Weekend Reading #355
This is the three-hundredth-and-fifty-fifth weekly edition of our newsletter, Weekend Reading, sent out on Saturday 14th March 2026.
To receive a copy each week directly into your inbox, sign up here.
*****
What we're thinking.
One thing none of us should be doing is judging this war after less than 2 weeks. The world today judges things in real time as they (and falsehoods) unfold. One of my favourite thought exercises when looking at current events is to imagine what a history book written in 50 years would say. However long it lasts (apart from full on World War three) the swings back and forth will be irrelevant in the end to the outcome. But markets have to adjust daily to constantly guess which way this thing is going to go. A headline here, a video there. Etc. And especially with the oil price being so critically important (and fertiliser, chemicals etc). In the real world we're starting to see rationing in Vietnam, Pakistan and Bangladesh for example and Australia is running out of jet fuel. In the absence of some kind of imminent agreement for ceasefire, one has to assume this is going to get worse economically. Yet if you think that the Americans and their Gulf allies didn't plan TOGETHER for an outcome such as what we see now then I don't know what to say. Military planning takes place in infinite detail with every scenario war gamed. It doesn't necessarily mean things are going 100% as planned. How can it ever in a full-blown war?
The key variable in this war as we see it is whether the US and Israel are able to successfully create conditions for the Iranian people themselves to seize power. History shows us, as Prof Jiang has quite correctly elucidated in his more grounded lectures, that air power alone does not win battles. There is however a scenario in play where US and Israeli intelligence plus unmitigated air access can assist in empowering the people to take over. It would go against history per se but the history books in 50 years’ time would also then write about how up until then it had never been done!
I had a long chat with my LLM about whether Israel's recent campaign to attack security checkpoints in cahoots with local intelligence can be a game changer. This is what it had to say:
"This specific tactic of targeting the Iranian security apparatus with the help of local intelligence is a defining characteristic of the March 2026 US-Israeli military campaign. Israel and the US have focused extensive efforts on degrading Iran's internal police state and security forces—specifically the Islamic Revolutionary Guard Corps (IRGC) and the Basij paramilitary militia—aiming to pave the way for a grassroots civilian uprising to dismantle the regime. Recent developments show that Israel has added a layer of civilian-supplied intelligence to its targeting apparatus. Ordinary Iranians on the ground are reportedly transmitting information about security deployments in Persian through Israeli social media accounts. After the Israeli military vets and verifies this crowdsourced intelligence, they use it to authorize precision strikes. This local intelligence has directly facilitated a new phase of the war: attacks on street-level security units. For example, on March 11, 2026, Israeli drone strikes successfully targeted Basij checkpoints that had been set up across Tehran. According to reports, Israel is deploying a new AI-equipped drone swarm system launched from a "mother" platform, which uses facial recognition and a large target database to precisely eliminate forces involved in domestic repression. By systematically blowing up the regime's security infrastructure, checkpoints, and internet censorship hubs, the US-Israeli coalition aims to clear the streets of loyalist enforcers, neutralize the state's monopoly on violence, and create the physical space necessary for the local population to mobilize and take over."
One must also be able to hold two conflicting thoughts in one's mind. Despite a possible outcome which would reset the landscape of the Middle East and usher in an era of untold peace and prosperity, until we get to the point of change of control, we could have chaos. There are so many outcomes in play here we once again say DON'T BE A HERO (yet).
What we're watching.
After spending almost every moment "monitoring the situation" in the Middle East this week, I took a breather to watch the first few episodes of Mr Mercedes on Netflix. It is based on a Steven King book, which I read some years ago. And it is dark. Very dark. And weird. But it's GOOD. DC
We came across this short video on YouTube about robot legs for retail use in China – as always, there are two way to read this: on one hand, it’s a great enabler for the elderly and disabled, assisting them in walking (importantly commensurate to the effort they put in themselves i.e. not complete dependency) which would arguably help to maintain mobility and independence; on the other, there are obvious military applications, pretty much like something out of a Tom Cruise movie. Would almost argue that the Chinese version looks a lot less unwieldly than Tom Cruise’s. EL
What we're reading.
In a probably horrifying glimpse of what the future might hold in an AI dominated world, RentAHuman.ai is a website offering exactly what its name suggests: humans for rent to do jobs that AI agents can’t do. Everyone knows that you can’t AI your way to fixing a lightbulb or a choked pipe, or cooking dinner. But if this is more than a fluke, then it’s truly a case of humans already willingly surrendering their autonomy to an AI agent, offering their physical world services starting from $10/hr, bookable via API by any AI agent. Browse more than 615k rentable humans here. I don’t know if this is a really bad joke or the beginning of a dystopian future but where I used to say that uber drivers were basically slaves to an algorithm, at least that didn’t seem so bad because at least the customers were human, and someone got a benefit out of it. Try charging an AI payroll tax. Payments currently done via Stripe (stablecoins soon?), but it doesn’t take much imagination to see where this is going. But hey, at least there’ll be jobs, right? It’s a new horror story for the kids – if you don’t run the AI, the AI runs you.
The other role humans play that arguably can’t be AI-ed away is that of legal liability – when something goes wrong, heads need to roll, but heads made of blocks of code don’t count. Does an AI have legal personality? Can it be jailed? Can it be fined? Unfortunately not, so that’s where humans step (back) in. In the world of financial services, that unfortunate job lands on someone like an MLRO (money laundering reporting officer) – a single individual personally responsible and criminally liable for ensuring the anti-money laundering responsibilities of a regulated firm are fulfilled. It looks like that’s about to be the blueprint for every other AI-enabled business i.e. almost every surviving business by the time this AI wave is done. Starting with Amazon. In case you missed this, last month the FT printed an account of how Amazon, in its zeal to push AI agents to ship code faster and cheaper, ultimately led to the AI agent deciding that the best way to improve their code base was to “delete and recreate the environment”, leading to a 13 hour outage on AWS. Amazon’s response now is to put senior engineers (at least the ones left after its massive layoffs) on the hook for new deployments, requiring senior engineer sign off – and their acceptance of liability for things going wrong. At least MLROs won’t be alone now.
On a lighter note, if even possible, given everything happening in the Straits of Hormuz, especially mines being laid by Iran in the Straits of Hormuz, here’s some practice in Sweeping the Straits from an old Windows 3.1 classic. EL
