What caught my eye this week.
Ever wonder what happened to that great little money or travel or cookery blog you read for years?
Yeah, it closed down – or at least its creators stopped updating it.
Why would they? Nobody visits anymore.
But maybe you didn’t notice because Google no longer takes you there, anyway.
Google search results are now dominated by big corporates who create content that’s made more to appeal to Google’s search engine than readers, and who can afford to pay SEO experts to keep up with Google’s endless rule changes.
To some extent Google’s apparent capriciousness was well-intentioned, as it’s fought to outrun spam sites and AI slop, and other gaming of the system.
But the end result is that the death of the independent web continues apace, with the damage done by waves of Google algorithm changes now redoubled with a kick in the balls from AI chatbots and AI search.
You don’t have to take my word for it:
- ‘AI is killing the web‘ warned The Economist earlier this month.
- Venture capitalist Tom Tonguz shared data showing how OpenAI already gets about a quarter of the queries Google gets every day.
- In response, Google is including more AI summaries in search results. And people tend not to follow these links to see where the data comes from.
It’s all an existential threat for anyone who creates the content that AI outfits train their models on, which they then regurgitate to their users freely instead of directing back to the original source.
We’re all in it together
Fortunately I ignored Google – and some well-meaning people who worked there – when they assured me that small publishers had nothing to worry about.
I realised the future of Monevator must be as an email newsletter – where we can sidestep search and the web altogether – and with our best posts reserved for those who pay to support our ongoing efforts.
After 15 years of sharing our knowledge for free it was a hard decision to make.
But honestly, introducing paywalled Monevator content two years ago has kept the lights on at Monevator Towers.
Ever more of you have become Monevator members. We’re grateful to every one.
Our member-only content archives are getting pretty well-stocked, too:
- Our Mavens articles (for all our members)
- Our Moguls articles (for active investors – or simply our biggest supporters!)
Only a small percentage of members ever cancel (‘churn’, in the industry lingo) which is also heartening.
So if you’re not a member, please do sign up. The price is still the same as two years ago. Annual Mavens membership is a steal!
A greater share of our content will probably have to go behind the paywall in the years ahead – perhaps including Weekend Reading.1
I don’t like it either. But I don’t run a tech giant and I didn’t make up the rules.
How to get Monevator membership for free
On the grounds that some of you must like what we do, I’m introducing a referral program.
This way you can tell your nearest and dearest, and bag yourself an ongoing discount as a small thank you from us for your troubles.
To benefit, you must be on an annual membership plan. Not a monthly plan.
Most of you are already annual members. But if you’re not then please switch before referring anyone, as it won’t count otherwise. (Annual is cheaper, too!)
- See the membership FAQ for how to change member plans
Already an annual member? Great, then by referring friends and family who would enjoy Monevator to sign-up to our annual memberships via your personal link, you can get a recurring discount on your own membership as follows:

Yep – find ten lucky people who become Monevator members on annual plans and you’ll get your own membership plan for free, forever!
- You will find your unique referral code to share via your membership account settings. See the membership FAQ for more details
I stress again, both you and your referees must be annual members for the referral to count.
The software handles referrals automatically and I can’t change things later, so please do take note of this.
Monevator versus Skynet
Obviously I hope these referrals will get us a few more members. But I also see the discounts as a tiny way to thank our most loyal readers.
I know it will take a long time for most people to get to ten referrals, if ever. (Though if you run a website – or have a lot of money-savvy workmates – who knows?)
But I love the idea of our top supporters getting even just a few quid off.
So again, please see the membership FAQ to learn where to find your unique referral code, and give it a go.
And thanks again for supporting us fleshy and bloody humans over the robots!
P.S. I forgot to include the link to my new property newsletter Propegator last week. Not very good at this promo malarkey, am I? Anyway it’s just a fun hobby project, but if you live in London you might like it so please do take a look.
From Monevator
Gold miners: do they improve your portfolio? [Members] – Monevator
Going without versus doing without – Monevator
From the archive-ator: How to invest in sectors, themes, and megatrends – Monevator
News
Average house price registers steepest drop for 20 years – Rightmove
UK vehicle making hits lowest level since 1953 – BBC
Rachel Reeves’ capital gains tax changes backfire as receipts fall – MoneyWeek
Government to conduct review into State Pension age – Sky
City trader wins bid to overturn LIBOR-rigging conviction – Standard
Brewdog to close 10 pubs across UK, citing ‘ongoing industry challenges’ – BBC
BNY, Goldman agree tokenised money market fund deal – The Block
Wise co-founder slams ‘inappropriate and unfair’ US listing move – Standard
Payment on account deadline: 31 July to avoid HMRC’s 8.25% penalty – Which

Wall Street is starting to warn about the stock market – Sherwood
Products and services
FCA softens remortgaging rules in dash for growth – City AM
How does NS&I’s new one-year savings bond compare to rivals? – This Is Money
Get up to £2,000 when you switch to an Interactive Investor SIPP. Terms and fees apply. – Interactive Investor
Eight questions answered about cash ISAs – Which
How to save money on airport parking – Guardian
Get up to £100 as a welcome bonus when you open a new account with InvestEngine via our link. (Minimum deposit of £100, T&Cs apply. Capital at risk) – InvestEngine
Carmakers discounting EVs on fears they’ll not qualify for new grants – T.I.M.
Cut the cost of using your mobile phone abroad – Be Clever With Your Cash
Homes for sale with wild gardens, in pictures – Guardian
Comment and opinion
Selling sales – Money with Katie
Could the triple-lock really push state pension age to 74? – This Is Money
“What I learned from being disinherited” – Daisy Goodwin
Restrictions on capital flows should be considered [Paywall] – Financial Times
Small, value, or small/value? – Klement on Investing
The market has a short memory – Apollo Academy
What if gold went the way of diamonds? – Abnormal Returns
If active investing is the loser’s game, what is the winner’s game? – Morningstar
Avoiding investment scams [Podcast] – Rational Reminder
To Bitcoin or not to Bitcoin? The corporate cash question – Musings on Markets
Gilts (again) mini-special
Investing veteran Peter Spiller: brace for a breakdown in UK gilts – City AM
New OBR research paints a painful future for gilts – Interactive Investor
Britain is at the mercy of the bond markets as debt spirals – This Is Money
Naughty corner: Active antics
Should investors take another look at the rebounding property sector? – T.I.M.
Japan’s dividend and buyback surge – Verdad
Betting on GLP-1 – Market Sentiment
Palantir’s valuation is insane – Sherwood
Buffett’s intangible moats – Sparkline Capital
An interview with renowned short-opportunity spotter Mark Hiley – FT
Kindle book bargains
The New, New Thing by Michael Lewis – £0.99 on Kindle
The Tipping Point by Malcolm Gladwell – £0.99 on Kindle
The Everything Store: The Age of Amazon by Brad Stone – £0.99 on Kindle
Essentialism by Greg McKeown – £1.99 on Kindle
Or try one of the all-time investing classics – Monevator shop
Environmental factors
How can we negotiate with autocracies on the climate crisis? – Guardian
Gene editing could save species on the brink of extinction – The Conversation
Why Filipinos keep getting married in flooded churches – BBC
In Ukraine’s bombed-out reservoir, a huge forest has grown – Guardian
The right is waging war on climate-conscious investing – The Atlantic [h/t A.R.]
How plastic industry swamped vital global treaty talks – Guardian
Robot overlord roundup
OpenAI and UK sign deal to use AI in public services – BBC
The hater’s guide to the AI bubble – Ed Zitron
Gaps in our knowledge of ancient Rome could be filled by AI – BBC
Content and community – Stratechery
ChatGPT advises women to ask for lower salaries, study finds – The Next Web
Not at the dinner table
About that US-Japan deal – Paul Krugman
CBS’ cancelling of The Late Show looks political – Daring Fireball
Biting the Fed that feeds you – Sherwood
The anti-immigration backlash comes to Japan – Noahpinion
Fascism and first-time founders – Tech Dirt
Life is (increasingly) luck mini-special
The Ovarian lottery – A Wealth of Common Sense
Zero-sum thinking and the labour market – Kyla Scanlon
Off our beat
Should old home movies be canned? – Next Avenue
The sinister truth at the heart of the Coldplay kiss story – Guardian
If GLP-1 drugs are so good, should we all be on them? – Derek Thompson
Tales from the island of illness – More To That
The rise and rise of country music – Stat Significant
Seeing the lottery – Seth Godin
And finally…
“If you’re the first out of the door, it’s not called panicking.”
– John Tuld, Margin Call
Like these links? Subscribe to get them every Saturday. Note this article includes affiliate links, such as from Amazon and Interactive Investor.
- It takes more than a full working day to compile it, and for my trouble I’m advised Monevator is probably penalised for hosting what looks like a link farm. So it actually hurts our traffic! [↩]
A big thank you for the great content. I can’t even imagine the amount I have saved on financial advice from what I have learnt from these pages. And taking a hands on approach means that I actually have confidence in what I’m doing – I’m not sure I am comfortable to just accept “I am a professional and everything looks good” from a financial advisor.
Just a note, I originally stayed away from the Moguls subscription as I don’t want to be dragged to the dark side, however there’s some amazing not-active-investing content there which might be worth highlighting e.g. the discretionary trusts article.
Thanks for doing what you do, Monevator crew. There was a good interview with the CEO of the Atlantic about AI and the future of the media:
https://www.exponentialview.co/p/ai-and-the-future-of-media-nick-thompson
Present company excepted, people are blurring hobbies and hustles. What we’ve seen is people’s failure to turn a hustle into a sustainable business. What the capricious tech gods giveth, they can also take away. The core of an online business should be the website, and email list – ie stuff you own. Any other income or traffic from other platforms should be treated as gravy as it will likely end up enshittified by their owners so they too can maximise their return.
Not every passion has to end up making money (or even washing its face), or getting tons of eyeballs and likes to feel valuable. I don’t ride a bike, and think maybe I should monetise it with ads or go out running, hoping for applause from the crowds I pass.
Think also a dose of collective reason in the influencer-sphere needs to be applied. I probably read a dozen or so blogs, and listen to another dozen or so podcasts, and the same again of video channels. All extort me to pay them each $10/month. I might pay them $10/month in total, but I’m never going to spend $300+/month on content.
@Oninion — Thanks so much for the positive comments, and for being a Mogul supporter!
Yes, that Finumus article and the comments were pretty amazing. (I can say that because I didn’t write it haha. I’d love to run more of his writing, but alas he’s busy being a high-powered captain of finance on the 9-5).
@G — Cheers for thoughts. I’m not blind to the practical/cumulative implications of multiple subscriptions, but I would point out our Mavens membership currently works out at less than £3 a month on an annual membership. Obviously people will have to decide what that content (and the wider near-2000 free posts supported by membership) is worth, that’s a personal decision.
To be clear though the days of Monevator being a few hours a week side-hobby were long ago. (My new property newsletter is definitely just that!)
Writing today’s style of Monevator articles basically takes at least a day, and some of TA’s with the number-crunching and so forth take 2-3 days or more.
My Moguls ones take at least a week, when you take into account the background research and so forth.
And yet writing is the least of it nowadays! I can easily spend a couple of full days working on Monevator without writing anything. Admittedly some — perhaps a fair bit — of this is related to monetisation, in that it’s related to member support stuff, affiliate partners, or whatnot.
But vast amounts of it is comment moderation, reader email follow-up, and the like.
When you have a site that attracts a couple of million of visitors a year and many thousands of email subscribers, it’s a lot of work, basically.
I agree we have no divine right to exist and people can and will make their own minds up. Clearly most people don’t pay for most content on the Internet. I’m acutely aware I only pay subscriptions/similar for a fraction of what I read.
But make no mistake, you are paying for all that stuff you think you’re getting for free.
Google is a $2.2trillion company on the back of hundreds of billions of dollars of revenue. A lot of that is it’s cloud business, but a vast share is advertising revenue that has basically been transferred from the media industry to Google’s bottom line.
That’s why we used to have vibrant local newspapers and bulging magazines and a vast Cambrian explosion of independent websites which have all now gone to the wall.
Again, whether this matters to any particular individual is a personal thing. But I don’t think the media industry was equivalent to someone going on a bike ride for their own fun, and I don’t think people really understand what we’re losing (or have lost).
Thank goodness we have the BBC and the license fee to pay for it in this country. That shouldn’t be the controversial statement modern politics has turned it into, but the fact all political sides are annoyed with it is a good sign it’s still doing something right.
Hi TI! I’ve been reading your articles and the weekend links for a decade now and this is my first time commenting. I just wanted to say thank you and I am going to sign up today. I’d not given much thought to how Google + AI is making it ever harder for non-corpos to reach a new audience online.
A rare hello from one of those that reads the content from here by mailing list rather than the website… just to confirm we exist! (I don’t even remember the last time I opened the site!).
Ignoring the issues with search results, rankings, and AI search summaries, it’s hard to tell these days what’s a real blog made by humans rather than AI generated content. Feels like the real ones have already mostly gone.
>To some extent Google’s apparent capriciousness was well-intentioned, as it’s fought to outrun spam sites and AI slop, and other gaming of the system.
Google gave up and accepted SEO years ago, roughly when Matt Cutts left. Part of the problem is the fact that they have both search and display ads so even if you didn’t click on an add when you searched you are still presented with Google ads at the destination site.
What’s happening now is they are biting the hand that feeds them. Who is going to create content when they know it’s going to be gobbled up by some tech company and regurgitated for $20 a month.
Personally I’ve pretty much done with Google, I keep my documents local rather than online, most of my searches start at wikipedia or stack overflow and those that don’t go to Kagi a paid search service.
I’ve also stopped producing open source for much the same reason, I don’t want my hard work co-opted by some huge faceless corporation.
Is substack one way out of this problem?
“I knew it in my heart. You can buck the system but you can’t buck the dark forces that lie hidden beneath the surface.” (Bob Hughes, played by Matt Dillon).
Google and Meta are the Weyland-Yutani and Tyrell Corporations made real. 🙁
They’re the leading dystopian extraction mechanisms, funnelling ‘content’ one way, and ads and misinformation the other, all whilst carrying out algorithmic PysOps on a scale even PKD and Gibson could scarcely have conceived of. Parting writers from the financial fruits of their works, which are then mechanically and methodically mixed up, regurgitated and recycled by the corporate bot hordes.
They’ve done more than anyone to ruin the 1990s Brave New World of the internet ‘information superhighway’, just as Altman and Musk will surely ruin, and forestall, the liberating potential of AI.
“Don’t be evil” must have been meant sarcastically.
You can buck system, but not the primary trend which both underlies and drives it.
The dark forces of the enshitifiers can’t be compromised with, nor assuaged.
Time to strike out on your own 🙂 Good luck!
https://www.theguardian.com/business/2025/jul/26/why-early-retirement-isnt-good-for-uk-plc
Wow! The above came too late for your links, but might raise some hackles!
“Giving up work in your 50s is a wealthy indulgence that should be discouraged – the new pension commission needs to widen its remit”
@Delta Hedge #8
> the liberating potential of AI
Eh? What liberating potential?
> then mechanically and methodically mixed up, regurgitated and recycled by the corporate
botAI hordes.fixed that for you 😉 Other than this minor detail I’m with the general trend you identified.
@BBlimp #10: this problem with Google search long post dates the Referendum dramas. Many creators in multiple jurisdictions report problems in the last 2 years.
@ermine #11: “The ultimate hidden truth of the world is that it is something that we make, and could just as easily make differently” David Graeber. AI could be great. It could be Ian M Banks’ “the Culture”. But, sadly, that’s just not going to happen. It’s either going to be p(doom) or p(disappointing). We’re trusting the future to the likes of Zuckerberg and Altman. This won’t end well, even if we make it to the otherside of whatever’s coming down the pipeline for us.
@Pendle Witch #9, Phillip Inman (author) certainly doesn’t believe in freedom! Wow!! Infuriating stuff …
In view of the dire warnings about the state of the UK economy, and the bond market, I’d be interested to know where readers are holding their ‘safe’ allocations?
@Pendle Witch (#9) @Lalta (#12)
Certainly raised my hackles!
“Every western country needs their more mature workers to keep going, if not full time, then part time. And if not paid work, then unpaid voluntary work that acknowledges the luck that flows from being a 21st-century baby boomer in good health.”
I wonder what compulsion he is contemplating to get me back to work whether for pay or for free. Confiscation of assets, prison or what?
And besides, after multiple decades of working 50 hour weeks that, eventually, took a toll on my mental health sufficient that it took nearly two years to recover after retirement, I have little interest in going back.
Is there a possibility of a rival search engine gaining traction by focusing on human generated content. I’d certainly make the switch but no idea if its techichally possible and like you said, most of the content is already dissapearing/ indistinguishable from ai anyway.
@14 – Alan S – I don’t see any issue in people retiring early at all. They are just going to have to pay their way. Taking retirement at circa 60, living for another 30 years say, paying very little tax during that period whilst taking up a significant portion of public sector spending through state pensions and healthcare spending isn’t going to cut it. It may have worked whilst the retirement pyramid was 4:1 but at 2:1 it isn’t sustainable. It doesn’t matter if people in that position agree or disagree – their view isn’t particularly relevant. The only entity that matters is the bond market, current gilt yields suggests the market isn’t happy and we are clearly edging closer and closer to a major fiscal event. Which is why pretty soon the state pension will be subject to 20% income tax as it breaches the tax free threshold.
@G I think, increasingly, people will either have to pay for content or make do with the AI word salad that floods websites these days. It’s not necessarily about making money, it’s about protecting the time invested in the research and preparation of articles.
My own experience is very similar – blogging for a decade, spending more time than I probably should have done, with several thousand (free) subscribers and the associated workload of maintaining a site, answering emails etc. Not my day job, but done because I was interested and people appeared to value (=like) what I was writing about.
Then, a couple of years ago AI content started appearing. My content was scraped, regurgitated and misquoted. Some text that was clearly mine was patched next to complete garbage, often contradictory. Some sites simply filtered it through an ‘Americaniser’ (changing garden to yard, for example). No credit of course.
All of this is theft, but worse, readers unable to make the distinction follow the idiotic AI-generated ‘advice’ … some then emailing me when things went wrong.
So, I introduced a paywall and increasingly place content behind it. I’d hoped for 2-3% conversion to paying subscribers, I achieved more. I probably now get the minimum wage for the content I produce. I no longer feel as though I’m simply creating stuff for the AI scrapers to purloin. My novel or good ideas (very few of either, but that makes them even more valuable) can be aired without being immediately stolen, they can be expanded in future posts, fleshed out and — assuming they stand the test of time (and most don’t) — developed into something others can properly benefit from.
I recently had cause to explore a related topic to mine, needing some information for one of my posts. I reckon that 90% of the content I was offered was AI generated. Those that cited anything, usually included other AI content. Where scientific papers were cited, they were often misquoted or interpreted wrongly. I never found the information I was really looking for (perhaps it’s not known?), but the closest I got was in a 1964 PhD thesis available online as a PDF.
Finally, with the way the US and UK are biasing things in favour of AI, either through deregulation or changes to the copyright laws, you can expect more word salad and more paywalls.
And, frankly, £3 a month is less than a cup of coffee …
@substack … it’s a solution if you like contributing 10% of revenue to Substack. The advantage is that the platform is easy to start on. If you expect a dozen subscribers it doesn’t feel too bad paying Substack a tenner a month. However, when you reach a thousand, you’re paying £12k a year for a rather basic publishing platform. There are lots of alternatives. I use Ghost.
Cheers
PS @TheInvestor … any chance of allowing Markdown formatting in comments?
@SF (#15)
The need to increase [direct] taxes has been odds on for years. IMO it is now near certain as the indirect/stealthy/3rd party options are all but exhausted. To steal one of @ermines favourites: this is what decline looks like!
@Seeking Fire (#15)
I continue to pay tax (although not as much as when I was in work) but no NI (unless I want to top up my SP) and will be able to continue to enjoy retirement at my current living standard without a SP (if I live long enough to receive it) and close enough to it even if/when income tax rates rise. However, I’ll admit that if the NHS is dismantled and medical insurance premiums required (~£8k per year per person if similar to the US – although medicare is cheaper), then I will die younger than expected (private medical insurance in the UK is currently cheap because it tends to exclude chronic conditions).
Current gilts yields have returned to the levels they had from mid-1950s to the end of the 20th century – in other words, not entirely unusual. I also note that the bid to cover ratios at auctions in 2025 have ranged from 2.58 (4 3/8% Treasury Gilt 2040 on 3rd April) to 3.69 (for the same gilt auctioned on 23rd July), therefore it would appear that the market is still happy to invest in gilts but only at lower prices than until recently. However, it will definitely be time to worry if those bid to cover ratios drop. Of course, rolling maturing bonds will now cost more together with increasing coupons on ILGs if inflation rises.
I started reading the ‘hater’s guide to AI’ and now wish I’d put it through an AI summary program.
I do agree with him though. I probably have more exposure to AI than most as I frequently have to test various flavours of it, and it’s …interesting… as a technology.
It’s far from vaporware, but it’s also very far from intelligent, and it doesn’t really seem to be improving hugely despite the ongoing work on it. I think it’s entirely disingenuous to claim that it’s somehow going to bridge the gap to AGI, it’s not playing the same game at all.
I also agree with Ed that it replacing the workforce at large is also just not on the cards – again, it’s really not intelligent at all, it is not truly reliable at even the simplest of tasks.
It’d all be fine if it was a quiet innovative corner of tech trying to break out and gradually finding more footholds in particular areas where it can be useful, but it seems utterly out of whack that it is priced so gigantically in the US stock market as if it is the saviour of the innovative future.
I think perhaps the ongoing spending/investment into it is a symptom of a lot of loose money flying around, and a sharp recession will see off a lot of that, as it simply does not deliver anywhere near enough profit.
@PendleWitch. Being a Guardian comment, the article downplays the real point of that IMF paper which forecasts UK spending rising around 8% of GDP by 2050 to pay for an aging population (5.5% for other European countries). It recommends cutting the pension triple lock and replacing it with an earnings lock and some people copaying for NHS treatment. Both sensible policies which seem unlikely to be taken up because of the continuing delusion of the voting public.
Re: Krugman piece on Japan deal. Basic arithmetic does seem beyond US voters who do seem to swallow the nonsense from Trump. It’s just unarguable that current account + capital account = 0. Yet, people prefer fantasy to basic math.
@far_wide > and now wish I’d put it through an AI summary program.
mustelid powered summary –
AI sucks. It’s not all it’s cracked up to be. It is burning money for everybody but gold rush shovel maker Nvidia. It is being used as an underhand distraction for Western companies to outsource Western workers, using AI as a convenient scapegoat.
I asked perplexity to summarise the URL in 150 words. I can’t argue with the summary, but it was curiously unsatisfactory. Picking out the essence of those 150 words
It’s more technically accurate than my summary but I reckon my first three sentences are a purer extract of what Zitron means. It’s possible I smooshed together the jobs bit from a different EZ diatribe, as reading his post again I can’t substantiate that reference in this particular post 😉 Messy pups, human thinkers, eh?
But the AI summary is like drinking Huel rather than going out to a decent restaurant.
I enjoyed the long-form content, but I have a tendency to prolixity myself 😉 My mother taught me to read before I entered primary school so I don’t mind long-form. Conversely I can’t abide podcasts and Youtube talking heads*, because the data rate is so horrifically slow, less than a fifth of reading speed and I get bored, and I can only understand things up to ~1.5 x speed. I do accept I am probably an outlier. I expect in 50 years time we will lose the art of writing and become an oral tradition, apart from a thin elite, comparable to the priesthood of previous eras.
* Youtube scores where it shows you how to fix an engine or get into something. Show don’t tell writ large
@The Apiarist I think there’s a certain irony that a couple of my favourite reads each week are essentially a curated set of links to other sites which (mostly) offer their content for free. I imagine we get to the point where the curated lists and all of those sites linked to are behind paywalls, and that the commenters below posts such as these will also want paying for their comments…
I’m sympathetic, because I once had a blog or two which started as a hobby, grew, took up more time, became more hassle, got monetised etc – and eventually I folded them all as while I’d quite enjoyed them as a hobby, it wasn’t fun as a business (and after a few good years, wasn’t paying enough to compensate the amount of “spare” time they were taking up). Get the frustration of copy cats too.
I understand that running a successful (however that is defined) site takes time – but current situation of all hoping to be paid for their time does not seem realistic.
@Martin T — Re: risks to the government bond market from worries about the UK finances, I don’t think there’s really any credible prospect of the UK not meeting its obligations / repayments in nominal terms.
The portfolio return risks would more manifest in more issuance / higher than otherwise rates, especially on long bonds, and inflation eroding the real value of your investment. The way to combat this in a gilt portfolio would be to keep duration short and to hold index-linked gilts (of short duration) in size alongside conventional gilt holdings.
I’d also want gold and cash in my ‘safety’ cushion (and I’d also hold equities and all the rest of course, including plenty overseas which would helpfully diversify away from Sterling, which is unlikely to do well in some of these downside scenarios…)
@all — Thanks for the solid discussion as always, and also of course to @Chatters and others who have become members on the back of this update. I don’t find it easy to write promotional content and I don’t do it half as much as the experts recommend, despite how it might feel to @G and some others 😉
So, really appreciated!
(I see a few people have converted to Annual plans, and so I hope to see a few referral discounts getting enacted soon too… 🙂 )
@The Apiarist — Oops, meant to add a special thanks for sharing so much info. It’s always good to hear corroborating evidence, and heartening that you’ve made a good go of it.
RE: Substack, yes I’m not convinced, especially as it increasingly seems to pivoting towards video and indeed more being a platform. Time will tell I suppose. I’ve set-up my fun property blog on Substack to feel it out, and potentially to try to capture some readers there from its famed recommendation engine who might enjoy Monevator too.
RE: Markdown, I do deliberately keep the comment system here as simple as possible. That said you can include some HTML and whatnot if you’re know what you’re typing, which perhaps you do from running your own blog. (I tend not to as I edit and comment in the dashboard).
Off to have a look at your site. 🙂 Cheers!
@The Apiarist: before the Mordor of Meta and the Autarky of Alphabet the internet was awash with wonderful, novel, quirky, beautiful, enthusiast created websites and blogs like yours. Thank you 🙂
@far_wide #20 and @ermine #22: the LLM hype machine is a fire hydrant compared to trickling stream of criticism from EZ and the likes of Gary Marcus.
I’ve been trying to update the comments to last years’ “First they came for the call centres” piece on Monevator, and have ended up (after much going back and forth) broadly triangulating with @ermine’s estimate on SLIS of a 60% chance that LLMs are BS, and that this isn’t a “I’m telling you this is it” (a la Jeremy Irons/John Tuld in Margin Call) moment for humanity.
Notwithstanding the incontestable fact that, to date, accuracy in LLMs has only scaled to the one twentieth power of training computation, there’s always the kernel of doubt that we could be missing the next big thing, as (only today in my YouTube feed) this LLM scripted and generated take on April’s “AI 2027” scenario illustrates:
https://youtu.be/ZxvPdYMw_Sw?feature=shared
> in my YouTube feed) this LLM scripted and generated take on April’s “AI 2027” scenario illustrates:
This is a a movie. We’re suckers for stories, it’s the same as anything on CNN. I watched it at 1.75 times. And it stank worse that the most terrible B-movie I’ve ever seen. Fire that LLM. I didn’t care enough about the narrative to get to the end. Joseph Campbell would like a word 😉
It’s one way we could die – without the humour
Must admit I’ve not found much use for the enterprise LRM (large reasoning models) from the major commercial providers (OpenAI etc).
They are marketed to me by the tech teams as very good at “summarizing all the research you get from banks and analysts”. I get over a thousand research emails a day and, yes, ChatGPT will summarize their content for me. The issue is that I don’t want to summarize that research. If I’d ever wanted to do that I might have hired some grad analysts. I didn’t because it adds no value.
What I do want is for them to trawl through those thousand plus emails and extract the key insights that nobody else noticed, synthesize them and construct a trade idea. ChatGPT cannot do that. It has absolutely no idea how to do that. Nor can the grad analyst.
ChatGPT is great at replacing a whole bunch of activities and jobs. It can produce as many pitch books in Powerpoint as you could ever want. All very nice. But most of those activities seem, frankly, to be BS jobs. The sort of thing we give to clever grads to do to take up their time so they don’t actually do any harm while they learn the ropes.
So far we seem to have invented a way to replace humans doing BS jobs. Rather than actually eliminate the BS jobs. Is that progress?
I see a certain convergence in these comments.
One of the things driving me to avoid OMY is the daily assault of corporate BS around AI. I’m sure the bright grads are already harnessing it and not putting in the time they claim – I’ve noticed quality of output seems to have increased as the amount of manual copying and pasting has been reduced.
As I work in a people services business it doesn’t take much to spot that while for the owners of the capital “AI” is a great way of driving payroll down it’s also a great way to ensure that the business drives its model out of existence. It ultimately won’t take much for customers to wise up and say – hang on you’re charging me this rates pasted on old world timesheet type justification but you’ve now got agents and datasets doing it at basically nil marginal costs.
I feel when the great reckoning comes there I’m better being out, perhaps with a different (hobby) hustle that eschews AI slop and emphasises personal creativity and/or human connection. And yes that might mean I’m a less “productive” taxpayer but possibly more content in a number of dimensions.
@Alan S,
FWIW, I tend to agree with the OBR forecast re gilts. The linked II post is silent on probably the largest driver: that life insurers (who are busy gobbling up DB schemes) far prefer corporate debt and on transfer of DB schemes invariably sell all/most of the erstwhile schemes gilt holdings and re-risk.
@TI #24 “I’d also want gold and cash in my ‘safety’ cushion” …..
Safe indeed! https://warwick.ac.uk/news/pressreleases/superheated_gold_survives_temperatures_far_beyond_its_melting_point
In my shop AI is no longer quite the flavour of the month. After a good 18 months of the firm singing AI’s praises our customers are now asking where their share of the savings is.
While the firm has tried to avoid publishing precise numbers it has been suggested that AI tools are saving practise staff 18 hours a week. Clients who are billed by the hour want to see a similar drop in the hours they are being asked to pay for.
While I doubt that the benefits of AI have been too wildly overstated, some commercial reality applied to the savings will be appreciated.
I must admit I probably wouldn’t have taken out a paid subscription if you were running them through Substack, given the company’s apparent comfort with hosting extreme right-wing voices. I wouldn’t want 10% of any subscriptions going to them.
I also get the sense they’re gradually creating more of a walled garden, somewhere that it becomes harder to leave. And while its cross promotion of newsletters can maybe generate more subscriptions, do they convert to paying subscribers? See, for example, the Garbage Day newsletter from last year https://www.garbageday.email/p/its-time-to-leave-substack
“In the last year, I had 5,000 people sign up for Garbage Day through the app. Of those 5,000, only 37 people converted into a paying subscription.”
They eventually moved to a different platform and have gone from strength to strength.
It baffles and saddens me when I see people assume Substack is the best or only option for setting up a newsletter when there are so many alternatives, which will help achieve your aim better – owning your content and subscribers, being free of large companies. Not just Ghost but Beehiiv, Buttondown, etc.
Hi Investor
Became a Maven over the weekend thanks to your prompt. Been lurking a long time and only started to appreciate that independent content is now sidelined when I google and all I get is AI slop / sponsored links. Like how you link to other independent financial bloggers.
I’ve been “delving” (useful spot check whether an AI article or not) into how to use AI on LinkedIn Learning. The key takeaway I gleaned was that AI can automate the drudge and amplify the good stuff you do at work. It comes down to the quality of the prompts, but AI is no genuine substitute to human creativity.
@TI, thanks for awesome articles. My yearly subscription from early this year has been rewarding as I can access older articles, and what’s more is I can support the good work you guys are doing.
I bumped into the site accidentally well before covid and words of wisdom and courage has kept me sane from staying away from logging onto my brokerage account and hitting that sell button during huge market drawdowns.
Maybe lifetime mogul subscriptions? hehe kidding.
From the DT today: “Google is overhauling its dominant search engine with an “AI mode” that will no longer provide links to other websites, in a major shift expected to cause turmoil across the web…. Instead of showing links to websites, the AI mode generates its own answers using information from around the web.”
Watershed moment. “Content creators” (we used to call them writers) don’t have a choice now. Either push back hard against the Google gargoyle, and both gate content and block the giant’s web crawlers, or face being turned into the website equivalent of one of those ‘human batteries’ from the Matrix.
@Al Cam (#31)
While you are right that, at around 24%, DB schemes are holding a lower proportion of gilts in issue (e.g., see https://www.actuarialpost.co.uk/article/defined-benefit-pension-rundown-cuts-gilt-demand-23059.htm) than in the past, I do wonder whether recent increases in yields together with increasing savvy in retail investing (e.g., ILG ladders and individual high coupon gilts for income) might at least partially offset this.
McQuarrie’s work on US corporate bond returns (https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3740190) which found that the 10 year rolling return premium for corporates was only about 15bp (and highly variable) and much smaller than the yield premium is interesting (it is a long, long paper!).
Maybe I am just a suspicious git, but I don’t trust anything AI says, it makes Boris Johnson look like a paragon of the truth. Fundamental assumption in the output of either is that there’s something wrong in some/all of this, and it’s necessary to find out what.
The web search mode of ChatGPT is fine because it does show you source links, and I right click on these to see where GPT’s claim comes from. Sometimes the source looks credible, sometimes the source just looks like an AI slop listicle. And sometime ChatGPT just didn’t get properly what the source was saying.
People who take any source of online information without trying to substantiate or second-source it are asking for trouble, people who take an AI summary without demanding to know on what do you base this deserve all they get. Validation/due diligence is imperfect and nobody has enough time, I get that, but heck, you don’t just charge across the road without looking. A lot of AI output just doesn’t pass the sniff test.
Good at sifting widely, and there is something positive and useful about that. But taking it as gospel? Mad.
@ermine
I absolutely agree. Currently trying to refine plans on a extended trip I’m taking. I absolutely do not trust the Gemini summary when it comes to things like local public transport services, opening times, costs of things or even times of year and locations for peak wildlife spotting.
I can see it or other generic AI search might be really good at cobbling together a rough itinerary from a standing start etc but fear it could leave me stranded in some manner by not picking up specifics of road closures, border crossing access hours or wildly misled about availability and price of local taxis.
I might be missing something very obvious (please correct me if so) but…
If everyone refers to AI slop summaries, and if noone clicks through to real sites written by real humans:
– Then those sites get no ‘traffic’ (and so no ad revenue etc).
– Like a shop without footfall (but with window shoppers, who then buy on Amazon instead), the sites will end up closing down.
– And then they’ll be no sites left for the AI slop to summarise.
Final result: unhappy internet users, no ‘content creation’ or creators, and, ultimately, even the AI slop gets canned as having nothing left to plagiarise and distort.
Can’t see how this is going to work out for anyone or anybody in the long term.
Talk about killing the golden goose.
Google effectively created easy access to the internet from ~1998 (remember using web crawler search before?), and now it’s about to wreck it.
Ultimately, like the misnamed ‘AI’ they use, Big Tech is too stupid not blow itself up, taking everything it touches with it.
@Alan S (#38):
You may be right about ILG ladder’s etc, but I would imagine this will be small beer cfi the erstwhile DB funds demand.
Re: McQuarrie’s work on US corporate bond returns:
AFAICT, this work effectively strives to compare apples with apples – but, of course, that is not quite how it actually works IRL for the LI business in the UK. In general, LI’s indulge in (sceptics might even say some of the LI’s gorge on) regulatory arbitrage via a mechanism known as the Matching Adjustment (MA); see e.g https://monevator.com/money-market-vs-bonds-which-is-best/#comment-1887949 and associated comments in that @M post.
To my eyes (at least) this information is the missing link as to why LI’s do what they do and re-risk [in spades] erstwhile DB’s bought up in BPA deals.
FWIW, the first link provided in comment #52 (in the @M post linked to above) [to eumaeus*] makes some eye-popping claims and I feel that it is probably fair to say that the authors of that blog are deeply sceptical about the MA.
*sub-titled “keeping an eye on things”; IMO it is worth reading their “about” page to understand their choice of blog name – apologies in advance if you are familiar with Homer’s Odyssey
“cf” not “cfi” in para1 of comment #42 – my bad!
@Delta Hedge #41
> Can’t see how this is going to work out for anyone or anybody in the long term.
I don’t know. I am not sure that I will live to see it, but perhaps the reductio ad absurdam will give us a chance to redeem the original sin of the Internet which is the surveillance/ad based model.
This did not arise because there were inherently a bunch of misanthropic gits who planned it from the get go. The early WWW (1992 to ~ 1996, arguably through to the dotcom bust) had no easy model of collecting payment. I worked on one system in the early 1990s, so help me God, it took data from a Perl cgi form and stuffed it into GPG and emailed the output to the customer to extract with PGP for entry in the back office to a manual terminal. Yes I know, security risk just about everywhere in today’s world, but it ran until 2001 when it was thankfully retired and there were no issues 😉 Another organisation I did work for 30 years ago used http POST of the purchase order into a form for some American credit card processor, which at least didn’t have security problems we were responsible for. That firm just had pretty usurous fees.
And so the original sin was born. You couldn’t pay small amounts with money, so you got to pay with attention. It skewed my thinking – compared to young ‘uns who pay in app, pay for mobile subscriptions and are more than happy to have subscriptions I don’t do them in the information space. This site is the only content I pay for(explicity, as opposed to implicitly, though I block ads ruthlessly), and that is not for the usage or even the articles, many of the paid for ones are way beyond my pay grade and very seriously not my tribe, but as thanks for the tuition and the guidance, and the tips that aided me to crawl from the wreckage of the GFC and get ahead.
I am not short of the £80 p.a. but it gives me serious cognitive dissonance because it is the one, single exception. I have just unsubscribed from Duolingo because it is enshittifying with AI and also my German has improved enough I can get away with free 😉 Not only that, I paid Mrs Ermine to add me to her subscription, so I could honestly tell myself I wasn’t a subscriber
We ended up with the crappy surveillance and ad-based model because of that historic problem. Maybe we have to let it all gum up and burn to the ground so we can start over.
This Nielsen article from 1998 shows you the world we could have had. I disagreed with him when I read it in 1998, because IMO there is a non-trivial cognitive cost to thinking ‘do I want to pay anything for this’. I do acknowledge perhaps there’s something wrong with me 😉
Your previous comment
> Either push back hard against the Google gargoyle, and both gate content and block the giant’s web crawlers
may be one way it will go, Substack and its ilk may be another way it goes balkanising things into walled gardens, or it will just jam up, and we will read books again.
Somewhere in some Asian kid’s basement the putative solution will be built in 20 years’ time. Having lived the alternative. Nielsen’s proposal looks tolearble
I did take the kick up the backside of your comment to tell Cloudflare to shut out AI crawlers. Unlike TI I don’t try to monetise, it is the learning and interaction with like minds that is the way I am rewarded. It vaguely grates when I see someone dupeing the copy. ChatGPT learning to speak ermine, well that’s one way to throw sand in the wheels I guess
I don’t want @G #23’s tongue in cheek proposal that TI pay me
presumably on a all subscription site. I hate gamification is all of its forms.
Perhaps I am just an ageing hippy still holding onto the notion that information wants to be free. The pre-Millennial web sucked in many ways but it was human and much less homogenous. A guy I worked with observed that as the bandwidth of internet access went up the signal to noise ratio went down, he was not wrong.
/OMYAC
@ermine — Well obviously I’m sorry you feel that way about @Finumus’ article. But to be honest, I’d run it again. 😉
I will always include a proportion of articles targeting the (sometimes very) wealthy.
Three reasons!
Partly that’s because these do represent a sizeable slice of the Monevator audience.
But it’s also because I want to know what the rich are doing, even if I can’t do it myself, for any clues as to how I can think/act like they do. I’ve seen enough in life to know that those born into/adjacent to money have advantages from seeing how rich family members operate that helps them have a very useful relationship with money quite different to how I was brought up, what I saw around me, or to be honest most of my 20s before I did a lot of reading.
If I can narrow that knowledge/experience gap between the haves and haven’t-yets, I’m all for doing so, within reason.
Finally I will publish such things for political-ish reasons, for want of a better word. I would rather people had some idea about how, say, IHT can be mitigated rather than they just parroted what they read about rich people not paying IHT. It’s a more informed and nuanced discussion about policy then.
It’s kind of ironic that my post about getting referral bonuses has coincided with this dissatisfaction from you, and when I was crafting the referral tiers I was often thinking about you and 2-3 others as potential ones who could hit the ten-and-then-free level. 🙂 Ho hum.
Re: paying for subscriptions and the like, well I’m afraid I don’t agree at all. And this isn’t just out of self-interest, albeit that is obviously in the mix as discussed in the relevant recent Weekend Reading.
We are all of us (even you!) living vastly more of our lives in what we might term mediated environments. Some more, some less for sure, and quite happy to believe you’re at the less end of the spectrum (IIRC you didn’t pay your TV licence?) but anyway, I think it’s completely unrealistic to, say, consume more media than ever (as a society) but pay less for it than ever, which is what your statement implies if applied by everyone.
In 20 years it’s extremely likely that huge swathes of even our normal days will go by with some kind of augmented layer on top, most of the time, too. Personally I’d want to adjust my mindset before this arrives, lest I’m (not) paying for the Easyjet or Google Adwords version of what then passes for reality.
Finally, I used to buy 2-3 magazines a month, a newspaper most days, 2-3 CDs a month, and go to the cinema 1-2 times a month. Now I mostly do none of those things, so should I expect whatever economic incentive helped make that media happen to come to me for free? I don’t think so. Why should creators / journalists / whatever not be paid, while I’m still paying for my sofa and my ramen noodles? I don’t think that’s fair.
So I do pay, within reasonable limits, for a range of subscriptions, and I’m extremely happy to do so. Not least because it takes up much less space than all the above generated!
Obviously it’s up to you what you choose to pay for and Monevator might be a close-call on many people’s lists, I accept that. We can’t pay for everything, or even a slice of everything. But I think it’s almost a moral imperative to pay for something, or at least if you don’t then IMHO one is excused from having much of an opinion about the quality of output that results. 🙂
In any event, thanks for your membership to-date, and cheers for considering carrying on with the Mavens sub, given your views on all this. 🙂
The problem is the medium really is the message now.
Must have spent £1k as an undergrad in mid 90s on CDs (at the time, IIRC, £5-£7 each 2nd hand, against 2 or 3 for a quid these days).
Despite keeping them all, and having a working CD player, I still just end up listening to exactly the same tracks on YouTube; notwithstanding all its intrusive ads constantly interrupting.
Why?
Because it’s frictionless.
Open up YouTube, and within 5 secs the track is playing.
Whilst to find the same track you have search for a particular CD, get it out of the crappy plastic box, switch on the player, and then get it playing – all of which takes the best part of 5 mins.
The world’s dematerialised.
Travel to an office to have meetings over Teams with colleagues who are invariably elsewhere.
Sending emails rather than speaking F2F.
Its all now intermediated through the electronic ether.
And the more it gets intermediated the less real any of it seems, and the less willing people are to pay for any of it.
If people can still grab hold of it – a book, a newspaper, a newsletter, a record or a CD – then they’ll open up their wallets fairly willingly.
But an article on the website for the Atlantic, or an essay on Aeon. Not so much.
Now, they might be happy to stream. All the world’s music for a bit over a tenner a month and no ads. That’s quite a deal. Even a Scrooge like me can see that.
But getting them to cough up on a per piece, a per artist or a per writer basis?
Well, that’s a tougher ask all round.
Especially when people have gotten well and truly conditioned by the ad based model of the past 25-30 years, a model where they are the product, and Google is the consumer – a modern version of one of those Hieronymus Bosch paintings where some poor peasant soul is getting eaten by a monster.
I suppose it’s like our politicians and policies. We get what we’re willing to pay for and commit to. Less input worse output.
@TI #45 Apologies, I didn’t mean (and I don’t think I said) that Finimus’ article shouldn’t run 😉 I said it pulled the black tip of my tail, but that’s me. You run a grand site and good luck to y’all.
My history is probably why I struggle with subscriptions, as it happens I did get a TV in the pandemic I think. So for clarity no criticism intended. That was how I felt, and you can’t please everyone all the time.
Plus since checking I am up rather more than £80 due to, ahem, tips in the last year so I’ll leave it be. Obvs I know they’re not tips, no sir, but you know, a fellow’s gotta do what he’s gotta do.
There are easier ways to pay for the subs, in the case of this specific site. I didn’t mean to be boorish.
Once again, sorry for any offence given!
@Delta Hedge #46
Fair cop, guv. That’s me 😉
Eh? Is not the ad stealing your time major friction? Breaking the flow? Doesn’t it make you want to punch the wall? Not only that, you end up listening to a narrower selection? On a minor point, I’ve never found a better solution than a CD for classical. I use a LAN streaming system from a local server for ordinary music but artist-title-album just stinks for classical. It’s so much easier to pick it from a shelf than drill down for it. While you’re up you may as well get a glass of wine, what’s not to like?
> All the world’s music for a bit over a tenner a month and no ads.
I considered Tidal. Losslessly compressed and compatible with my hardware. But I just couldn’t make myself do it. It’s not the money. It’s the principle, it’s the lock-in, bah. Not. Going. There.
As Kyla Scanlon said, friction has value
As a teenager I read some Asimov SF story, think it was called the Feeling of Power where the protagonist rediscovers doing arithmetic by hand – everybody used what we would now call smartphones – there was something ineffable about the mastery. Friction is not always bad.
@ermine — No harm, no foul, Sir, but thanks for clarifying — and doubly so for re-upping on Moguls! 🙂
@ermine, thanks for that Nielsen link, fascinating in many ways. I do remember the late 90s internet when people started talking about this amazing new search engine called Google- where did it all go wrong…
@Larsen #49 > where did it all go wrong…
with the smartphone IMO. They could yank Your chain everywhere. At least on the PC You had to go to Them. I don’t carry a smartphone regularly but pretty much everyone I know does. I often see children, dogs and even romantic couples* in restaurants being phubbed. But it’s a lost cause.
*At least they’re often doing it to each other.
The central defining paradox (or dilemma, both for ‘content creators’/blog writers and for consumers/readers) is just the same now as it was back in 1984, at the time of the first hacker convention, where Wozniak & Whole Earth Catalogue’s Stewart Brand first summarised it:
“information wants to be free, but it also wants to be valuable”
Digital data is infinitely, effortlessly and costlessly copyable; endlessly mixable, malleable and compostable; ceaselessly manipulatable; inexhaustibly aggregatable, capturable and storable; and of course instantly (re)transmissible.
It almost boundless. There will soon be a septillion bits of data in existence. That’s oceanic.
But the right bit of information in the right hands at the right time can be priceless.
The consumer should pay a negotiable and freely agreed price depending upon the utility (value) offered, and the producer should receive.
But the search/social media/LLM provider-advertising-industrial complex have gotten their blood funnels betwixt the creator and the consumer.
They got in there first (FAPP).
They weren’t the smartest. They certainly weren’t the best. But they were the first.
They forestalled the micro payment economy that Nielsen foresaw, but which never stood a chance.
Their relationship to both the production and consumption sides of the internet is best understood as analogous to the face hugger in Alien.
They will not be easy to displace. They know our weaknesses only too well now.
And it’s not just the emergence of the enragement-engagement economy.
The algorithmic gatekeeper has replaced credentialled experts, public intellectuals and mainstream politicians. It’s a real world torment nexus.
You can just switch the bugger off and touch grass, y’know 😉
> But the search/social media/LLM provider-advertising-industrial complex have gotten their blood funnels betwixt the creator and the consumer.
Yeah, but the consumers and creators surrendered the end to end principle for the convenience of platforms. This was amplified by people surrendering the open web for apps because smartphones are/were just a little bit crap. Walled gardens in the platforms, walled gardens on the clients, it was always going to end like this. There’s friction in maintaining your freedom, and the results are in. Most people CBA. As Moxie Marlinspike, he of Signal, quoth, people don’t want to run servers. Fair enough, your ass will be owned by the people who roll up their sleeves, spit on their hands and do run servers. Search was so easy, now we can’t find ‘owt because all the earlier methods had friction. There is an argument to be made that Weekend Reading is one alternative to the big G, it’s one or two guys poking around for interesting Stuff. It’s another way, indeed, it’s the way web search started.
I’m not so sure you can blame all the rot on Them. You listen to a narrower selection of music because you CBA to change the CDs (or select some other way) and your flow is interrupted by ads. That’s perfectly your call, but ease comes with a price. In the limit that price is going to be your feed filled with Youtube’s AI bands that sort of riff off what you’ve listened to before.
Many small ways lead to perdition 😉