But you had tutorials each week, and if your tutors thought you weren't doing enough work they could set you exams mid-course called 'penal collections' and if you failed them you could be thrown out. They were rare but definitely not unknown.

It's very hard to predict from a half hour interview precisely who will do well in future, and in fact some people might do better having gone to other universities. Oxford and Cambridge aren't the best choice for everybody - I'm sure being a bigger fish in a smaller pond suits some people better.

I would test myself on the main steps of some of the proofs before exams but that was the nearest I got to memorisation. It felt like if you needed to memorise definitions then you hadn't used them enough. Even 30 years later, not having done maths for most of that time, I could still tell you the most of the definitions and could still do an epsilon-delta proof - they feel like conceptual things rather than memory things.

Perhaps I was in a way doing spaced repetition just by trying to solve lots of problems and looking things up if I didn't know them but it didn't feel like a memorisation process, more a process of trying to really really understand the mathematics.

You can set age in Google Accounts, but they can access pretty much anything when they reach age 13 (which is about the age a lot of these things start to become a real problem) and as a parent you don't have any way to limit that. They don't even need to ask to install new phone apps.

Children also use computers as well as phones and you don't want to stop them using a web browser! You also obviously can't supervise them every moment they are online even if you wanted to.

The hospital only had two rooms suitably equipped for giving birth in a squatting position, so I was lucky to get one second time.

I seem to have struck lucky with a child who didn't want a mobile phone, but he was one of only two or three children a year of 200 at school that didn't have one. I can see that it's hard to deny your child something that every other child has and phones are useful when your child starts becoming more independent. I certainly wouldn't do a long journey without my phone myself. We obviously managed without phones as children, but if your child is going to school on public transport, I totally understand wanting them to have a phone in case of problems.

The apps and websites often have good (or relatively harmless!) parts to them e.g YouTube has tons of brilliant educational videos. There is no easy way as a parent to give them access to the good parts without access to the more harmful parts. You don't want to be a total killjoy and never let them play a computer game (plus playing games with them is fun!). Schools require internet usage for homework. You end up in lots of awkward grey areas. Is watching a video about building logic gates in Minecraft ok? If so, then what about Minecraft videos generally? You can't pre-vet everything they want to watch and then supervise them closely to make sure they don't stray from that.

It feels different from smoking/alcohol because there are so many grey areas and there isn't quite as clear evidence as to what exactly is harmful. Ideally you try and prevent your children smoking and drinking alcohol with education and setting a good example rather your wrath. That's trickier with social media.

I think the current proposals are non-ideal but a bit of a desperation measure because nobody has come up with anything better.

Most apps and sites are terrible in terms of the parental controls they provide - they tend to be all or nothing. E.g. I'd like to have a group for our family on WhatsApp but I don't want my children to be able to join random WhatsApp groups and there isn't any way to have the former and not the latter.

There is also a cost socially, which is hard to navigate as a parent. If everybody is talking about minecraft every break and playing it together in the evenings, then it's hard for them if they haven't even seen the game.

For example, my sister is a translator and she says that checking AI translations is actually harder in many ways than doing a translation in the first place, but the agencies pay less for checking than actual translation.