The idea here is that you take a bunch of best sellers in a particular category and you sketch out the plot line of each book by highlighting a passage and basically summarising it: “Joseph comes home to find his wife murdered”. So you end up with a bunch of these statements.
You then take all that, and use some Machine Learning to essentially generate text the other way. So you sketch out a basic plot line, and the algorithm spits out some plausible template like text that gives it the right structure. Depending on how good it is, it might even be good enough to be called a terrible first draft that just needs a lot of editing.
An extension of it is that it actually turns your plot into a questionnaire to generate the book: “Describe Josheph’s home”, “Describe the murder scene”, “Enter dialog, Joseph”…
Publishing books is one industry that seems ripe for disruption. You can already self publish through a range of digital formats, most notably kindle & amazon, but this leaves anybody developing a children’s book which is meant to be a tactile thing you hold in your hands at the mercy of publishers. The main problem is that if you want to get your book published, and in bookstores, you need reach, and advertising. So I’m wondering if a digital advertising agency, or “creative agency” (e.g. mude.com.au) can hire a guy who’s knowledgeable about publishing and use all the digital advertising techniques to get the book at the forefront of people’s mind, and set up a system that gets the book sellers to order directly from Amazon Print on Demand.
The point is just that they replace the traditional book publishers with a new method that achieves the same goal but passes on a bunch more money to the actual author.
So this is a combination of two previous ideas. One is the smart mirror / mirror mirror on the wall idea where you have an AI like presence in the mirror that can display certain things. The other idea is this concept that recording a message for yourself when you’re up can help you when you’re down.
I’ve just learned about “mirror work” or “mirror play” – essentially it’s about self acceptance. And you stare at yourself and say “The world loves me”. The idea is that you begin to accept yourself. Now I wonder if you can make a smart mirror that can see you, notice your expression, and pop up a version of yourself that is happier, smiling, or says something like “Simon, the world loves you”. It’s technology for technologies sake, but perhaps there is an angle where an AI coach can be “watching out” for you, recognise your mood when you look in a mirror, and give you the prompts for you to recognise and fix the scenario.
So in places like England, it’s entirely plausible to live out in the country and catch a train into say London for work. But the commutes are long. I guess similar to living in Gosford and commuting to Sydney which is 1.5 hours. Now wouldn’t it be great if there was a car on the train that was full of exercise equipment (treadmills, bikes, elliptical trainers etc) that you could use on that journey. That way you can hit the gym and do some exercise every day on the way to and from work. Any electricity you generate for the train is deducted from your gym fees. You’d put a couple of showers each end, ask patrons to respect others by not being smelly bastards, and charge a motza for access!
I don’t know about you, but I lose the little microfibre sack that comes with sunglasses as soon as I take the glasses out of it. I put the case wherever, and it’s never seen again. So this idea is part magic trick, part gimmick, part practical. What we do is that we have that super fine microfibre sack attached to a tiny cord with a mechanism such that after you’ve taken the glasses out of the bag, and as you open the foldable glasses, it pulls the bag into the cavity in one of the arms of the glasses. When you fold it again you can see an edge of the case poking out of the arm of the glasses and can tug it out and wrap it around the glasses.
Take glasses out of case. Open glasses. Case disappears. Magic. There’s probably a bunch of physics getting in the way here. But perhaps if you paired with a professional magician they could build something like this – it is after all a very similar trick to all their others.
The idea was that I needed a way to capture our “big picture” features as a bubble graph with the size being the priority. Perhaps what this tool is is a multi-dimensional trello board where the size of the bubble can be represented in different dimensions.
So imagine this: I’m using Bubblello to map out the big picture features of our product. I might have things like “MacOS version” and “Collaboration”, and on this dimension I’m modelling the demand (or importance) of the feature. So perhaps the MacOS feature is highly sought after so it’s a big bubble and the Collab not so much. And then I can “flip the cube” sort of thing so that I’m now looking at the same bubbles but from a different aspect: From this angle the size of the bubble represents the complexity (or time) of this task. So the MacOS feature might be small and the Collab big.
The idea is that you can create a bubble chart which you can look at from multiple angles. Perhaps one of the dimensions is a Trello like interface where you can categorise them into lists….
This stems from my skin-to-replace-plastic idea. Really what I want to do is encourage those with varied backgrounds to collaborate to make something happen. This is the open source software model: People who are expert software developers spend their time developing software on the side as an open source project. GitHub is central to this (though one would argue that GitHub is a relatively recent phenomenon in open source).
The point is this: For someone who is spending some time developing software that is open and consumable by others, GitHub provides a valuable and free resource for them to do so. Science does not have the same resource. What if there was a GitHub style collaboration space for science? Would scientists contribute their expertise to little side projects with the aim to build a better world?
I don’t know what this might look like, but I imagine it might be a bunch of people contributing to scientific docs and research plans, perhaps booking out research equipment and the like.
The idea here is that the DCEO app is capturing sales, revenue, expenditure, any NDAs and agreements are filed etc, and a we provide a dashboard and chatbot such that an investor can ask questions of a company. “What does growth look like?” -> hockey stick graph of sales over time… “What about churn?” -> add churn to the graph…
DCEO allows you to create a company in a chosen jurisdiction (e.g. Delaware, US), and has an opinionated Shareholders Agreement and Company Constitution. From there it allows you to take actions that are allowed via the agreements and the system manages them for you such that it complies with the agreements set out. For instance, you can do share sales digitally, where the system is sending out letters to the various parties (including preemptive rights etc), you are putting in entries for acceptance, number of shares purchased, it’s basically guiding you through the process similar to how an accountant or lawyer would.
This is probably a bit like stripe atlas, but I think stripe atlas just helps you set things up, as opposed to actually helping you do share issues and things like that.
The idea here is to connect those seeking insurance with those seeking a high risk investment opportunity. Here’s how it works. Those who want insurance sign up to the site and they put up their profile. Age, number of accidents, car etc. Then a bunch of punters come along and they stake $100k of their money, and the browse the profiles and bid to insure that person. If they look like they’re a great driver, competing bidders will drive the price low. The insured person pays their (now lowered) premium, which gets distributed to those that have taken on the risk of insuring them (after we’ve taken a cut). In this system you essentially have rich people wanting to make lots of money “gambling” on whether people are good drivers based on their profile. Perhaps we can make it more open to the people by creating syndicates. So you and a bunch of mates can get together, chuck in 10k, and insure some people. It would make roughly 12% per year, but of course you risk losing it all if the driver makes a big claim.