• The Entrepreneur Forum | Startups | Entrepreneurship | Starting a Business | Motivation | Success
  1. Join 40,000+ entrepreneurs
    who are kicking butt and
    winning their dream life.

    Unscripted™ Entrepreneurship:
    A Business That Pays More Than Money, It Pays Time.

    "Fastlane" is an entrepreneur discussion forum based on The Unscripted Entrepreneurial Framework (TUNEF) outlined in the two best-selling books by MJ DeMarco (The Millionaire Fastlane and UNSCRIPTED™). From multimillionaires to digital nomads, the forum features real entrepreneurs creating real businesses.

    Download (Unscripted) Download (Millionaire Fastlane)  Register
    Registering for the forum removes this block!

EXECUTION Data on how my habit building app improves several motivational variables (Scientific Study)

Discussion in 'Progress/Execution Threads' started by MarcoSto, Apr 30, 2018.

  1. MarcoSto
    Offline

    MarcoSto Contributor Read Millionaire Fastlane I've Read UNSCRIPTED

    Messages:
    21
    Likes Received:
    31
    Joined:
    Apr 28, 2018
    Gender:
    Male
    Location:
    Germany
    Rep Bank:
    $368
    Hey fellows,

    As promised in my intro - you might want to check it - here is some data of my yet unpublished psychological long term study on habit building.

    15 months ago, I had...basically nothing but an idea for a study for my doctoral thesis: I wanted to help people build study habits to reduce motivational interference while studying. Motivational interference is a phenomenon that occurs when you basically want to do something else than the current action - we all know that. You study for a test, but want to watch netflix. That destabilizes your learning process (you get into a bad mood, have less persistence, think about other stuff, switch tasks etc.).

    I learned app development with udemy and some learning apps and was able to create a simple app that helped students define and track a simple (5-20 minutes) study habit and would give them feedback on how strong their habit has become psychologically (it might read like it was not so difficult, but it was a tough grind). Took me 3 months from zero to app.

    Then I contacted several professors and asked them if I could advertise my study in their lectures. After another 4 months, I got around 100 students that took part in my study, which was basically: Use my app for 6 weeks and lets see what happens. The participants could choose their own, individual study habit. Some examples are: lecture summarizing, reading literature, language learning etc.

    I collected data on how the habit developed and got more automatic over time. When I first analyzed the data, my feedback loop triggered like crazy. Check this (excerpt of my results mail that I sent to the participants of the study):

    The key result: Habits work an make everything better :).

    Check out the graph I attached for you. It summarizes the most important trends. Means over all participants over time are plotted.

    On the x-axis, you can see the number of habit repetitions and on the y-axis, you can see the different outcome variables. For example, at habit repetition = 40 you can see, that the average score for motivational interference was approx 1,9 (vs. 3,6 after the first habit repetiton).

    The variables explained:

    Scale 0 (= not true at all) to 10 (= absolutely true)

    Automaticity: Habit strength. Describes how automatic a habit action feels. This measure basically shows the habit building process. You can see a continuous, linear growth, which means that the performance of the habit action actually got increasingly easier and more automatic for the participants.

    Motivational_Interference: Motivational destabilization during the habit action. If you are in a bad mood, do not persist long and are easily distracted, this value is high. The stronger the habit, the more resilient the habit becomes against motivational destabilization.

    Awareness: Awareness during the habit. This measure consists of the two facets acceptance and present-orientedness. A high value here means that one is in the present and feels balanced while performing the habit. Participants became more and more aware while performing their habits with increasing habit repetition.

    VA_intrinsic: "I like doing my habit.". With each habit repetition, participants actually liked their habit more and more - a positive spiral.

    MC_should: Should conflict. The feeling one should do something else (e.g. going to the gym, do chores) while performing the habit. Should conflicts stayed constantly on a low level during habit performance.

    MC_want: Want conflict. The feeling of wanting to do something else than performing the habit (#netflix). The more often you did your habit, the less you feel the craving for other actions that are more pleasant to do - this is huge.

    But, check out the graph for yourself.

    Habit_graph_Xampler.png

    The whole thing took 15 months to complete (idea, planning of the study, app development, testing, getting the participants, data analysis, presentation at the psychological faculty). Now, I'm at the point where I want to monetize this app with a freemium/subscription model. I will describe the value skew and traffic strategy in detail in another post.

    If you have any questions about this, feel free to ask - I love talking about this.

    Keep up the good work.

    Marco
     
  2. ApparentHorizon
    Offline

    ApparentHorizon Gold Contributor Speedway Pass

    Messages:
    667
    Likes Received:
    1,884
    Joined:
    Apr 1, 2016
    Location:
    Greenville, SC
    Rep Bank:
    $6,615
    So your app tracks how many times you do a certain task, in order to maximize repetition. Thus, by being consciously aware of how many times you do said task, you're forming a long term habit. And by the time you reach 42 repetitions, it has solidified.

    Taking that feeling of...

    "I know I should be studying, and I'm feeling agitated that I'm watching or thinking about Netflix, therefore I can't concentrate on anything. And I don't want to do anything now."

    ...and helping you attain the goal of becoming more productive, by giving you a roadmap.

    Am I understanding that correctly?

    If so, what are the mechanisms at play here?

    We know repetition is key: The more you do something, the more efficient your brain is at doing said task. (IE the 4 stages of competence)

    But we also know how hard it is to stick with a new pattern long enough for it to become a habit.

    After all, why let our brain use all of that energy when we can still have food and shelter without the extra effort.
     
    MarcoSto likes this.
  3. MarcoSto
    Offline

    MarcoSto Contributor Read Millionaire Fastlane I've Read UNSCRIPTED

    Messages:
    21
    Likes Received:
    31
    Joined:
    Apr 28, 2018
    Gender:
    Male
    Location:
    Germany
    Rep Bank:
    $368
    The app helps defining a new study habit (nudging the user to create a habit that is very likely to succeed by implementing several psychological techniques). Here is an excerpt from my exposé that describes the habit definition process - 4 steps:

    ****

    1 What. The participants are asked to describe their new habit here: What shall be the new habit? The two constraints are:

    • It must be for the university
    • It must be a new habit

    To clarify this step, the participants are given examples like reading relevant literature and summarizing a lecture.


    2 When. The participants are asked to describe when they want to perform their new habit. The two constraints are:


    • It must be performed daily
    • It must be pegged to another daily activity

    To clarify this step, the participants are given examples like after breakfast, after coming home from university and before brushing teeth.


    3 How long. The participants are asked to describe how long it is approximately going to take to perform their new daily habit. The two constraints are:


    • It must be at least 5 minutes
    • It must not be more than 20 minutes

    To clarify this step, the participants are given examples like 10-15 minutes and 20 minutes.


    4 Goal. The participants are asked to set a goal for the daily habit repetition. The two constraints are:


    • It must be attainable within the previously defined time frame
    • It must be measurable

    To clarify this step, the participants are given examples like read 5 pages and summarize ONE whole lecture.


    ****

    You DO NOT have to be consciously aware of the action in order to form a habit. A habit needs 2 things:

    1. A stable context
    2. An automated cue-response chain

    That means: Reading literature for university at different times in the day and at different places? NOT a habit. Doing so every day at the café after your second coffee - more likely to become a habit over time (if the reading behavior is somewhat automated).

    After each habit repetition, you answer some questions about your behavior and experiences during the todays habit repetition. The app then calculates an automaticity score and gives realtime feedback on habit strength (i.e. automaticity).

    The habit helps more less with the "I know I SHOULD be studying.", but more with the feeling DURING the study habit: "I so desperately WANT to do something else." by helping to form a stable study habit.

    Repetition is key, you are correct. Thats why it is so important to keep the habit very easy, short and simple in the beginning to ensure that the first 14 repetitions are really performed daily. After that, the majority of participants reported significant increases in automaticity.

    A mechanism that drastically improves the likelihood of actually doing the habit in the beginning (before it has become more automatic) is implementation intention: By formulating the habit in such a clear manner, pegging it to another daily action with a clear goal for the habit repetition and a specific (short) duration. Implementation intention paired with a relatively easy task is your entry ticket for a new habit that you can then expand and use for more difficult and more time consuming tasks. After the first 14-40 repetitions, automaticity takes over and it will be far easier to start, let alone finish your habit once you started it.

    There is no magic involved. Just statistics, probabilities, psychology and TINY steps.
     
    ApparentHorizon likes this.
  4. ApparentHorizon
    Offline

    ApparentHorizon Gold Contributor Speedway Pass

    Messages:
    667
    Likes Received:
    1,884
    Joined:
    Apr 1, 2016
    Location:
    Greenville, SC
    Rep Bank:
    $6,615
    I should clarify:

    Say for example you're trying to form a new productive habit. Most people (myself included) will just try to take the same action every day. Without thinking about the number of times I've completed it. Nor how many times it takes to turn it into repetitive task that becomes second nature.

    On day 11, I'm not exactly sure where I am in that process. So by giving the user a number, let's call it 30 times on average (based on Duhiggs' book Power of Habit), I'm able to better focus on the end result.

    (The problem with his book is that he says do something for 30 days and it becomes habit. But the points you're touching upon provide so much more context)

    Sounds like anchoring

    So a follow up question, in reference to the tiny steps...

    The saying is that perfect practice makes perfect. And the mistake I make when I try to force a new habit is going all in. Meaning being very detailed and exact in the steps, which is exhausting.

    In those 20 minutes of studying and during the first 14 repetitions, how important is it to get the task right?

    You sit down for 20 minutes, but you fidget for 10. And manage to read the other 10.

    Does this eventually translate into 20 full minutes of studying by day 40?
     
    MarcoSto and Argue like this.
  5. Leo Hendrix
    Offline

    Leo Hendrix Bronze Contributor Read Millionaire Fastlane Speedway Pass

    Messages:
    439
    Likes Received:
    257
    Joined:
    Apr 15, 2014
    Gender:
    Male
    Location:
    WWW
    Rep Bank:
    $1,457
    Do you have any copies of your app available for download and use now?

    I would like to test it out - sounds like something useful and beneficial.
     
    MarcoSto likes this.
  6. MarcoSto
    Offline

    MarcoSto Contributor Read Millionaire Fastlane I've Read UNSCRIPTED

    Messages:
    21
    Likes Received:
    31
    Joined:
    Apr 28, 2018
    Gender:
    Male
    Location:
    Germany
    Rep Bank:
    $368
    Glad to see you are so into this topic :).

    Yes, the feedback you get is a crucial factor and helps to keep you going through the desert of the beginning and shows you later how far you have come in terms of automaticity. So, yeah, you are absolutely correct about the conscious-making of the habit forming process in that sense.

    Two concepts need to be differentiated here.
    Anchoring: The influence that especially numbers that are simply activated in our brains (e.g. by simply saying them) have on our thinking and decisions. In the field of the psychology of habit, a context activating a habit would be called triggering: A cue (trigger) that activates an action. The more automated this cue-response scheme has become, the more likely this certain behavioral pattern becomes in certain contexts. Let's assume you go into the cinema. This might trigger "eating popcorn". The stronger this popcorn habit is, the more likely are you to show the behavior of buying and then eating popcorn while watching the movie in the cinema.

    Great question! So its perfectly fine if you, by day 1, do 10 minutes of fiddling, not really getting into it and taking half of the time to get the material right etc. That is one of the many things that will have a steep automatization after the first few repetitions and you will see a lot of progress. If you cannot reach your goal, then its too high to begin with. But it is important - especially in the beginning - that you reach your goal and thus reinforce your behavior. If there is no positive feedback for your behavior (e.g. the feeling of "yeah, I successfully completed my learning habit for today"), no habit will form.
    You should not set a time goal (e.g. learning for 20 minutes), but a measurable, attainable goal. E.g. reading at least 4 pages and quickly write down the key takeaways. If you could not finish your first habit rep, or it felt arduous, set a lower goal: 1 page, 1 key point. You have plenty of time to grow your habit.

    I have - obviously - several habits. Writing scientific text is very hard for me and is as fun as eating glass, in my opinion. However, this needs to be done. My approach to write my journal article started like this:

    Look at the document for ONE MINUTE. Thats it. Otherwise, I would not have opened the document, but would just have procrastinetflixed this problem away.

    Guess what happened. I looked at the document (there was some text already) and thought of the next step I needed to take: Write an outline. Ok. I wrote "Outline", saved, felt okay about my habit (I did more that I needed to do by writing a whole WORD).

    Next day: I had to look at it for at least 2 minutes and I would have to write one point for the outline. Guess what happened: I wrote several points for the outline and started looking at articles of my colleagues to compare their structures and figured out some points for my outline.

    Next day: 5 minutes....

    You get the gist. This is a very difficult habit to build, and as I am now in the "10 minutes and write something down"-goal-space, it becomes increasingly difficult to increase the goal difficulty. But thats okay. I am almost done with the first draft of my outline within a week.

    If you cannot take steps continuously, you need to make them even smaller. This is a minimal intervention paradigm that a lot of clinical psychologists use for behavioral change. I just adapted it to habit building. This will be included in the app as a kind of continuous, intelligent adaption.
     
    ApparentHorizon likes this.
  7. MarcoSto
    Offline

    MarcoSto Contributor Read Millionaire Fastlane I've Read UNSCRIPTED

    Messages:
    21
    Likes Received:
    31
    Joined:
    Apr 28, 2018
    Gender:
    Male
    Location:
    Germany
    Rep Bank:
    $368
    Hey,

    you just triggered my feedback loop :). Unfortunately, the new version of the app is currently in development and there is no official version in the appstore (the study was conducted via the beta test program of apple).

    As soon as the app is in the store, I will obviously inform you guys and discuss the mechanics, business model and everything you want to discuss.
     
    Leo Hendrix and ApparentHorizon like this.
  8. ApparentHorizon
    Offline

    ApparentHorizon Gold Contributor Speedway Pass

    Messages:
    667
    Likes Received:
    1,884
    Joined:
    Apr 1, 2016
    Location:
    Greenville, SC
    Rep Bank:
    $6,615
    Lao Tzu said, "Mastering others is strength. Mastering yourself is true power."

    Plus you radiate energy, always fantastic to see people excited about their craft :D

    Yeah exactly! The other part of the puzzle is doing the right work. For example, you actually gave people examples of what to do. Not just "form a habit." But form one to study by doing XYZ.

    I setup a new site for a client recently and one of the contact forms had a field, "Do you have any requirements?" It had less than 10% fill in rate. Whereas, when examples were added, it shot up to 71%. The majority of which were not the examples given! (Small sample size, but significant change)

    Oh that makes so much sense now.
    Anchor = thought
    Trigger = action

    But they're not mutually exclusive.

    Internal vs External.

    Have you read up on Kaizen?

    The Japanese philosophy of continual improvement. Starting out with just 60 seconds of an action, is how I've heard it described. Just like you began with the writing. (What do you think makes it hard for you to write the papers in the first place?)

    I did this with video games. Going from 12 hours a day playing MMOs to only a couple a week now. Slowly playing 5-10 minutes less every week or so. But it also increased efficiency.

    Games are effective at providing you with goals and giving you that strong dopamine release. So getting the same release in 5 hours vs 10 is a testament to how effective we can be.

    Jordan Peterson did a Q&A recently, describing the most successful people (Bezos and Musk) are not just smart. But incredibly efficient with their time. Whether they cut out nonesense and/or they cultivate the most efficient habits possible.

    By the way, I'm not sure if you've seen this post, but it might be interesting if you don't already know about most of these tricks: GOLD! - Guru Cults Exposed: The Tactics "Experts" Use To Pull You In & Suck You Dry
     
    MarcoSto likes this.
  9. MarcoSto
    Offline

    MarcoSto Contributor Read Millionaire Fastlane I've Read UNSCRIPTED

    Messages:
    21
    Likes Received:
    31
    Joined:
    Apr 28, 2018
    Gender:
    Male
    Location:
    Germany
    Rep Bank:
    $368
    A quote I'll keep in mind. Like it.

    Thank you, very motivating to hear that :). What's your craft, btw? Webdev?

    Yeah. It was a tricky thing to give them guidance and a broad direction but still make sure everyone had enough space to define a self determined "that's MY habit" kind of habit action.


    Awesome how such tiny changes can do some magic, right?


    U seem to know some concepts - like it. Yes, I had my hands on an old Kaizen-book some years ago. I did not dig into it deeply, but I recognized when I did something kaizen-like and was like: No step is too small and the improvement cycle will always keep on rollin' - kaizen style. So I used it for self talk sometimes, even though I just know the basic concept.

    I was at 9 hours of DOTA 2 when I had to write my master thesis back in 2014. I'm on 2 hours HOTS in the evening now. That's about okay for me. But, I know the struggle - been there. Can be a vicious cycle.


    I like Peterson a lot. Where can I find the Q&A? I listened to some podcasts that featured him. One with Jocko Willink was pretty neat.

    Thx, I'll check it out.
     
    ApparentHorizon likes this.
  10. ay47
    Offline

    ay47 Contributor Read Millionaire Fastlane

    Messages:
    54
    Likes Received:
    90
    Joined:
    Mar 29, 2018
    Rep Bank:
    $1,063
    How would I would go about getting your app? I need better habits :p
     
    MarcoSto likes this.
  11. MarcoSto
    Offline

    MarcoSto Contributor Read Millionaire Fastlane I've Read UNSCRIPTED

    Messages:
    21
    Likes Received:
    31
    Joined:
    Apr 28, 2018
    Gender:
    Male
    Location:
    Germany
    Rep Bank:
    $368
    I‘ll definitely let you guys know when the app is in the AppStore. I’m working on it every day and hope to get it online in like 4-5 weeks.

    But that should not stop you from building a new habit right away. Once you understood the basic principles, it will get easier and easier to develop good habits.

    What is one thing you would like to learn, improve, get done by the next 2 months?
     
    ApparentHorizon likes this.
  12. ApparentHorizon
    Offline

    ApparentHorizon Gold Contributor Speedway Pass

    Messages:
    667
    Likes Received:
    1,884
    Joined:
    Apr 1, 2016
    Location:
    Greenville, SC
    Rep Bank:
    $6,615
    Programming, with webdev being a small subset.

    Oh yeah I remember part of that one.

    He has a YT channel with some lengthy videos. Great listen while doing mundane work.


    View: https://www.youtube.com/watch?v=4GcU9LjuVOo
     
    MarcoSto likes this.
  13. MarcoSto
    Offline

    MarcoSto Contributor Read Millionaire Fastlane I've Read UNSCRIPTED

    Messages:
    21
    Likes Received:
    31
    Joined:
    Apr 28, 2018
    Gender:
    Male
    Location:
    Germany
    Rep Bank:
    $368
    Thx, I’ll check it out :).

    Oh yeah...you programmers have a lot
    of work in these days, right?

    I like programmers...they often think in a very clear, rational, logical and structured way. Comes with the kind of work that’s associated with it, I guess.

    Are you freelancing, employed, entrepreneur, something in between?
     
    ApparentHorizon likes this.
  14. ApparentHorizon
    Offline

    ApparentHorizon Gold Contributor Speedway Pass

    Messages:
    667
    Likes Received:
    1,884
    Joined:
    Apr 1, 2016
    Location:
    Greenville, SC
    Rep Bank:
    $6,615
    Own biz with a few VAs to do the repetitive and boring stuff.
    Depends on the type. Things like web and app dev drove down prices hard b/c of the Asian markets. There's still money to be made, but it's a fraction of what it was 5-10 years ago. (Unless you make your own projects of course)

    The real money is in enterprise SaaS, Analytics, and Artificial Intelligence.

    AI is super fascinating. We don't even know how it works half the time! What companies are pushing for right now is AI that tracks buying behavior online.

    Forget what Facebook was giving to Cambridge Analytica. They're going to know when you woke up by the pitch of your voice, and pop up a coupon for a hamburger.

    I think someone like yourself could add huge value in this sector. Analyzing what the black box spits out and creating a service around your findings. It wouldn't be hard to do. Team up with a few people running their own ecomm stores, and offer behavior analysis.... anyway, I'm getting ahead of myself lol

    "Programming is ordering chaos"

    It's a lot easier than most people think.

    If you look at a painting and say, I could never do that. Well, you could learn some of the basics, but you'll always feel that something is off.

    With coding, it's much easier to put a site together if you're an artist.

    IE: why so many people are in web dev. In fact, you'd be better off artistically inclined in this situation.

    Are you putting together your app or having someone else do it?
     
    MarcoSto likes this.
  15. MarcoSto
    Offline

    MarcoSto Contributor Read Millionaire Fastlane I've Read UNSCRIPTED

    Messages:
    21
    Likes Received:
    31
    Joined:
    Apr 28, 2018
    Gender:
    Male
    Location:
    Germany
    Rep Bank:
    $368
    Nice. Seems you are already going the right way :). I'd love to work with a VA by the end of the year, too. To be honest: Thats something that I put off many times, because I think its not a very easy thing to find a VA who is trustworthy and get the workflow going. Actually, I have a VA for my business (I just remembered :D), but she basically only handles my business related letters.

    Yeah. I worked with some pakistani coders two years ago for a SQL-Thing. It is what it is so often: You get what you pay for. It was horrible and one day, they simply stopped answering. Meh. Thats the problem with the trend: They set new price tags to coding tasks, but (at least some of them) dont deliver but the customers get attached to the low price and dont want to play you 70$/h anymore for your quality work.

    Oooooh yeah. I think so, too. Are you into it? Did you do something with AI, yet?

    I was very interested in this and took some udemy courses in machine learning, AI (and I'm, by craft, relatively cool with analytics - at least with datasets that are not too big (<10k cases)). I cobbled together some ANNs and made some simple predictions.

    I wanted to train a RNN with my longitudinal data and then implement it into my app to tailor/optimize habit building. The participants would get optimized goal setting instructions (i.e. "too optimize the probability of successful habit formation, you should lower your current habit time from 20 minutes to 12 minutes".). But for the first version of the app, I will use an old school regression model to derive recommendations. But an RNN would be such a nice value skew for the app :)...

    AI is really sick, you are right. Its something we all need to understand in order to understand the society that is going to from in the next few years. So much already runs on AI but people just do not recognize it, yet. Its useful, but self developing AI that can write its own code might be a serious issue for security.

    Correct. I heard of a voice analyzing software that puts together several thousand data points from 10 seconds of you talking and can tell an absurd amount about you, your mood, age, gender, intentions, IQ etc. It's not scifi anymore...

    Thanks! That's what I try to do in the realm of longterm behavioral change - or habit building :). A lot of psychologists run on the 1990s OS and did not yet understand what can be done if we bring together psychological findings and paradigms with modern tech (AI, analytics, Apps etc.). I see myself as a translator in this realm: Psychologists find out some interesting sh*t or developed a really well working behavioral model, but then simply publish it in a scientific paper and it is never used in the real world. Thats a waste of research.

    Yeah. What I love about coding is: It's always your fault. While this may sound a little frustrating at first glance, its pure self efficacy. You can control nearly everything and if something does not work, its because YOU did not follow the logic. It's always outfigurable.

    I taught myself last year using some apps to understand the basic concept of coding. Then I took a udemy course and cobbled together my first spaghetti-code app. BUT: It worked :). From "What is a variable?" to working app it took me 3 months. I'm still like in the top 15% of worst iOS coders out there, BUT: I somehow got it to work. That was really very empowering. But on the other hand, I realized what I do NOT know and that let my respect for coders grow. There is so much you have to learn and you have to constantly adapt. So...a lot of kudos to all the coders out there from my side.

    For the AppStore version of the app I now have some support by a not very cheap, but good coder that will help me with some coding tasks and give me advice for the launch and stuff. But we will both work on branches of the app as a team. I need to understand the app entirely to implement my psychological model exactly the way I want it to. So I want to understand every line of code of the app. But in order to speed up the dev process and increase my commitment, I decided to get some paid help by a professional.

    Do you have currently own software projects in progress or are exclusively delivering for your customers?
     
    ApparentHorizon likes this.
  16. ApparentHorizon
    Offline

    ApparentHorizon Gold Contributor Speedway Pass

    Messages:
    667
    Likes Received:
    1,884
    Joined:
    Apr 1, 2016
    Location:
    Greenville, SC
    Rep Bank:
    $6,615
    I have a fantastic lady from the Philippines, who speaks better English than some of the locals here lol And 2 great coders both from India.

    It's all in the way they answer questions.

    For example, write a CSS class that vertically aligns a table and colors in every other row with a transparent gradient.

    50% will answer with, "Yes I can"
    30% will write out the code
    20% will write the code and explain it
    Then every so often, 1 will write, explain, and offer alternatives and why they would be good for specific situations

    It helps to know the subject at hand when interviewing. Otherwise, if you're delegating a subject matter you're not familiar with, an in house local would probably be better.

    hmm. Now that's an interesting concept. The others are easy for us to pick up on....but intentions. That's more body language reading than voice.

    That's an interesting idea! taking research papers using "normal" methods of data gathering and analysis. Then feeding them through an AI.

    Or like Facebook's AI that started talking to each other in a language they invented. They promptly shut that down lol

    That's the way I do it too, and so do a lot of coders. You just scrap together until something works. Then you launch. If it works well, hire someone to optimize it or rebuild it in a better language for the specific task.

    That way you can focus on building the business around it and raising money.

    Both. Right now I'm developing stuff in house for my own use. And if it's good, exporting them to Enterprise level companies.

    I'm prototyping a FFNN for simple prediction. RNN are much more complex, and computationally expensive.

    Check out FFNN for your own app, it can get you a good part of the way there that an RNN would. But much easier to implement. Then once it takes off, hire someone to optimize it.
     
    Siberia and MarcoSto like this.
  17. MarcoSto
    Offline

    MarcoSto Contributor Read Millionaire Fastlane I've Read UNSCRIPTED

    Messages:
    21
    Likes Received:
    31
    Joined:
    Apr 28, 2018
    Gender:
    Male
    Location:
    Germany
    Rep Bank:
    $368
    It's an art to find good peeps, keep them and delegate effectively. A lot of people underestimate this. And I have to get better at this, I guess. But I'm working on it. :)



    Well. I dont think its that difficult, actually. An AI could analyze the content of what you are talking about, then cluster the topics and calculate probabilities for lying in certain areas (e.g. when talking about things that are associated with 'money' or the persons self) by anomaly detection in speed of talking, voice pitch, use of vocabulary, stuttering, filler words and the other 5k datapoints the AI has at its disposal. An intention to be derived by that could be "The speaker wants to hide his true self.".


    Thx! I have a feeling, though, that everything becomes interesting if you combine it with AI :D


    heard about that - that was really a bit scary actually.






    Nice. Seems like you have some things going for you. Hope, it pays out as planned :)

    I'm not sure how clever it is to try to implement longitudinal data in a FFNN. Actually, for a first step, I'll simply use an even simpler version of a FFNN - a simple Linear Mixed Model. It accounts for inter- and intraindividual variance and for development over time while integrating personality variables. It's not as flexible as a NN, but its interpretable and thus its easier for me to derive behavioral interventions. But once the app with this (compared to NNs) primitive statistical model is up and running and generating data, I can dabble with and feed NNs, I guess. Thats at least my rough plan. For the customer, not much is gonna change. In one version its gonna be: "Your habits will be statistically optimized." and in the other version it will be: "Your habits will be optimized by AI.".
     
    ApparentHorizon likes this.
  18. ApparentHorizon
    Offline

    ApparentHorizon Gold Contributor Speedway Pass

    Messages:
    667
    Likes Received:
    1,884
    Joined:
    Apr 1, 2016
    Location:
    Greenville, SC
    Rep Bank:
    $6,615
    You may still be able to claim it as AI, even if it doesn't use NNs. It's still machine learning, however "simple" it is. As long as it's getting results, of course.

    Anyway, keep us posted on your app. A progress thread would be interesting to watch :)
     
    Siberia and MarcoSto like this.
  19. MarcoSto
    Offline

    MarcoSto Contributor Read Millionaire Fastlane I've Read UNSCRIPTED

    Messages:
    21
    Likes Received:
    31
    Joined:
    Apr 28, 2018
    Gender:
    Male
    Location:
    Germany
    Rep Bank:
    $368
    haha...no...iIts only AI if its complicated and hard to implement :D

    You are absolutely right, of course...

    thats a good idea. And it would give me some accountability so that I would feel a little bit worse when watching netflix than I currently do anyway haha...

    I'll browse through some progress posts and see how people set them up. Thx for the hint...
     
    ApparentHorizon likes this.

Share This Page