The painting and the love affair are interesting, but, as a landlord, I found myself wondering: How does an apartment remain untouched for 70 years? How was the rent paid without interruption? How were rent increases handled (or did the rent remain constant for 70 years?)? How did the landlord not have to access the apartment when the other parts of the building/infrastructure were renovated/upgraded over the decades? Also, Pigalle was/is a bit of a rough neighborhood, so why wasn't there a break-in over those 70 years?
The people who get their coffee to go and the ones who sit at the tables are different segments (although there is some overlap).
Also, the sitting customers can be segmented as well: most of them will sit down for 15-30 minutes and leave, but there are some who will stay longer. It would be difficult for someway staying 4 hours to spend as much as 8 people staying 30 minutes each, so that's why djt talks about the need to turn over tables.
Good to hear there's a HN meetup in Toronto! Any idea why it's set up as a private group? All the other HN meetup groups I checked on this page are public.
I think that if the content is relevant then there's nothing wrong with self-promotion. I looked at the comment history and he linked to tldr in 7/10 of his last comments over the last week. That doesn't seem excessively repetitive, although a disclosure might be appropriate.
I've found MT results quality to be a problem. 95% of the workers are great, but you do have occasions where you get a bot completing the task or someone being deliberately careless in the hope that the HIT will get automatically approved. To combat that, I give the task to different workers until 3 answers match and that works well. It's worthwhile once you've automated, but that's worthwhile only if you've got a fairly large number of tasks.
There's an interesting book called After Virtue by Alasdair MacIntyre that does a good job of going through how these confusions over language arise. I haven't read it for years, but I liked how it explained how and why the language of philosophy has changed over the centuries.
I've been a landlord in Ontario for 10 years and had a sublet request a few months ago. In Ontario, there are 2 stages: a tenant has to make a general request to sublet, and then a specific one. They can break the lease if the landlord refuses at either stage, but otherwise they have to provide a legitimate replacement tenant (as opposed to putting a random person forward as a pretext to break the lease) or stay until the lease ends.
Of course you could just stop paying rent and your tenancy will be terminated much faster and you won't have to go through the hassle of finding someone to sublet.
Edit: The above applies in the case of legitimate sublet request - a two week AirBNB guest wouldn't count.
I had an hour-long chat with a guy from Canada's Federal Economic Development Agency who I met at a startup event and still keep in touch with. I suggested to him that they should copy the Startup Chile program to get entrepreneurs into Canada and create some buzz. He politely insisted that Canada's current programs are innovative and adequate. I know this is just one anecdotal data point, but it fits the cliche of the don't-rock-the-boat/cover-your-ass government employee and I wasn't too surprised. I wonder how Chile was able to overcome government inertia and get their program off the ground?
My understanding is that this comment is stating that there is a trend towards products having more or less identical functional features, but at different price points and that they will only be distinguished from each other by marketing. I would argue this has always been the case, from the day people first started trading goods instead of making everything for themselves. Whenever there are competing, nearly-identical goods in the marketplace, someone invents ways to distinguish them based on pricing, logos, fatuous claims, etc., but that happened a long time ago.
An interesting thing about FMCG companies and profitability:
BigCo selling a bar of soap for $4 has a similar contribution margin to GenericCo selling a bar of soap for $1. Why? BigCo has larger overheads (i.e., they have to promote and advertise, do consumer research, etc.). In fact, BigCos don't make much profit from categories like soap and laundry powder, but they need them to cover overheads.
Can you give us some examples of who has done this and how they avoided getting caught on Mechanical Turk? I've had legitimate HITs removed for apparently violating policy recently, so in my experience AWS is fairly vigilant in looking out for HITs that attempt to juice SEO or promote.
I guess it could work if you abstracted it enough so that it's not apparent to the worker, but it wouldn't be trivial to execute. Using VAs on Odesk might work better.
There's no way to guarantee you won't behave like that (I don't specifically mean you, but anyone in a similar position). People have the best intentions, but things change as time goes on.
In fact, charging money might make it all the more insulting if/when a company later does something to piss off its developer community.
Platform integration is a risk, but I think it's better to take that risk than not pursue an idea because some day the platform will cut you off from the API.
It's basically an affiliate business model: you get 5-10% of the cost from the doctor, hotel, and airline (depending on your negotiation skills). $500K revenue means $25-50K in true gross revenue for the company. Their net profit is $3-5K on that revenue, depending on how efficient they are.
I put my business on the back burner for 2 reasons:
1) 5-10% of the cost of the package is your true revenue before overheads, so the it's not very profitable once you consider your expenses.
2) It's a high touch customer service experience, so you can't scale it profitably (people will not buy an operation via e-commerce like they would an e-book: there's a lot of personalized contact and questions, examination of records, etc.).
It's possible that instead of going after happy DB users, they'll go after current non-consumers by offering them something that meets an unfulfilled need. I can't provide a specific example of what that use case would be, but the addressable market is large and diverse enough that a company could make some dough with a $10/month subscription model if they find that niche.
I agree that $750k to go head-to-head against DB would be futile.
What's a typical percentage rejection rate for your tasks?
I find that for some classification tasks my rejection rate is around a third (this includes careless/sloppy answers, so the rejection rate of honest answers is lower).
Even though there is always a right answer, some answers are almost correct, but they get rejected and the worker's stats are affected.
In addition to the iterative approach, you can also offer a base rate and accept all the tasks, even if they're not quite perfect, then pay a bonus to the workers who really nailed it.
That way you're not rejecting the workers who are making an honest effort, but the better performers are being paid accordingly.
This article reminds us that you can't just look at historic growth rates and extrapolate economic prospects. Environmental sustainability, the effectiveness of political institutions, and demographics start to matter more and more once you get to a certain level of prosperity.
Also interesting is the effect of demographics on technological development. For example, one of the factors driving robotics technology in Japan is probably the size of their aging population and the need to take care of them. Another case of 'necessity is the mother of invention'?
I think that as consumers our reflex is to send a text when a missed call might be faster/more convenient. For example, if I'm picking up a friend, I'll text "I'm here" or "here", whereas I could simply make a missed call and my friend sees that it's me on the display and goes outside because she's expecting me.
Making a missed call would just not occur to me, but when I moved to Lithuania for a few years it became a habit because that's what everyone does over there (and in other places where there is a missed call culture and people have several cell phones to take advantage of cheap/free calling plans within a given carrier network). It's faster by a few seconds than sending a text or calling the person and telling them.
Once you're in the missed call mindset, you see all kinds of situations where you can use it, but until you get in the mindset, the reflex will be to send a text or make a call.