4 or 1979/80 (it takes just 51 US gallons now to generate a $1k of US GDP compared to 131 gallons in 1973, in inflation adjusted terms), the SoH is still very much closed, and this is an inflation impulse. With fiscal dominance setting in both here in Europe and in the US, unless there’s going to be full fiscal repression, rates will rise and, as your linked to Verdad piece hints at, when the perceived payoff from so called ‘AI’ lies so far in the future, any increase to discount rates badly hurts DCF valuations for ‘unprofitable for now’ type growth tech.
– Something may go wrong with passive flow sooner or later. This is the unspoken elephant in the room. Latest on Green’s prognostications here:
https://open.substack.com/pub/tscsw/p/cocaine-on-the-tables
As Big Tech is so much of the index now (Nvidia $5 tn out of $58 tn in the SPY), a blow out in the index is going to be brutal for the tech names. Remember FB with the 78% draw down (IIRC) to $90 in 2022, off of the Metaverse fiasco?
On the other hand, headline earnings are up by lots, the forward PEs of some of the biggest names are actually compressing as their valuations increase (NVDA), and the US ain’t in a recession yet.
But what happens if the music stops (or just slows, as it was put in that boardroom scene in Margin Call) on the Capex front.
Credit markets frozen in 2007-9 in a way not though possible beforehand. A slowing of AI Capex could have all sorts of unpleasant second order effects
*NYT columnist Thomas L. Friedman suggested in 2000 this phrase was first used in 1970 by Ahmed Zaki Yamani, KSA Minister of Oil.
@DH — On the other hand, OpenAI just dropped this mathematical proof:
https://openai.com/index/model-disproves-discrete-geometry-conjecture/
According to mathematician Tim Gowers, as quoted by OpenAI:
So much of this is irresistible force meets immovable object.
It’s going to be a very jagged edge to progress, adoption, adaptation and change.
Less Asimov and his Laws of Robotics, and more William Gibson, John Brunner, PKD and, I fear, ‘Children of Men’ type vibes.
If only this tech could be kept under close control and supervision until we understand, with full confidence, what it can and can’t do; what it does and doesn’t mean; what it risks and what if offers; whether it is safe and aligned, or not; whether it will change the world for better, or for worse.
This is not like the Manhattan Project or an arms race. It’s more like man learning to speak and the emergence of symbolic language.
If it fails, then maybe it takes down the world economy (with over investment). If it succeeds, it could still take down the economy with mass un (and under) employment.
It might kills us all, or just drown us instead in a sea of slop (or perhaps that just precedes the former?)
As Gregory of Tours puts it (in the opening to his History of the Franks) “A great many things keep happening, some good, some bad”.
I’ve located that Green, Krishnan and Sturm 18 March 2026 paper (A Model for Passive That Breaks the Market), as referenced in the TSCS Substack link under my #33 above.
Here’s the SSRN link to it:
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=6438678
Per the abstract: “we will present a plausible model for the dollar size of the US equity market that incorporates passive share. Our model relies upon a small set of assumptions that are nearly universally accepted, or at a minimum widely used by quant practitioners. While our assumptions are seemingly innocuous and uncontroversial, the implications of our model should be a cause for significant concern. That is the main point we wish to emphasise in our paper. Once the passive share reaches around 65%, index volatility may increase sharply. At 90% share, an increase in volatility at cubic speed is nearly inevitable, leading to exaggerated boom and bust cycles.”
Not sure if this or an ‘AI’ bust is a bigger threat to markets; but they both have me worried.