Canvey Island, Cheap Eats 10 rules, and the Week the Vibe Shifted
Ten mental models for eating out with a focus on value for money. Life writing and curious notes from a bus trip to Canvey and how I think about home ed. Why “agentic” AI feels different this time.
Food: How to choose restaurants, 10 rules
Life and home ed: Bus to Canvey, Essex
AI: Eating software, eating the world?
Health: Walking benefits
Podcast: Actor Sally Phillips on family life, Down’s and acting; Elon on GPUs in space.
Links: Elon podcast, Civic Future Fellowship, Mercatus fellowship,
I’ve been busy with the investment world dealing with the fallout of the latest AI advances. It is possible that historians will look back and pinpoint this week, as the date of a turning point in history (although of course it is not a point in reality), the vibe I judge from equity markets certainly feels like this might be the case - although extremely hard to judge while living through historic moments. It could be just a normal week in 2026. I’ve been at 9/11, 7/7, great financial crisis, intern bubble and bust; thai beach tsunami and I know it’s hard to know when history is happening. The vibe was not quite like then, but part of the equity market vibe was. I will comment a little below. More fun, I’ve been asked how I think about choosing food when considering value for money; and I took a trip to the end of Essex, UK, and I lean into what we can learn in a life / home education way.
We went to Canvey Island, at the end of Essex, which no one really recommends. These are six items I learned about Canvey as we meandered to the end of the bus line. I wrote up some draft life notes and took some very very ordinary snaps. My friend comments:
That’s the most flattering account of the island, I’ve read…
Canvey is a reclaimed, ultra low-lying estuary island. Sea walls and flood defences are not background infrastructure, they are the condition of daily life.
The island’s modern town grew out of plotlands: small cheap plots that became DIY bungalows and holiday shacks, later hardened into permanent settlement.
A niche global-history footnote: Canvey played a role in early (gas) LNG-by-sea importing and storage, helping prove an industry that later scaled worldwide.
Canvey Wick is a conservation oddity: a former industrial brownfield that became an SSSI largely because of endangered invertebrates, a reminder that “wasteland” can be ecologically elite.
Labworth Café is an architectural surprise: a 1930s modernist seafront building and a rare piece of Ove Arup’s work as an architect, not just as an engineer.
Canvey now has a small but established Haredi (Orthodox, including Hasidic) community, largely relocated from Stamford Hill for cheaper, larger housing within reach of London, and it has built local infrastructure (shul, kosher provision, schools) as it has grown.
We took the 27 the way you take a local bus when you are not trying to “get somewhere” so much as trying to be somewhere. The bus as method. The bus as ritual. Travel defined by attention rather than arrival.
Southend loosens its grip slowly. You can feel it in the spacing of the houses, the streetlights, the way the sky looms. The 27 just keeps going, patient and ordinary, which is often the most revealing way a place can be.
More here as draft life notes.
This is also how I try and approach a lifelong education and home education. I’m leaning into learning about where I am and why and, to some extent, I pass this on to my autistic son. Following his curiosity and interests and seeing where they lead, and using that to find out more about how and why we are.
This is an expression of the “be curious” mantra given out. I keep swinging back on if you can “teach” curiosity. I think in some ways you can and that is something I think Home Ed / Unschooling can lean into. At least, in this case though, I think the learning companion has to be - or want to be - curious as well.
By popular demand here are some thoughts on eating out. I have loved food from young. I love eating in all forms: from street food and home cooking to extortionate, celebrated destination restaurants. I love cooking, shopping for ingredients, and thinking about the history and culture of recipes and gastronomy.
I think food represents human flourishing. It captures vital elements about a nation, city, or culture that you can miss from books or even travel if you pass over the food. Eating with people is fun and taste is pleasurable!
These are some ideas, riffing on other foodies (from economist, expected value thinkers such as Tyler Cowen, Nassim Taleb, to food critics and cooks like Anthony Bourdain and Jay Raynor) about how to approach eating out (eating in is a somewhat different matter, to which I will mention only one idea here: cook at home what restaurants struggle to cook.)
It’s worth noting I’m not a snob on pop art food nor a reverse snob on high art food. If I don’t understand a dish there is a good chance it’s my taste, not necessarily the dish. There are strong moral and environmental reasons to eat less meat but I’m afraid my cultural leanings have left me very omnivorous. I do think if we’re going to eat animals we should respectfully eat all parts presented to us.
Here are 10 ideas to help you think about eating out, particularly in London. The ideas lean towards “value for the food”. Many people will eat out for other reasons beyond food (eg to be seen in a cool place! Or to be in a well designed space), and I note uniqueness as a quality potentially worth paying for, but this note focuses on value for food. In zone 1 you can just about do much of this for £15 but hard to get below £10 although there are notable places you go below £10. Outside the centre, you can eat well for £10
The Mental Models
1. The Economic Equilibrium of Rent (Adapted from the Cowen Rules)
Tyler Cowen, the economist philosopher, suggests that quality food and high rent are enemies. If a restaurant is in a gleaming glass tower in Zone 1 with high footfall, the equilibrium suggests you are paying substantially for the real estate, not the ingredients or the cooking.
This applies also to places where people go to look beautiful and see beautiful people. Or, a restaurant with a beautiful view, high up.
There are plenty of exceptions where unique food experiences combined with the other intangibles you are paying for make the restaurant worthwhile but you should know what you are paying for.
Paying for a view may well be worth it, but you are paying for it.
(Side note: in my opinion, a good value view in London is in the Tate Modern restaurant; food is passable - you are paying for the view).
Nassim Taleb expresses similar views and extends the thinking towards Michelin starred places. Taleb argues you are paying in excess for such things.
I’m not a reverse snob - some of those places are unique experiences but you may be paying over the odds food-wise. That said, excellent and unique London food can still be inexpensive.
The Heuristic: Look for the “Low Rent” signal. If a place is in a basement, (in the U.S.) a strip mall, an industrial estate, or a “bad” neighborhood yet still busy then the money is likely going into the food.
All over London, there are cultural communities such as Nigerians in Peckham. Inexpensive but popular restaurants there are good places to try with this method as well (see below).
2. Don’t pay for the interior design and furniture; also consider survival or “lindy”
This extends to how a restaurant looks. If the chairs are uncomfortable, but the place is packed with locals, the food is often exceptional and cheap. That is a reason it survives.
Nassim Taleb has spoken about the idea of good things being lindy. They survive through time. [The name comes from “Lindy’s Law”, an old showbiz rule-of-thumb that got written up in 1964: comedians and theatre people hanging out at Lindy’s delicatessen in New York City noticed that a show that’s already run x weeks tends to have an “expected remaining life” of roughly x more weeks. So look for a place that has endured.
There are differing reasons why a restaurant might be lindy but it can be a useful signal. Although it can also fool you because the lindy quality might be its location, not its food.
Restaurants in this category in London might be Wong Kei (50 years, Chinatown; a cheap eat), St. John (30 years, although perhaps too young, Farringdon; moderately priced at bar), Cockney Pie and Mash (Portobello, 70 years; cheap eat); Sweetings (137 years, Mansion House; not cheap but pretty unique), Quality Chop House (156 years, Farringdon; moderate priced). Many of these are cheap eats
Even if the place is surviving for a non-food reason, it often makes the place unique and worth paying because of that.
3. Psychogeography and food hunting by community.
For good value food, you might think of London as a doughnut (H/T Jonathan Nunn of Vittles). The centre is an efficiency trap due to rent and plenty of people (tourists, maybe office workers) who will pay up as they have to.. The value is in the outer ring… the “outer boroughs” (zone 2+) often where diaspora communities live.
Now, I’ve had to travel all around London by bus with my bus-obsessed son, I’ve seen this first hand and has allowed me to try excellent Ethiopian, Nigerian, Kurdish, Bengali and Romanian food amongst others.
The Heuristic: Go where the community eats. Some London examples: Peckham for Nigerian, Walthamstow for Ethiopian, Green Lanes, Haringey (Kurdish, also Turkish; as an side, the communities used to fight a little, also see Kingston / New Malden area for Koreans and Japanese conflicts; as they brought their own nation state conflicts into London!)
4. Seek food by dish or cooking techniques
One good way of hunting food is by a specific dish. My family has talked about this and I’ve often found it true speaking to Asians who will debate where the best version of a specific dish is found. Then go to the place which is excellent for that specific dish.
You can do this in London by all manner of dishes.
Where is the best Cantonese-style roast duck? (arguably one of the Queensway restaurants: Gold Mine or Four seasons)
Where is the best prawn noodle dim sum? (Royal China Club?)
Where is a great Spanish egg omelette? (Barrafina?)
Or, overall technique… you can find pretty good barbecue in London in all sorts of styles but you are hunting barbecue technique and a place that specialises in that. The specialisation often gives the best or highest end technique for the money.
This may not find you the cheapest meal, or cheapest dish; but it will find the highest value of that particular dish. For this technique, it’s even better if the place only specialises in one or very few dishes.
5. Uniqueness and gourmets in other cuisines
Fuchsia Dunlop notes that Western palates often prize ease: boneless, skinless, soft; or at least do not seek out the textures a Chinese gourmand might.
It’s worth seeking out foods that challenge you on their own terms.
I also think it’s worth seeking out unique takes on food where a chef is building on a tradition. You can eat excellent Jollof rice in Peckham, but you can try a modern high art take on it in an expensive place (eg Asoka). You won’t be getting value, but you will be getting uniqueness.
This cuts across Taleb and the strict expected value economists but the food ideas and combination of techniques can’t be found anywhere else are still worth seeking out. This has a touch of the modern art about it. Understanding and appreciating modern current art takes an amount of learning but can be worth it. Your own mileage will vary and I think you will derive more from this the more food history and culture jives with you.
It’s a reason Sushi Tetsu, for instance, is on my list and why I like many of the tiny sushi master restaurants. They are offering unique insights and experiences which are worth paying a premium for if they are meaningful to you. (Although some really are very expensive as you are paying for more than just the food).
Uniqueness might also come from setting and can be a signal if the setting is cheap (see above), I am thinking in particular railway arches in London (40 Maltby is in this category); there is a Pakistani place which turns from a sweet shop into a resturant on the weekend that serves amazing halwa puri in a firefly chaotic setting!
6. The Flowery prose / Adjective Density Menu Test (derived from Jay Rayner)
Critic Jay Rayner warns against menus that over-describe. This one is more tactically and if you find yourself in an unknown place with unknown choices. If the menu outside is very embellished it’s typically a red flag.
Excessive adjectives and florid type faces are often compensating for a lack of confidence in the kitchen. (Some Asian places break this rule because they have one technique (eg wok cooking) but many ingredient combinations and no long food prep chain); but European style restaurants follow it (as they have long prep chains). In general, Long menus bad. Lots of type faces bad.
7. The Wine Markup (The Unbundling Rule). Avoid.
Restaurants bundle high-margin alcohol with low-margin food. You can buy wine online for £15 that costs £55 in a restaurant. (London wine mark ups are 3x or more). You cannot, however, easily replicate a 36-hour bone broth master stock at home.
I am basically now leaning to the advice of never or almost never buying wine or alcohol at a restaurant. It is rarely unique and always very marked up. Tea (in e.g. Chinese restaurants) or tap water is a better bet. There are two narrow exemptions: a truly unique wine, or a unique drink-food pairing.
Still that is for best value. You may just like wine at restaurants because it’s fun. That’s fine I used to like this as well, but you are paying over the odds. And my current advice now is avoid where possible for value terms.
If you will try wine, always ask to taste first if possible, especially if deciding between single glasses of bottles already open and go for something you wouldn’t get at home. But overall in terms of expected value. Avoid.
8. The “Weirdest Item” Strategy and also what to ask the wait staff; and how to eat
If a restaurant keeps a dish that sounds unappealing to the general public (e.g., “Fermented Bean Curd”), it is likely there because it is delicious and the locals demand it.
Bone marrow toast at St John’s is a good example.
Also, you might want to ask what the waiter recommends. They should have a view and why. But the better additional questions should be: what does everyone else (the regulars) order and why? What’s most under rated and why ? (That might pick up the odd dishes you won’t know) and whatever the waiter’s personal favourite is. If you get a good vibe, you can also see if there is a dish the chef recommends trying.
Order the thing you don’t understand. Avoid the “Roast Chicken” (the safe choice where the kitchen can coast).
(Although of course some days you just fancy a roast chicken that someone else cooks; but your Mum will cook it better…)
You may want to consider the cheaper eats at a well established place. Again, the bone marrow at St John’s while sat at the bar, a lunch sandwich special will gain you quality food and experience at a decent price.
9. The no reservation place.
I chatted to a front of house person, post-pandemic lockdowns, when they moved to a reservation system from a previous no bookings system. This move hit their margins, as they had more empty spots, for longer spans of time. To get away with no reservations, you have to have ideally 100%+ demand over your serving time, this means your covers are always full and more effort has to go into your food to (a) ensure you have 100%+ demand and (b) as your margins are better so you can reinvest profits.
London Examples. Roti King, Kiln, often tapas places e.g. Barrafina, Dean St (Soho), Koya Soho, Dim Sum Duck, often many dim sum places (which won’t seat you until the whole table has arrived!)
10. The Home Cooking Inversion
Restaurants struggle with time-sensitive simplicity (a roast chicken served immediately out of the oven) or premium ingredients that become unaffordable with markups.
The idea: Eat out for things that require scale (deep frying, tandoors), time (24-hour broths), or extreme specialization (dim sum). Cook the simple, high-quality ingredients at home. (Actually, high quality steak is good to do at home).
It’s the same in reverse. Until recently many Chinese homes did not have an oven and Chinese went out to buy roast and barbecue meats from specialist places. Use competitive advantages!
I will do another section of how some London restaurants fall into this category at a later date. In the meantime for London you can check out Vittles; for thoughts on food check out Tyler Cowen’s food specific (ethnic dining) blog and his recent podcast with Soumaya Keynes.
I’ve been drafting some thoughts on AI, but events are happening faster than my blogging time. Below I highlight a summary of some of the camps/clusters of thinking around AI and its implications, which are tentative. For now, I wanted to give you a flavour of the investing week.
There were large stock moves, where $100s bn into the low trillions of market cap value was shifted around. The anec-data coming from the latest iterations of Claude (co-work) and Open AI codex following from the impacts Gemini and previous work have convinced many that Agenetic AI - AI agents autonomously being set and completing complex tasks - is happening now. This will fundamentally re-shape every single business model with some software models are now broken or part broken (eg the seat model of SAAS software; see Ben Thompson), and re-evaluation of which business models might benefit (as costs go down) and so profits are immune or at risk. And which management teams can manage this change.
On one day this week, I had younger members of my team look at their computer screens in a somewhat awe struck manner. The vibe is NOT the same order of magnitude fear that I lived through during the 2008 - 2009 financial crisis but the absolute dawning that the world is changing / changed and that this was week where one set of people - investing market participants - really felt it. Prices are the signals, but these signals herald the narratives and stories we tell ourselves.
There is not a single moment which makes up this change in narrative but a continued cluster of anec-data. Such as:
The broad AI thinking camps, where they cluster and why they disagree
I guess huge caveat…This is a map, not a forecast expect the unexpected….
Camp A: Agent era compounding (capability turns into autonomy)
Scaling keeps working, and “scaffolding” (tools, memory, code execution, browsing, workflows) turns models into agents. Biggest story is diffusion: who redesigns organisations fastest, and where reliability becomes “good enough.”
Camp B: AGI soon-ish, but control is the bottleneck
Autonomy arrives faster than interpretability, evaluation, and governance. Expect intense focus on “scheming”, deception, audits, and enforceable constraints.
Camp C: Current paradigm is not enough (expect lumpy progress)
Scaling helps, but the hard part is robust generalisation: world models, planning, grounding, memory, and maybe embodiment. Progress comes in step-changes, not smooth curves. (This camp fights Camp A about whether agents are “enough.”)
Camp D: Bottlenecks are physical and human (bits do not replace runways)
Even if cognition gets cheap, reality is constrained by atoms: permits, trials, power grids, runways, construction, logistics, institutions, and coordination. AI can help, but cannot abolish physical throughput constraints.
Camp E: Economics and political economy (who gains, who loses, how fast)
The core dispute is macro impact (huge vs modest), distribution (wages, inequality), and market structure (winner-take-most compute/data/distribution).
Camp F: Consciousness and moral status (the swamp, now with measuring sticks)
Not “LLMs are conscious” vibes. It’s theory-derived indicators, structured uncertainty, and policy caution about welfare and manipulation.
Camp G: AI sovereignty and self-reliance (the China posture)
AI framed as a pillar of national strength and tech independence. Driven by state-guided industrial policy, military–civil fusion, and insulation from Western chokepoints. Emphasis on local foundation models, regulatory alignment with state values, and global influence through tech export.
Camp H: Governance and oversight first (the Brussels–DC split screen)
The EU sees AI as a normative frontier—regulated like chemicals or medicine—with formal rules around risk classes, rights, and audits. The U.S. mostly prefers post-hoc enforcement, voluntary commitments, and innovation-first posture. Both view regulation as how you steer outcomes, not just react to them.
Camp I: AI as sovereign diversification (the Gulf thesis)
AI is the post-oil moonshot. Gulf states are building compute clusters, hiring top labs, and embedding AI into national planning. It’s less about alignment or consciousness, more about industrial strategy, control over infrastructure, and becoming indispensable in a post-carbon economy.
Over this week, the Camp A is showing ascendancy. But, I think it is worth highlighting the D and C camps. True, many software models need to change or are gone, and jobs will go and change but the impact into GDP and the physical environment (will this improve bus service in America?) are still on-going debates.
It’s worth hearing Elon on robots, and GPUs in space on the Collison / Dwarkesh podcast. You can skim the transcript. Elon comes across as deep technical knowledge but his thoughts on higher level systems thinking seem to be vibes (simulation theory?). Maybe they are vibes for all of us though…
Current other thought are on the enduring power of live performance and the ability for humans to imagine the lived experience of others.
Also, a couple of studies both meta and individual which show the power of walking for health benefits.
Lastly, from the archive a podcast from Sally Phillips on her acting life and thoughts on disability.
Links:
Fellowships and roles:
“Emerging Scholars is an initiative of the Mercatus Center aimed at supporting innovative classical liberal thinkers in becoming this generation’s public intellectuals and policy innovators. We are seeking to hire Research Fellows for two-year, full-time roles at our offices in Arlington, VA, to start in September 2026.“
CF:
