The incentives for "as long as you're happy with our services, you'll keep paying us" may not be 100% aligned with user interest (see: services difficult to unsubscribe and so on) but they seem much better than the alternatives.
470 karma · joined November 8, 2015
Risk selection is a very driver of prices in health insurance. If you can spread the risk between healthy and sick people, the price goes down. If you can't, the price goes up.
Secondly, having a population that has no health insurance means less prevention, i.e. worse health problems down the road, again driving the costs up
If I understand correctly the biggest savings come from the mandatory nature of the coverage - everybody is covered at least partially by the State, + employees by their employers, + private offers for the others + universal coverage for immigrants and the most destitute. Risk is shared between the healthy and sick people, and prevention can be maximized.
1-1 bots are too close to regular apps, with some advantages and some drawbacks. 1-n bots are in a new, significantly different space
Group bots will be immensely more relevant, in the same way that successful Slack apps often solve group problems more than individual problems.
Oh do I hate this line. But it's a two-way street. I'm very fortunate that for one of my first freelance gigs, I read the contract and specifically pointed out that I would be using personal/open sourced code snippets, and that I might want to reuse/open source some of the code I produced for them (only snippets with general interest).
They agreed without a flinch, and thanks to that experience, I'm going to hold every other employer to the same standard "I never had a problem with this request with any other employer, why would you want to refuse?"
And then when pregnant women ask for seats out of exhaustion and embarassment, they hear "no, that's my seat", and it's the elderly lady two rows farther who gets up.
Media is hyping the AI part because it is catchy, but using mostly/only quick replies and buttons gives a much better UX IMHO. So, fully agree with #4.
There are only a handful of cases where text actually reduces friction instead of increasing it. Entering a date, for example : "in 1h", "tomorrow morning" or "next friday" is really faster than other input types.
For most other cases, Messenger bots just got access to webviews. Really under-utilized so far, I think.
this is not very hard to imagine
Do you realize you used 'just', talking about countries that were at war with each other for their entire history? When I said "we don't realize", I was talking about you
The shared attachement to democracy, peace, to solidarity versus exacerbated nationalism, refusing death penalty, of some version of social democracy in the broad sense of the term, is much stronger than what divides us.
I hope the best for Europe and UK on their separate path. Given the mess, a clean cut is probably in the best interest of both sides
I always feel like messaging platforms bots and AI get presented as one and the same, but I think they are very, very different things. Bots = a way to provide a service on platforms on which users spend 60% of their time. Absolutely fundamental trend. AI = a way for these apps to communicate using natural language. Insignificant at this stage
If these guys confirm their trajectory, it would prove that the sharing feature recently added by the Messenger team was a key element. I would have thought that this sort of explosive virality would have to wait for bots allowed in groups, but I am very happy to be proven wrong. This is very good for the whole ecosystem.
As far as finding an interesting problem to solve. PG explains in a follow-up essay, I think, that most people get used to suboptimal situations. We, as human, are very good as rationalizing painful situations, processes, etc, as normal and good. You need to take a step back when something is long, boring, repetitive or annoying. Is it really necessary? Couldn't it be automated? Question your assumptions :-)