What I Learned in 2019 (and you might find interesting)
Every year, you start the calendar with blank, grey months to fill in with colors.
You may never notice this colorful change daily, but you’ve probably done a lot in 2019. You’ve colored your calendar.
I’ve put together my lessons learned that might be useful for others as well. Kind of like a 15 seconds of fame for each month. Watch the clip below (if you haven’t). It gives you a brief summary, then read whatever you’re interested in below.
Details are below:
Read only what you’re interested in based on the list. It’s been a long year. It’s a long post. Oh, and the answer to the question from the clip is 7.
- Using a game engine for quick prototyping
- Recharge. Reflect. Celebrate.
- Future of work. Data. Value. Measurement.
- ATD ICE: Smart Dust LXP, LDX, AI, ML
- Visualization of data: storytelling
- eLearning is here to stay
- Open mind, not empty head.
- Data science: data literacy
- What are the odds?
- Intention vs. Implementation
- Who to follow?
Then for months, it gradually grew into a more comprehensive list of examples, including the challenges for the ATD TK session.
What I learned is that there are roughly 3 types of elearning designers out there:
- “I have zero coding experience but I’m good at copy pasting.”
- “I’ve been trying to get this interactive done and I’ve found the missing piece. Now, I can tweak it to my needs.”
Here’s a couple of things you can learn how to do:
- Transparent lightbox
- Toggle menu on and off anywhere automatically for users
- Copy text to clipboard
- Showing numbers like: 1,234 with automatic comma insertion
- Random numbers
- Monitoring text entry real time: as the user’s typing they answer you can count characters, words, enable and disable buttons based on character length, etc. WITHOUT waiting for a submit button
” I’ve used it to do some software emulation for my company and it’s hugely appreciated. As a non-programmer, there’s *no* way I would have been able to hack together a work-around. “Nick
2) Using a game engine for quick prototyping
February 2019. First week is actually ATD TK, where I did two sessions. One was a hands-on introduction to machine learning through kaggle.com. If you’re interested in data, what ML can do for you, and you want a place where you can not only learn but run code on the same page, try kaggle.com.
At ATD TK there was a lot of buzz around xAPI as well. We even did a live cohort report out for Megan Torrance. Interested in xAPI? Join the next cohort. You don’t need to be a geek. We need all kinds of people. You might even meet the Gone. The xAPI gone that is.
And this is one of the reason you should go to some conferences: networking, meeting people from online.
I put this picture (Diane Elkins and me) to an AI test later to see if it recognizes what’s in it:
You’ll be the judge how AI did…
In Febuary, I also did a webinar on Using a game engine for eLearning.
If you haven’t used game engines and you’re curious what they could do for you, I suggest playing with Construct. There’s a free version out there to test (limited number of actions and features but still good for a test drive).
Why Construct? It is a 2D engine, so it has less of a steep learning curve than the real things like Unity or Unreal. It has built-in functionality that authoring tools do not: physics, collision control, creating elements on the fly, etc. And finally, it has a visual and a logical interface. You don’t code, just build logic how elements should behave.
We looked at the following use embedded as a Webobject in Storyline:
- Mini-game inside Storyline that can receive information from Storyline (reading variables) and send information to Storyline (setting variables)
- Quick prototyping tool (you can build an interaction in minutes to playtest, and then if the client approves it, you could build it out in Storyline in weeks)
- Hidden driver: you can embed the game engine in a webobject outside the visible area and use it as a hidden driver. It can run “simulations” behind the scene and report it back to Storyline. For example, you can build one “customer” with certain traits. Then you can duplicate this character into hundreds of them, with each having similar traits. Now you have a crowd to work with in your simulation. They can “talk” to each other, so if in your customer scenario you don’t do the right thing, it would affect many of your customers.
Another interesting hidden driver use case: path finding. Construct can calculate the fastest route between two points based on obstacles. Could be a maze, house, roads, process with obstacles in the way..
3) Recharge. Reflect. Celebrate.
In brief, don’t forget to unplug. Recharge. Reflect. Celebrate.
4) Future of work. Data. Value. Measurement.
April 2019. With my Kineo colleagues in Chicago, we attended Trish Uhl’s future of learning workshop. It was a day with fellow practitioners and panel speakers to look at the state of our industry from a practical perspective. What’s happening? What’s working? What are we doing? What should we be doing?
What we learned from industry leaders from real L&D implementations (as opposed to white papers, blogs, influences, content growth engineers, etc. who may or may not have ever tried to implement their theories) is that we’re not waiting for this big change anymore. We’re in the change. It’s messy. Technology changes so fast that you should keep in mind a couple things:
- Don’t wait for the perfect platform. LMS and other platforms may last two years, and then new features completely change them or you buy a new one.
- Don’t rely on ONE platform. Use what works today and look for integration.
- Collaboration is key! But features do not collaborate. People do. If you spend all your energy and resources on technology and features, you may be disappointed in your “social learning” platform.
- Start small. Start specific. Pilot a program where you have buy-in already and you can control the execution. Learn from the pilot and expand from there.
- Learning is top of my mind today for leaders. But “learning” does not mean taking courses. It is the end result of the mean: able to do more, reskill and upskill. You should think learning and performance ecosystem, not a course in the LMS.
And finally: learn basic data literacy. Stats may sound like high-school all over again but you need to be able to speak the language of information.
5) Smart dust: ATD ICE and LXP, LXD, AI, ML
May 2019. ATD ICE is one of the biggest event with participants from all over the world. This year, I didn’t have a session. I only attended the expo, roaming the 400+ vendors all day.
Articulate was swamped all day. Seems like Rise 360 is taking off along with Storyline. But if I have to name one thing. One thing that was palpable in the air; hot as an air balloon, widespread as wind… That would be SMART DUST.
Smart dust is actually an emerging tech. You can see it at the bottom of Gartner’s Hype Cycle.
But what I’m talking about in the L&D space as smart dust is the magic potion brought to you by the letter “X.” Learning Experience Platforms (LXP) was the hottest thing. So hot off the press that one of them was handing out champagne to celebrate the entrance to the market right there.
There’s nothing wrong with LXP. Literally. And that’s where the smart dust comes in. This is the same feeling we had for the LMS 15-20 years ago. We started associate everything stupid, bad, and old-fashioned with the old tech, and think of nothing wrong but the sheer beauties of the new, exciting, shiny thing.
Again, the problem is not another platform. The problem is we assume this smart dust will solve all of our problems. Technology is never the solution. Right now, it feels like we invented Alexa, and it/she/he should be in our car but we don’t really know what problem it/she/he could solve. But it’s cool. It’s smart.
If you’re in the market for an LXP or you’re an LXD (learning experience designer), just make sure you don’t fall for the smart dust: AI-driven. Everything that can execute a SQL search is now AI-driven. (Not to mention that machine learning itself is not AI.) Here’s an example of this smart technology. This AI-driven algorithm was supposed be able to finish your sentence. Not only finish your thought but write a paper on it:
And WOW, it does. It reads like a real person. So, technically this AI-driven algorithm does what it was programmed to. The problem is that it learned from us. It learned from all the nonsense we (as in humans) put on the internet. The problem is not with the technology. It works. The problem is with DATA. AI needs data. Lots of it. If the data is biased, inaccurate, or wrong, the result will be the same. Vendors sell. Sell a product. Keep this in mind all the time when roaming the platform.
Speaking of vendors… I love data. I love when your solution not only works in theory but you also have some tangible details of an implementation. But when you show that, please, keep your story straight!
On the positive side, I finally was able to take a selfie with the amazing Cara North. She’s been doing more for our learning community in the last couple of years than you can even think of. If you’re not connected with her, I’d strongly recommend.
6) Visualization of data: storytelling
I learned the best way to visualize the process of converting between different coordinates. Specifically, to cube coordinates.
If you’re into games and gamification like me, this site is an amazing treasure trove: www.redbloblgames.com
What I also learned recently that that trick “?rel=0” after a YouTube link no longer works as expected 🙁 — It used to hide “related” videos at the end. You know, those random, HR-embarrassing clips…
7) eLearning is here to stay
July 2019. I visited Hungary, the country I’m originally from (land of Gulyás, Budapest [pronounced as “sh” not “s”], Ernő Rubik aka Rubik’s cube, Puskás Ferenc, etc.). As my daughter was taking a summer course at the alma mater I graduated from, I discovered something inside this beautiful old school…
In the “basement” we I studied for so many English exams, on the wall, I found this poster:
In this poster: in the field a young couple is taking an eLearning course, while two cows are checking out the event. One of the cows say: “Ez marha jó!”
The eLearning is for KRESZ (abbreviation for some crazy long institution for traffic and road safety). Basically, it tells you that you can prepare for your learner’s permit test anywhere. eLearning is here to stay.
Being Hungarian, there must be some clever word play: hence the cows. “marha jó” says the cow. “marha” is officially beef (hence the cow). “Jó” is good. But in this slang expression marha becomes an adjective for something fun and exciting. (Something like “FUNtastic!”)
Like my light breakfast in Debrecen, Hungary:
Anyway, eLearning is here to stay. My hope is that it’s becoming what it always should have been (check out the serious e-learning manifesto.)
8) Open mind, not empty head.
September 2019. I participated in a data visualization workshop with my one of idols in the field: Edward Tufte. (Check out his books on data and evidence.)
Anyway, one of his favorite sayings is: “You approach novel research findings with an open mind, not with an empty head.”
We discussed “drawer bias” in research publication. A “drawer bias” basically means that publication with positive results ARE MUCH MORE likely to be published than negative ones (hence they stay in the drawer). Why is that? Because many research projects are sponsored. When a sponsor finds out that your findings are not positive, they might not want to support it anymore.
Another warning Tufte expressed is using machine learning or AI applications without know what they’re doing. Today, it is a line of code in R or Python, and the selected algorithm spits out the answer in no time. According to Tufte, these algorithms should be labelled as medicine. Only used based on prescription and for the specified symptoms. Otherwise, you’ll might get side effects.
NO DATA coming from any algorithm should be used to make important decisions unless you know how the algorithm works.
I started my Harvard Edx online Data Science certificate while I was in Hungary. It contains seven courses from statistics to machine learning, all hands-on in R.
9) Data science. Data literacy.
Not just me…
The prevalence of data and analytics capabilities, including artificial intelligence, requires creators and consumers to ‘speak data’ as a common language,” says Valerie Logan, Senior Director Analyst, Gartner. “Data and analytics leaders must champion workforce data literacy as an enabler of digital business and treat information as a second language.”
My two cents is that this is not an all or nothing game. My advice to people who ask me if they should just ignore any of these because “they don’t work” is that in order to ignore something because it doesn’t work, first you need to know the fundamentals how it works.
Learn the basics of data. You don’t need to be a data scientist but you should be able to ask good questions. Not to be sold, not to be misunderstood. Like a foreign language in a foreign country.
For example, your LMS admin says: “The average score in your course is 85%” Now what? Sounds high? Any action items? Passing score is 80%, so are we good?
What if you also learn that the standard deviation was 5? Or 10? Or 20? Which one you should worry about? Does it make any difference? How about sample size? Mean? Medium? What if Joe says he’s got 82% and wants to know the percent of people who were better than him? Or, your business stakeholder asks: “What are the chances that someone can pass just by simply guessing?” There is an answer. And we should know that.
10) What are the odds?
October 2019. Speaking of chances… What are the odds that you see R everwhere when you’re working on your R code to build a linear regression…
Just one warning about correlation vs causation. Just because two things seem to correlate in time, it does not mean one is the cause of the other (see learning -> performance).
Not convinced? Check out this site (and book) or spurious correlations: https://www.tylervigen.com/spurious-correlations
25 years ago my thesis was to build an artificial neural network that learned adding numbers together without any human intervention (besides providing the data). It took me five months to write the code and three months to train my neural network. Today, the same thing would take two lines and 10 seconds. Think about this pace!
Finally, I’ve also finished working on ATD’s TD at Work edition coming out in January. It is about game thinking (games and gamification) and ten years of my knowledge in ten plus practical pages. Look out for the publication!
11) Intention vs. Implementation
I used this example from my home country during the game engine webinar for Training Magazine. What do you think the function of this road is?
There is a big difference between intention and implementation. When we see a proposal for a problem, these two are combined. We often reject the implementation (what we see as the proposed solution), while the other party sees our reaction as an attack on the intention (the true intent to solve the problem).
In this case, just by looking at the curvy bicyle road, we might think this Nyíracsád place is going for the Guinness records of the shorting bike road in the world (implementation).
However, digging deeper, the intention is revealed. On the other side of the road there is a bike road and bikers often turn left. When they turn left onto the auto road, accidents may happen. Or they can hold up traffic. (Let’s give the benefit of the doubt that this road gets really busy.)
To avoid that, they decided to continue the bike road on the side, where bikers can make a little half loop, and then stop. Once traffic is clear, they can go.
This is an example of GOOD intention, BAD implementation. This is not how reality works, folks! Next time when you present a solution, set clear expectation: what is your intention? What is your implementation?
Next time you criticize a solution, ask: What are we solving for (intention)? How does your solution solve this problem (implementation)?
If you disagree with the intention itself (this is not a problem), then your discussion should focus on the whys and not the solution. Otherwise, you’ll look like two politician trying to solve two different problems arguing that that others are wrong in their implementation strategy.
12) Who to follow?
I had a couple of mentoring conversations this year with new, talented professionals getting into the learning field. What I learned that we ended up with the same question at the end: Who to follow?
My two cents I learned can be helpful:
Balance your network between three types of people:
- ENGAGE (NOW): Peers at your level. They’re figuring out the same things you’re doing now. You can learn and share daily tips. This is a down-in-the-trenches support thing.
If you ONLY have engage people, you may get stuck spinning your wheels. It may feel like every project is a struggle. You may burn out never popping up from the trenches.
- MOTIVATE (SOON): Find people who are months ahead of you. These people have fought their battles, found their ways. Use them to find where to put your energy and resources.
If you only have motivate people, you may see solutions but for every day issues you’re left alone.
- INSPIRE (FUTURE): These are people years ahead of you. Check out the trend, explore their vision. Use them as a compass to walk towards the right direction.
If you only have future people, it might seem you’ll never catch up. That you’re way behind where you should be.
Where to find these people? LinkedIn, Twitter, Facebook, conferences, local ATD chapters, Articulate eLearning Heroes community, bloggers on elearningindustry.com, movers and shakers list, etc.
Here’s a list to start with: https://trainlikeachampion.blog/training-influencers-to-follow/
Devlin Peck has put together a comprehensive list of how to become an instructional designer: https://www.devlinpeck.com/posts/how-to-become-instructional-designer