The usual collection of things which caught my attention in the last week (or so).
Coming to you from Amsterdam and our first overseas trip with a baby.
Product Delivery
Related (perhaps very tenuously) to getting the right things done.
The Browser Company (YouTube) and a good summary of the product on Verge, it’s pretty common to see people (content marketers) talking about how important it is to ‘build in the open’ as a way to generate a community and hype for a product. In practice though it seems that the main people actually doing this are content marketers and consultants effectively hawking their services1. Arc is a very notable exception with 73 (!) YouTube videos covering things like fundraising, feature choices, monetisation, competitors, and tons more. Launching a new browser might be even more challenging than a new search engine (#1-20) - and this business is actually a browser+search+AI hybrid. You interact with an AI chatbot (OpenAI & Anthropic) which then runs searches (Google) and then it either builds a page or shows a page (Chromium). It looks slick. But as a business the inherent challenge is that their suppliers are essentially all direct competitors in that they are either the incumbent (Google) or are directly disrupting the incumbent. Fortunately AI is a very competitive market and prices are falling (thank you infrastructure build-out and Meta open sourcing Llama) - however they are still very high for a product which is free. For search the real question is whether the product would work if they had to switch to DuckDuckGo or Bing.
Disclosure; long META.
GOAT is now a free audiobook (MR), Cowen continues his pioneering work in applying GenAI to writing (#1-40). This time directly working with OpenAI (and presumably GPT-4o) he has generated an entire audiobook version in his voice. Presumably the model was trained on (or able to take as a RAG input) the audio from Conversations with Tyler. In either case the result is uncannily good.
Disclosure; long MSFT and therefore OpenAI.
Google’s AI search results are already getting ads, this was inevitable in that the entire raison d'etre of the company is to inject ads into search results. The result is better than I expected, I wonder whether this kind of model will further drive ads towards a Jobs To Be Done model.
Nestlé to launch GLP-1-friendly products via Chartr, Ozempic and similar are clearly a giant strategic threat to the industries which exist to sell highly appealing and not very healthy food - aka ‘junk food’. As I understand it these drugs largely blunt the desirability of food which helps people eat less. Nestle seems to be experimenting with leaning into this and producing food which is likely to be less appealing (less sugar, fat, etc) but provides nutrition in appropriate portions at an aggressive price point - without being soylent. It will be interesting to watch how this performs in the market.
Time sinks and money sinks, an interesting look at the dynamics of different businesses as either time sinks (think social media) and money sinks (think irregular but high value transactions). Typically companies in the first category monetise by selling ads to companies in the second category. The NYTimes seems to have hit on an interesting bundle where they combine different types of time sink (games v reporting) which allows them to charge a premium subscription and also sell ads at a premium.
How Crocs became a clog-selling profit machine, somewhat implausibly Crocs do $3b a year in revenue while earning the same margins (26%) as LVMH. The explanation seems to be that people really like limited edition ‘collabs’ and they have mastered churning these out.
Teen fintech Copper had to abruptly discontinue its products, another fintech struggling because they relied on a Banking-(License)-as-a-Service provider. These risks cut both ways; banks are taking on meaningful regulatory risk when they rent out their license and fintechs are taking on key-supplier risk when they rent a license. As I’ve said before I think the smart money is in fintechs who obtain their own license, either newly issued or by buying a bank.
Taking risks, Tom Blomfield (Monzo) on the power and importance of taking risks because the downsides of losing 1x your money (VC) or a year of your life (founder/early hire) are trivial compared to the upside of 1000x returns. Refreshingly Blomfield is explicit that this advice is most applicable to elites - when your fall back is living in your parents (large) home and walking into a job in top-tier consulting, law, investment banking, etc it’s easier to take risks. He’s also clear-eyed though that it’s good for everyone when people take risks because that’s how you get transformative businesses and products.
Off-Topic
Not even tangentially related to product delivery, but still interesting (to me).
The worker shortage is over (maybe), it’s very easy to generalise from your own experience to the broader economy - but that will probably lead to the wrong conclusions. I’m personally primarily exposed to tech/finance/fintech all of which have been going through effectively their own mini-recessions as they adjust to a world of higher rates2 which has meant waves of redundancies, hiring freezes, etc and ultimately more jobseekers than jobs. That experience hasn’t extended to the rest of the US3 economy which has had more jobs than jobseekers - which has driven wage increases and inflation4. The argument here is that the tight labour market had nothing to do with excess monetary/fiscal stimulus in Covid and was instead the one-off impact of many people retiring during Covid and the ongoing impact of an ageing population. Essentially that the number of jobs was flat but the number of jobseekers dropped. Migration (largely illegal) has now bought in far more jobseekers so the labour market is loosening. The UK’s version of this is more fraught because Covid coincided with Brexit so the UK endured both a wave of retirement and the sudden disappearance of migrant workers. Sternstein’s prediction is that the US could now tip into rapidly accelerating unemployment which would cause a recession, this would force a rapid (re)loosening of monetary and potentially fiscal policy.
The economics of the New Cold War, more and more people are getting (over)excited about protectionism as a way to build a military-industrial capability sufficient to deter/defend against China (Noah Smith here). I think it’s important to be very clear-sighted that this isn’t some kind of free lunch where we get jobs, growth and (avoid) a war. There are real trade-offs, which we can minimise or we can pretend don’t exist.
I think it’s important to say that 100 percent of the economic arguments for free trade apply completely in a world of intensifying geopolitical competition. My argument is that, for geopolitical reasons, we should do something that is economically costly. Building an aircraft carrier “creates jobs,” but it’s not something that you do for economic reasons. Judged in purely economic terms, aircraft carriers and missile silos and nuclear attack submarines are all just giant wastes of labor and material. Just because it’s economically harmful to maintain a powerful military doesn’t mean you shouldn’t do it. But you also shouldn’t delude yourself into thinking it isn’t economically harmful.
By the same token, trade barriers with China are economically costly.
What we ought to be doing is trying to minimize the total amount of trade barriers, and thus the economic cost of pursuing competition with China, by reducing as many trade barriers [with our allies] as we can.
And of course part of this is choosing what matters and what doesn’t, as Tyler Cowen points out slapping tariffs on things like solar panels and wind turbines doesn’t help national security but does make transitioning to renewables slower, harder and more expensive. Which in turn robs us of capacity to do other things (like make semi conductors or shells).
How to Build 300,000 Airplanes in Five Years, sticking with the war production theme it’s very common to see people talking up things like the Defense Production Act and WWII as an example of the US switching from civilian to military manufacturing and saving the free world. The reality is not so simple.
US armament manufacturing was kick started by the UK and France placing massive orders and paying upfront. This meant that manufacturers were already ramping up by the time the US entered the war.
Production lines largely were not repurposed from civilian to military because the fundamentals are so different. Military equipment is larger, heavier, with many more parts, much tighter tolerances and much more complicated assembly. Massive new factories with new tools were required.
Banks wouldn’t lend to build massive factories which produced single-digit profits and would likely be obsolete in a couple of years. So the government directly funded almost all of them. FWIW the banks were right, in that they were sold off at cents on the dollar after the war5 - but that’s a cost of doing business that government exists to manage.
Manufacturing workers were largely not repurposed from civilian to military production lines because (a) there were nowhere near enough and (b) as fit healthy men then tended to volunteer or be conscripted. Manufacturers needed to bring hundreds of thousands of women into the workforce - and completely redesign production lines to make things simpler (the women had no experience) and physically easier (the women were weaker).
They never really achieved mass production in the way that people imagine because (unlike cars) the designs could never truly be frozen. Instead production lines were constantly having to change and adapt to produce new versions.
Despite all of the above the US was producing so many planes (and everything else) that it was often cheaper to throw away a damaged asset than fix it. Flyable planes were pushed off carrier decks, used for target practice or just abandoned.
Palmer Luckey Wants to Be Silicon Valley's War King (Youtube), interesting profile on Luckey/Anduril. Couple of insights;
Anduril operates like a tech company, not like a defence prime. They are self-funding R&D in the hope of subsequently being able to sell products to government. This makes them far more focused on efficiency and efficacy.
Luckey claims that it’s easy to recruit from big tech, simply by giving people the opportunity to work on things with bigger impact than increasing DAU on a streaming app. This makes a lot of sense, beyond a certain level of material comfort people are motivated by a range of things and different companies can thrive by offering different combinations of things.
Luckey sees software as a durable advantage in warfare because in his view hardware is easily replicated but software isn’t. I’m very sceptical on this. For a start Anduril is a very hardware oriented company. I suspect what he’s really thinking is that the US has an edge over China in software development which they don’t have in hardware and that this edge can allow the US to make their hardware more effective. Even that though doesn’t really stack up. Software can be copied more easily than almost anything else. But more importantly as we see again and again (Ukraine, Yemen, etc) shooting down low cost drones with high cost smart munitions and platforms doesn’t make sense.
Why Britain is the world’s worst on homelessness (FT), homelessness is a function of supply and demand. Else equal fewer houses means higher prices and more homelessness. It’s not that hard. Don’t overcomplicate it with social housing. Just let people build more houses and the market will sort it out.
Why Did We Stop Building Beautiful? The Economics and Ideology Behind an Aesthetic Revolution, fortunately private development is also (slightly) less likely to be ugly and unwelcoming brutalist/modernist structures because they actually need to rely on people buying them.
Strata SE1, although even in private development you do have architects overselling things which don’t work in practice. My latest favourite example the three 18m wind turbines in the middle of London which vibrate the whole building when in use so have basically never been used - and even if they were would power only ~8% of the building.
Trends in Law and Order: SVU (Twitter thread), old TV shows and especially long running ones represent an interesting time capsule / time lapse view of cultural trends.
Sly (Netflix), this is the natural progression from the Arnold doco (#2-19) and it ends up a fascinating exercise in the contrasts and similarities between the two men;
The narrative devices in each were quite different. Arnold was structured around a photo album whereas Sly explicitly used the Rocky series to shape the narrative. In Arnold the B-roll was feeding his animals, while in Sly it was packing up his house - which really felt like it was the original idea for the narrative that ended up being abandoned in editing. Using Stallone’s old taped interviews was genius though.
They both had terrible relationships with their fathers. Schwarzenegger ultimately credits his father with his success (‘what doesn’t kill you, makes you stronger’6), while for Stallone he succeeded as an actor almost as a rebellion and he abandons polo (twice!) essentially in spite of his father.
Stallone is much more focused on and serious about writing and acting as a creative craft, whereas Schwarzenegger was attracted to it as the business of spectacle. Essentially the difference between an artist and a showman. Schwarzenegger very consciously pursued business (and political!) ventures outside acting while Stallone was driven by wanting acceptance as a creative - the sequences breaking down Rocky’s fight with Mickey and explaining his confrontation with De Niro in Copland is especially informative on this.
Schwarzenegger was able to very intentionally and very successfully transition from action into comedy. In contrast Stallone’s career has effectively been a long transition from (admittedly muscular) drama like Rocky I and Rambo First Blood into pure popcorn-action like The Expendables. Although in non-trivial ways he pushed against this in films like Cop Land and Rocky Balboa.
One wonders whether Stallone is a more private person or he simply trimmed content to fit the narrative and duration (~90min vs ~180min for Arnold). None of his three wives were mentioned by name, there were no details of how the relationships, started, went and ended. Only one of his five children was mentioned and that was essentially to draw a parallel with Rocky Balboa - the fact that the son died was only revealed as a cut-away. His steroid/growth hormone use went unmentioned - unlike Schwarzenegger’s.
They both have questionable taste, but Stallone’s is definitely worse. His house was filled with memorabilia and statues/paintings of himself as Rocky.
Resources
Nothing new this week
Something Fun
Describing this newsletter, found in a decent read on thought leadership which didn’t make the cut for a write-up above,
This newsletter is no exception
As a short sketch of the impacts;
Tech/fintech - higher rates mean that present profit is (much) more valuable than future profit. Which breaks the basic business model of losing money upfront to build/scale a product and worry about monetising later. This means less hiring/building/etc now as companies cut costs and try to reduce the upfront losses.
Finance - investment banking is exposed to the tech rates problem because M&A and IPO activity has dried up. Meanwhile in retail/commercial banking higher rates devalue (on a mark-to-market basis) the bonds which dominate balance sheets (and grew especially during the pandemic) squeezing Net Interest Margins (crudely profit margins) and creating bank run risks (see Silicon Valley Bank).
The UK is a different story with a proportionally larger finance/tech sector, much more exposure to energy prices and the ongoing impacts of Brexit.
I love that O’Hare Airport was originally an aircraft plant run by Douglas.
For Arnold this is real in a way that almost no one else ever means. He explicitly highlights how his father’s parenting led to both his success and his brother’s death.