How can we reliably come up with more and better ideas? Creativity doesn’t have to be just left up to chance. With the right techniques, the process can be systematized. Google Brain’s Jeremy Nixon joins the regular panelists for a discussion and workshop on ways we can become more creative. Recorded 11 February 2018.

Image credit: Jeremy Nixon

Brought to you by Fling: Urban Drone Delivery. Get it fast. Fling it!

Shownotes

Guest: Jeremy Nixon
https://angel.co/jeremy-nixon
https://www.crunchbase.com/person/jeremy-nixon
https://www.linkedin.com/in/jeremyvnixon/

Jeremy
https://docs.google.com/document/d/1Ip7nkMnVpdgpG4X5i4O1x1cSGvFItZl0rtSgGXfx8G0/edit

https://docs.google.com/document/d/1RiRp08FXKhLWp03Gwwk3turlLKIaQ_x6VpdRhffyMPE/edit

quantification of our failure modes

James Altacher podcast

generate 10 ideas for a problem he was working on

time constraint (5 minutes, 10 minutes), you get an idea
let quality fall (expected value / volatility)
novelty is valuable in research, but in other things just execution is valuable
engage subtly different mental processes

how ideas propagate across networks
–> Richard Dawkins (meme) (comes from mime = copy)

e.g. euphemism treadmill
e.g. watch memes buttress one another

“idealist” over ways to turn memetics into a scientific field

use NLP to show track ideas over time

Michael (not discussed):
topology of meme communication: ants, humans (how large the individual brains are), how strong the bandwidth is between the individuals

Jeremy: back to creativity… load up a set of ideas, take a shower, your mind will move freely over those ideas. doing that intentionally on a daily basis.

Thomas Edison: fall asleep with a rock in his hand, free associate for inventions, wake up in a half dream-like state.

Michael: In the Middle Ages, they didn’t bathe/shower, perhaps that’s why they didn’t make any progress.
—————————-

Recombination
dissimilar domains
have a way to represent those concepts in that domain
machines are good at discovery of structure inside of data
can you abstract over it and generalize to find new data points
diffuse mode is hard to implement
all ideas in a shared mental space
hard to represent ideas in a machine so that it will see the similarities between them.

scrum/blitzkrieg == what’s the larger word for that process

— Michael (not discussed):
Any ideas I have seem to come spontaneously at random moments, since my subconscious was working on the idea.
—

WORKSHOP WITH JEREMY NIXON.
(8 minutes)

CONCLUSION.

Episode Machine Transcript (unedited and uncorrected)

Welcome to let’s future discussion about future trends technologies and their implications for society.

From. The world that brought by dot.

Delivery.

Really get.

This episode’s future trend discussion topic creativity with Jeremy Nixon.

Welcome to let’s make the future. I’m Daniel ends with I go alone me I MICHAEL CAREY My name is host saying quote I mean my name is Sarah fail and I’m Tanya and today we have Jeremy as a guest here can you introduce yourself currently working on machine intelligence research at Google or Google Graham but I have a time in the past year writing about extraction and information processing how to do concept representation affectively and a lot of I.D.S. war that have been generated by this model for automating creative processes I am really interested in basically finding more general principles behind generative modeling both a machine learning and human and behind creative concepts generation and really excited about optimizing the process both for humans and for the shins so this should be fun so you know Jeremy tell us how did you get into all this creativity business so I was a Harvard working on a lot of problems around behavioral economics and trying to sort of create mathematical models of human your ethics and biases and through quantification of our failure modes start new experiments that would allow us to overcome some of those failure modes and basically in trying to do problem solving in that environment I found if pod cast by somebody named James alter I don’t know if any of you have heard of him but basically he had this habit where every single day he would get a little notebook that he kept in his back pocket set on a timer and generate ten ideas in some contexts for a problem that he was working on whether it was ideas for creating a. Or solving a particular problem needing money needing to find the person needing to learn some skill whatever it was he would sort of take out this notebook write down ideas in that some demand and start to execute on them and it actually started to change his patterns of thought so that his intuitive reaction to a situation was no out here’s a problem one frame that I have is I’m coming up with ideas that will solve the problem and so he would basically fit inside of this basically set of control and understanding for him so control saying you know how do I basically modify the inputs to whatever the situation is such that I get the output that I want and immediately doing idea generation of that and turning that into a habit so that was his intuitive reaction to the situation rather than being something that required effort or that So there’s this like body of techniques that can exist outside of your mind where you actually effectively do it and it’s much slower and less intuitive and it’s harder to integrate all of the knowledge that you have with that mental motion if you haven’t sort of embedded it into your body of mental habits and so I was pretty impressed with it I think just because in planning and in problem solving it’s really valuable to say kick some goal that you have and generate ten or twenty different ways of accomplishing that thing and then systematically execute over the paths to it and a lot of the value that comes out of the generation of ideas is that there are sometimes pathways to a goal that can you know shrink the amount of time it takes by ten X. or by a hundred X. and sitting down and doing one of these for the schools and super useful for me so when I was basically growing in college and starting to get into machine learning I would basically use this technique to improve my body the problems of a clam and keep generate ideas in spaces that I care about that sounds pretty cool what was your experience with it like to chew follow the specific framework that you learned from how to true and if yes did that work for you in that way yes I think that there’s this beautiful thing to time constraints so I actually think of it as the sort of modular two or so I delisting eyes or my favorite techniques actually turned it into a daily habit and that was really transformative for my creative output the time. Constraint component of it is huge because as soon as you know you only have five minutes or ten minutes to accomplish something there is a pressure that’s put on you and so your idea of actually starts to come more quickly and this is your stick where like if you’re running out of time and haven’t gotten to the market you should let quality fall in quality usually looks like either expected value or volatility so by modeling like how new or how different or how unusual is the idea that you’re generating now it is really valuable say in research that may matter a lot less if you just care that executing that can tax you just want the highest expected value but as you sort of free yourself to generate ideas that have lower value because you just want numbers you actually start to engage like subtly different mental processes where you let yourself freely associate between ideas that seemed strongly disconnected in the past and so the time constraint axis is really nice and forcing action I think it’s really useful for journaling as well Jeremy I think we call that desperation don’t we when you when you have a few minutes left you desperately try to link a few concepts together maybe it works yeah that’s right well it is activated and process which is great I think we’ve all had the experience of coming up against some deadline and realizing that we actually have enormous amounts of self-control and the ability to focus that didn’t exist before and that and wishing that we could be inside of that time constraint state more often and I think that you can you can just create a system where you time things and you disallow yourself from doing things outside of time and so put yourself in touch intentionally I love the desperation state it’s very strong very strong there may. I am curious about this Jeremy because I hear you talking about it and I’m like This is an amazing coping mechanism for being under pressure and that’s kind of the idea or a positive way to respond to it and so it’s a semantic as well you talk about systematizing creativity what I want to say about those folks people can get in where you just kind of hit a wall and it can maybe be an emotional thing and I was wondering if that was what you were calling a failure mode can you go into more details about the failure modes Yeah the upside to letting. Your quality by a slit is that it frees you from the funk if you are open to everything then you actually start to generate bad ideas but as you start to do the generation you both realise that it’s possible and you slip into a mode where you’re basically rowing with ideas and you actually get into this creative state of mind and that actually is really really helpful for funds often the functions are there because you have this somewhat artificial extremely high barrier to entry for any idea that you come up with and because you’re putting that kind of pressure on yourself it’s hard unless you have very like reversed pressure of time pressure to force yourself to put things that are slightly suboptimal so that you can get yourself into a state where you’re generating things that are active the useful I would love to hear is to put it in a context like an anecdote you might have a personal account or one from someone you know where you know some of the quote unquote bad ideas they have that actually broke this wild out for them you have a trip in their home that it should be so one answer is that all of systematizing creativity was generated as an idealist and so the origins of this project was somewhat I also use it in research fairly often but one of my favorites is in understanding how ideas propagate across networks so I was reading a book by Richard Dawkins called The Selfish Gene and I ran into this concept of a meme trance or that comes from the intersection of Gene biological change which replicates and mime so to copy and so memes are visibly the description of the way the ideas copy themselves as they move from person to person and I was basically doing creative ideation over how means operate so for example I would come up with an idea like memetic dressed like mean start to get associated with the emotions they come around with and so if there’s a move that has some negative connotations and the past you know people would refer to someone who’s now African-American or black it’s colored and because there is like the negative association racially in the United States that term started to have this like racial connotation and so the people who want to Markham south as not being race will move to.

Instead of saying colored and now all of sudden there was this distinction between me and I guess after everybody who was both like raises the not sort of using the word Negro they moved to African-American after an African and I’ve been with black Amazingly there is this drift in Myspace due to the negative attention associated with the idea or like also not it’s even watch memes but trust one another so if two people have competing wishes of it where there are two sides of a political battle then they’ll generate controversy that we’ve seen huge argument but basically I’ve been reading The Selfish Gene the idea of medics came up and I decided to Idealist over ways to turn the medics into a fine to fix field and ended up with a number of ideas one of which was to create computational Maddux through machine learning and natural language processing so if you watch the way that ideas move across every special network you can start to isolate the properties that ideas have and all the properties that people and communities have and so there is basically a creative process that I ran over an idea in a book that I was reading that led to an interesting research front here that has central to put foundations under parts of social science that’s one from example I guess that if you look inside the sort of Korean box a nation I’d be a generation back when I shared in the group you can see seven examples We use of idealists to generate ideas whether it be unlike machine learning research or finding contrarian positions on important topics or for learning in whatever domain each era when it there are a ton of us text for learning more quickly and efficiently but also for problem solving generally so I have a bunch of example you want to see and here I have an observation one thing that I notice in science fiction there’s a trope that you visit a planet and it will be a bunch of other people that kind of look like humans or aliens that look similar basically every conception of the other of an alien or of some different race is still usually consists of a bunch of individuals and that’s what made for example the concept of the Borg in Star Trek kind of a powerful idea because it was a hive mind so I feel like this idea of tracking ideas. How they evolve how they get passed on between one another you can call the memes or call them ideas or call them concepts but it seems to me like when you study that you’re sort of helping to break down the difference between how mental processes work within the confines of one brain and then how they sort of start to mix and swirl around in the minds of many many creatures and so I wonder if we’re often limiting ourselves when we conceive of systems and societies in the future as being collections of individuals when really the boundaries between individuals are not as straightforward as that and we should really be modeling the future and how the future will look in a much more fluid sense where there is no real distinction between individuals and the full civilisation because ideas are just going to get easier and easier to transmit between brain to brain Yeah that’s right and that transmission mechanism says a lot about what idea of propagate were ideas don’t there’s a strong simplicity falter for example and so overly complex ideas don’t get communicated or at least on technical effectively there’s also a strong emotional need where you have to activate a desire to spread an idea in some other individual if you want it to prompt and there’s an emotional component to it and then died of a single man you can imagine communication not being about but something else and so I think with a tool that allows you to communicate that meaning as opposed to like the particular way that you describe the idea should change things so if you have no interface that lets you communicate immediately entered a plea pretty dramatic there was a process of abstraction and generalization and sort of ask myself what is this process and why is that helpful and what is it doing and how can we it’s a CLI find more tools that you need exact same thing so there is a body of importantly different approaches to idea generation and I think one of the most important is intentional diffuse not you can think of this as the state of free association when you know you’re on a walk. Taking a shower when you’re about to fall asleep and you start to dream certainly when you’re meditating definitely when you’re on psychedelics I guess there’s a mode of thinking that allows you to freely associate ideas in different domains with one another without being conscious of direct but what you can do is load up an idea or a set of ideas and then go take a shower immediately after and your mind will move freely over those ideas and after all generate really unique and fascinating insights and so doing that intentionally on a daily basis is actually amazing for my writing process and shares their anecdotes about the shower insights that really translate them for the character but I think that that mode of being a super useful if you’re falling asleep there’s this famous story of Thomas Edison habitually falling asleep with a rock in his hand inside of his chair so that he basically would have softly associate over it ideas for inventions and then once the rocket was in his hand at the ground he would wake up and he would wake up in this half dreamlike state where many ideas have been floating around in his mind mixed together in unique ways and would run off and run whatever experiments just come up with a certain entering diffuse what intentionally is extremely useful and came out of basically a simple search for other categories of idea that also would let you generate novel plots you have been reading and experimenting with this I would assume I mean I can agree to that shower I don’t like with Hiram or university where you had like assignments and then suddenly after having worked on a certain math problem for a week without any idea and then in the shower you suddenly solve it but I would imagine that if you try to systemize in that way and basically do something then go ahead and take a shower afterwards with the objective to get a good idea or get creative in some way I would imagine that at least in my head there would be some build up of pressure that would work contrary to what it should be like the free floating mind that you said do you have any experience with that yet Tucker Daniels wary Yeah once I fall into the hot sun just feels amazing and I forget what I was supposed to be doing something it’s really hard to be so. Rounded by warmth and basically was the pressure when I’m walking outside and that’s gorgeous I just notice that it’s gorgeous the pressure falls away so I don’t really have practical issues with it I think that if you’re trying it for the first time that might happen but I have a point I just feel like yeah going for a walk is raw so I don’t practice level hasn’t been an issue try it out and see if you feel that way if you do maybe it’s an issue but if you tried out and it’s not then you are free to just likely associate and generate ideas now and implementing a question and I’m wondering also why they’re on their raise yourself and those ideas you mission Sherry’s they want to buy their Titan system is that like process itself Carney’s question I guess that how do you operationalize that it’s just like noticing that you will basically have some working memory sitting around and if you heard ideas that you want to do with in the working memory the things that your own mind wanders over first will be the things that we’re thinking about there’s so much a way that you know you’re working on a math problem before you fall asleep and it ends up in your mind it’s really just in time associating it and then going and doing the diffuse mode activity whatever it might be OK I will definitely try that out I mean with my work on map after nine pm I am very very confident at least in my experience from university and afterwards that I will not be able to sleep for hours usually because Prime Time runs too quickly and maybe it’s preggers Interlochen.

And so I thought I meant mapping is actually better than sleeping for the most part and it’s often because it’s easier to wake up and either RAM or write before you going to deep sleep and have ideas it’s actually much better than trying to use like your late night sweep for creative ideation I find that some fields are better for creativity than others like with mathematics and programming I find that I get a bunch of nonsense ideas if I’m just sleeping on it like I find it’s time for very useful it’s better to walk through something very systematically looking at the mathematical problem the paper rather than rely on my unconscious but maybe I’m right yeah well I think that our math is also a harder case where you. There’s a system and it’s rigorous and it follows very concrete will swear if you’re say writing or you’re an author the idea that you generate say in fiction can be far fetched but emotional and that is not something that like if right or wrong country if you are brainstorming and you just need to filter and also can be really helpful is just idea generation but certainly when people are daydreaming you know when you’re on a walk in a shower it’s much closer to regress mode but there are still controls in place in the mind to make sure that everything makes sense and so it might make more sense to use those kinds of techniques if you’re doing something extremely rigorous and running they can have also any metrics that I can figure out like what are your creativity minds are at the moment and also lecturing at their spectrum reflected the creativity and highest So I think that certainly like abstracting over your problem or of your solution that is like finding what kind of structural problem that has and finding it in many other problems or finding other applications or solution and generalizing it to its farthest reaches of super useful basically recombination concepts of super useful Will machines be able to become creative How will it benefit them so I have a research front here in creative forms of machine intelligence and a lot of it is actually downstream of better representation of the world so there are a number of process it’s the same go through summer into it have some are not a computer some are it’s clearly a technique that you could run intentionally and basically a P.S.A. an automated clip to be is to be recombination so if you want to find new ideas you take sets of problems or solutions that are to exist and you recombine those solutions with one another in flexible ways to work on some other problems and the humans will usually say that some solution is creative if it’s a solution that is on the want or that other humans had not thought of before or between domains that are not very similar to one another at first glance so that we combination really just requires that you be able to represent the concepts inside of that machine and have those concepts and interact with whatever problem it is that’s being worked on and a lot of ways which ones are. Better at doing basically discovery of structure inside of data and so many problems will look like you know here a number of data points they all share particular types of feature a particular type of structure and can you abstract over it and generalize to new data points basically figure out which data points is it similar to using some how will notion of similarity and I guess the ability to basically train on a number of tasks and generalize to some other domain where you’ve broken up your understanding of the task in a sub tasks is something that humans do all the time that is totally implementable and certainly many people who are brain in a few months are working on doing this right now so I think that a lot of our creative processes fall into these pretty clean categories and machine intelligence where it’s like oh you know what we’re doing when we generalize can actually just be creative and what we’re doing when we basically we combine concepts can be implemented some things are much harder to implement so I think that if you smoke free association it’s very hard to implement humans are very good at basically integrating all of the experiences that you have in the systems that exist in a shared mental space where they can sort of do cognitive fit between arbitrary domains and it’s very powerful hard to replicate So yes it’s hard to represent problems to a machine and such a way that it can see the overlap between all of them and there’s a sense that you can make a metaphor between almost anything and the human will pick up lots of subtle similarities and structures inside of that metaphor and frankly the metaphor is usually doing tasks that abstraction where there’s some shared structure between two domains in the metaphor selecting them and we don’t know at a point at exactly what that trait structure is we just let it set inside of the metaphor So for example yesterday talking to a Stanford researcher about his creative process he pointed out during a small experiment and then iterating over and over again and that shows up in minimal viable products such as in the scientific method it shows up in the military close up and let’s create There are all sorts of metaphors you can make that kind of process and what you want is a label for the process and ideally you would basically generalize to other problem that was well fit but that solution and you start to figure out what kinds of problems are the solution and when you do it over dramatically different brains will say oh a computer is being creative in generalizing. That kind of solution is not to me so let me share a short block that basically has a number of approaches to automating creative processes in shimmering there are things that aren’t in here so I guess you can sort of see in the intersection between machine research ideas and sort of how the creativity processes that I think are actually really hard to implement everything the short term but a lot of the short attempt at an answer to the question of machine creativity I guess a lot of it’s borderline implementable are already implemented cool thanks I want to ask this question of implementing creativity for machines when you bring an example Jeremy that it was successfully implemented and a task that the machine solved using some creative strategies you have that frame is actually in my mind mostly about reference points where whether or not you call something creative is actually a function of your expectations and so what usually happens is you implement some creative process inside of the machine and quickly people are like oh no that’s actually just such and such an algorithm that’s not creative and so there’s this constant moving from Tir much like with the concept of artificial intelligence or the concept of creativity I think certainly almost all of machine learning looks fairly creative if you start to find the Long Tail say recommendations that work very well so once you say are Amazon and have huge amounts of data on what people like and are comparing people to other people you’ll start to make extremely creative looking recommendations things that people didn’t expect that they would like because it so happens that across the globe there are thousands of other people just like them who happened to also like this other obscure thing so I would point to almost all of those processes it’s a moving window so I can tell you things and they may sound I don’t press a lot of problem frames look like you have creative solutions because the optimal solution is actually very unusual So if you’re doing a linear programming problem really optimization problem there are going to be edges and optimization space that are your optimal solutions especially if it has the structure that can be fit into say a complex because you end up optimizing a loss function you’ll end up in like machine learning with all sorts of terrible behavior at the edges. Your space where the I will do something optimizer the last one sure but it’s actually very destructive and that’s actually very creative behavior because for the human it’s an unusual event if you don’t expect it to be framed that way you don’t expect a problem with it because you just set up this last option and had to reinforce it when you get it up to my eyes that you’ll say I have a racing but that actually just sits inside of the middle of the course collecting points because it found a way to collect points more quickly if through that mechanism then through actually doing the racing game so those are a few examples I can give you a lot more about the short answer Jeremy really wants to do this workshop and I’m excited to see what that’s like is a workshop like is it use our individual problem domains and we’re going to privately kind of work out things ourselves Yeah that’s right what you suggested time frame Jeremy if we can do it in twenty minutes Yeah didn’t you hear before it’s better to rush things right time constraints.

Or they force you to weed out those Everybody have access to the document that outlined all of the techniques and models for instance in a particular Yes from the Facebook Yeah that’s right it’s called sometimes creativity models of technique it opens with different categories of techniques and inside each category there are short lists of different versions of doing that technique and it should be actionable and should be specific enough that you can say I could go ahead and do you know abstract and transfer over some more problems and solutions or that you can basically do recombination the Republic seeing examples might also be helpful but I think that if you go to the stock you’ll be able to see explanations of accountants nicely OK Well let’s assume we’re all in this document so take it away Jimi so let’s do this I’m going to set a quick time in the next two minutes each of you run through the models and techniques and choose one that you think will be interesting they apply to different categories of problem so there are some problems that social solutions work well for like say crowdsourcing your creative process there are some problems that you really just want to do an idealist over summaries of the others if you are not sure ideals and works for almost everything transferring and attracting other summer solutions to problems works once everything is mode works for almost everything so you. Feel free to pick the top three if not yet choose one and basically read through the detail on how to do it and it may be short so feel free to ping the group with a slightly more because if you think you’re not sure exactly what you want me to question the prefix for as well if you want to start generating ideas that were so I’ll give you a forty seconds to think about it hopefully everybody has a tool or technique by now so the next thing is to generate either a problem or question or a solution that you want to generalize the principle of finding a target for the technique so only give us pretty seconds to decide exactly what you should use and so open up a notepad or a drive document and write down basically a technique how it’s likely to work as well as the target for the idea forty seconds more so go ahead and do the right that’s time setting up a document and write exams if everybody is set up I’m going to give us eight minutes for actually running the technique on the problem the solution the thought the idea that you have so we have one question that we’re supposed to answer and then we’re supposed to apply this technique to that question that’s what I’m going to give us it minutes and then after we can share the results of actually have ideation feel free to also do it over things you don’t share you don’t have to share it can be private those are often much much more interesting I’m going to certain three to one it may not come across six minutes last four minutes only criminal three men.

Time creative process over well and eight minutes on a technique we only have about five and a half minutes left so let’s do two shares if anybody is particularly excited and didn’t do anything dangerous and then we can have our last questions for me anybody excited actually I’m excited because I thought I was using technique number eight with the leading questions and the problem I had was a lack of standardize ation of American sign language because there are a lot of debates over legitimacy of certain interpret and product in the deaf community and so I tried to. I thought the sticker beliefs which were you know deaf people happen title meant to knowledge because of a cultural status one sometimes the cases their educational background doesn’t afford them the linguistic knowledge necessary to make the assessment same goes for hearing people obviously and then I thought of working backwards if it were standardized there might be a sort of police or governing body involved in that I thought and Steve says it’s like that big French language police that exists in France and they have something like that would be established I’m like the university Gallaudet University on a deaf university campus that would be a really interesting idea yeah that’s right and if you want to create organization you could do an idealist for basically a pathway it’s like it’s like if this or that really the question you know if they existed what would have happened and how can I actually make that exist so I don’t also Mr I can also share I was thinking about creating it until I found Diaz and Nick to read their conflict system raising their touch in creating ontology because sometimes I think it’s extremely hard problem to organize your ideas and I kind of like lists of ideas one for example was like abstraction you want to go into detail like when you have. To trade training or being in and there was a very simple what run you are next and King about the idea educating. And catch agrees eftpos and how you can essentially like a structure like different ideas there’s ground now also thank you Jen essentially have many different types of input data and example in the conversation it’s trying to find a many different categories are pandering and you should be used to from techniques and then you know I like reigning exhaling tax right there you are just like trying to win a game many different activities that that we do and interesting and you have main different types of imprint into an imprint in some sort of like an amp and then you can just say you know you can regenerate many different types of like and based on your input data a lot more as a kind of like a new shell and this. Sort of I mean it wasn’t a question. Thanks Terry I’m curious which of the models people chose to use in the workshop actually so I went ahead and did the composition so if you take an idea from domain and try to break it down in multiple ways and in this case I had just an idea somebody composition over IP A listing you know intuitively in the conversation that was my immediate mental motions like Oh you did the workshop as well Jeremy Oh that’s right yeah that’s what I was I thought I was also being creative that’s a good friend I noticed that I just started breaking down I do nothing in conversation because I started to install the mental patterns like I will it’s because of the subcomponents and maybe modularize it so that you can generalize and recombine the subcomponents into like new types of techniques but also understand like what’s a logical impact we have and you talked about psychological back of time and so that was my choice I did leading questions just like Sarah Yeah did you pick different ones which ones did you pick I just walked through the whole list of leading questions I think I got all the way to the point of what if question so I guess I got four down to twelve so yeah anyone else have any other questions Michael learning or you know there are a number of techniques you know you could how.

You can live with to crack the fiber speculation into what they need to walk for fact if you Isha so for example X. zero one four social technique for if it was legal out there I have to go I think to back up a go it’s for work to class work for more pay for their shells you know what they’re going to work for when you’re working with mashup example so there are six impact mix and all of them work and follow a different contract so the answer will be very very long if I try to give you my body of intuitions I don’t have a shareable Here are the things that all of the techniques are particularly useful for are not particularly useful for documents for you and so I think that that that’s the worth creating it’s actually on a pathway stock so there’s a like pathways for surprising creativity document which is basically in the fighting what direction we should go with these ideas and up of them are fairly fresh Well absolutely feel free to like pain if you have other where the emotions there are to you like twenty things worth doing with us one of them is deciding basically like how to map techniques to tie. I feel your pain maybe on that note I found it quite funny in the workshop why I chose was find a source idea categorize it and generalize it to finding more instances of a category so basic that’s now like reflectively I did a very very machine like behavior by asking a question coming up with like a starting point and trying to optimize from there by categorizing it into like different variables and try to optimize each one of them so it sounds really really rich Mick and the grass was how can we get tech people excited for an exclusive event here because we kind of have that issue right now it may call a big space so I came up with a good solution I will present afterwards I want to thank everyone if you have more questions than me on Facebook Feel free to ping me as they try to be here thanks Daniel for a bargain.

Thanks very much Jerry for your time and for your valuable input that was super interesting Yeah thanks everyone I was really fun.

Thank you thanks everybody.

Got a good morning good night.

See you guys next time and I thank you. Let’s make the future featuring the voices of Michael. Danielle’s balance whale.

Sarah.

Michael.

Eric.

Music to entertain Christian Pelton.

Let’s make.

Keywords

creative process (9) failure mode (4) machine learning (6) automating creative processes (2) generate ideas (5) time constraint (5) machine intelligence (3) problem solving (3) data points (3) idea generation (3) input data (2) falling asleep (4) working memory (2) intuitive reaction (2) systematizing creativity (2) free association (2) future trend (2) mental motion (2) mental processes (2) creative ideation (2) fall asleep (4) feel free (5) math problem (3) bad ideas (2) leading questions (3) freely associate (2) optimal solution (2)