AGC SXSW Digest // Thursday, March 15th

Ethereum Will Disrupt Everything // Andrew Keys, Co-Founder ConsenSys Enterprise

What is Ethereum? It's a transactional layer of internet 

AI Creativity in Art, Neuroscience, & the Law // Jessica Fjeld (Harvard Cyberlaw Clinic), Sarah Newman (Harvard), Alexander Reben (CEO of BlabDroid, MIT grad), Sarah Swettmann (MIT)

I wish this session could have been 3 hours instead of 1! The overarching theme was discussing how artificial intelligence can intersect art, how we respond to that in a human way (the tensions, fears, etc.), and how we distinguish and define who gets credit for the creations. The session was organized like this;

  • Each person introduced themselves and their work
  • then discussed how AI is currently intersecting with art, the law, neuroscience, and where it is likely to go in the future
  • spoiler alert : it's not going anywhere, it's only going forward, so we should think about and learn from

Jessica Fjeld was the moderator for the panel and introduced some questions and context as well as the model for how AI generated artwork functions. As of today, there are programs that have the capability to produce art that is indistinguishable from art made by humans without the aid of technology.

Some contextual references presented to frame this discussion:
One argument is that generative art has been produced for centuries. An example is the Mozart Dice Game which was published in 1792 and would use the roll of two dice to randomly select small sections of music to compose an entire piece. While the claim that Mozart was first to do this seems to be unprovable, the name stuck. You can download an app on your phone called Mozart Dice Game (it costs $1.99) and see the algorithm play out before your years. This raises a few questions:

Who is the author of the music created with the app on your phone?

  • It couldn't be your phone, that's just the tool
  • is the author the programmer who developer the app?
  • Is it the user of the phone (ie, you)?
  • Is it Mozart (or another composer) from the 1700s who first defined the process of using dice to create music?

The point being, the computer or machine operates at the behest of many humans. 

This taxonomy chart is from Jessica Fjeld and Mason Kortz A Legal Anatomy of AI-generated Art: Part I and was presented in the session:

AI taxonomy chart by Fjeld and Kortz

AI taxonomy chart by Fjeld and Kortz

Sarah Newman is currently an artist and researcher at MetaLAB at Harvard and has a background in philosophy and fine art. She explores difficult philosophical concepts through her work and installations. One of her installations was at SXSW and I went to it after the session. It's called The Future of Secrets, here's how it works:

  • you type your secret on a computer in the room (hidden characters, so your secret remains anonymous)
  • once you enter your secret, a tiny paper prints out of a machine someone else's secret
  • there is a projection of the secrets on the wall
  • there are headphones in the room that are playing the secrets in automated voices, both typically masculine and feminine

A little bit more about the installation:

  • It was inspired by the relationship of humans and machines
  • she noticed an uncanny phenomenon; people connect with someone else's secrets and apply meaning and a narrative to them. Some people even become convinced that the someone-elses secret that prints out is meant for them. Like the machine is a psychic or like the installation somehow connects with someones mobile device upon entry and then learns things about them and then produces some meaningful secret that wasn't their own, but that they can relate to. The people believe the machine knows more than it does. 

Some things I asked Sarah after the panel ended:

  • Did she have any expectations for how people with interact and react to the installation. Her answer: not really, but she did expect that people would be more hesitant or uncomfortable about sharing as secret, however, in hind sight, she realized that the people entering the room were volunteering, so that made sense. She didn't expect people to connect a meaningful narrative to their secret 



AGC SXSW Digest // Monday, March 12th

DIRTY LITTLE LIARS: WHY ARE THERE PRODUCT FAILURES? PRO TIP ! the recording is available if you click on the title link // Mahin Samadani, Charlie Burgoyne, Charles Marcus, Tracy Thirion #dirtyliars

Charles Marcus – Co-Founder, Vidlet (moderator)
Charlie Burgoyne – Founder/CEO, Valkyrie Intelligence (was an astro physicist first, and then left and went to work at Frog)
Mahin Samadani - Partner, McKinsey and Company (started on Newton team at Apple - one of biggest product failures of our lifetimes)
Tracy Thirion - CEO, Bamboo Worldwide

Session Description (directly from SXSW website):
Google Glass and Heinz Purple Squirt Ketchup. How can these products be such market failures? We are swimming in data, yet it's a fact that over 50% of products fail every year. Why is this happening despite having so much data easily accessible to us? Of course, much of the answer is that humans are messy and unpredictable. How do we cope with unpredictability and find meaningful insight? In this session, we will explore the rise of quant analysis and the fall of qual analysis. Are we better off?

Real life blatant example of when quantitative methods/research/analysis : 2016 election. Michael Moore was gathering qualitative data in rural parts of the US and the writing seemed to be on the wall, but we latched onto what we feel like is safer data (quantitative).

Techniques and machine learning thoughts:

  • In one of Charles Burgoyne's jobs, the company dropped traditional terms like gender, race, etc. and instead created new architypes
  • building tools that allow us to build things that will push you places you can’t even fathom, we'll be designing for 45 architypes
  • The next 10 years in AI : we will be creating products that conform to people vs. expecting people to conform to products because products need to be designed using iterative steps and AI can be programmed to do this without a huge manual time investment
  • bringing a human element to data
  • Duolingo example : instead of being obsessed w/ average person, they built a platform that adapts to each user. Changing pedagogies and taxonomies and changing entire experience based on the individuals
  • The products can and will fail, but they will heal when they do
  • Good designers are those who both understand customers and data and can empathize with both customers and data and data scientists (or be both in one)
  • The designer and the scientist become a more fluid, evolved idea. 
  • Tracy said she doesn’t think adults can learn empathy (I almost entirely disagree with this, but I could also be wrong)
  • we're loosing empathy in our society largely because of our digital tools we're building
  • example of woman cleaning house: quantitative analysis would provide things like; how many hours per week she cleans, what types of products she uses, her patterns. What would be missed is something important that was only revealed during conversation: she has deep rooted emotion tied to cleanliness and her mother, whenever she spoke about her mother, she would grab her neck.

Qualitative vs. Quantitative : qualitative is the spark, quantitative is the fuel. The qualitative research is the why, it’s the influencing and emotional impact. There’s a reluctance for clients to want to pay for the qualitative research because it’s more nuanced and uncertain, however it builds confidence through depth and understanding. It’s a harder sell to the boss. Or the bosses bosses boss. Problem in market industry: there are not a lot of good qualitative moderators available.

What’s the process of bringing qualitative research into quantitative research? It’s the blending of insight with analogies. Doing so can allow the processes to get directionally dangerous, allow space to fail, and working toward a solution. This will require a massive cultural shift to combine these methods and be ok with failing, to see it as a necessary part of the process and not as a negative.

Training people on how to sell ideas: create an emotional connection to the idea through compelling stories.

Why do products fail?

  • Because they lack empathy or       
  • There’s a lack of diving more deeply into the data

Contextual research: co-creating products alongside users. Story: Maheld’s firm has a project of working with a utility company to transition from analog maps/highlighters/binders to a digital system. The designers rode along in the trucks with the drivers to gain insight in how they use their tools to perform their job. The boss was certain that the designers wouldn’t be able to tell him anything he didn’t already know, but he was able to.

The most basic machine learning is the survey.
Qualitative, quantitative, qualitative, quantitative


Tension of a business decision vs. streamlining a process

The ethical layer – responsibility of the designer to help make sure the machines don’t do EVERYTHING and mess it up.

Why retro packaging works: it’s a safe bet, unlikely to offend anyone.


DESIGNERS WHO BROKE THE INTERNET // John S. Couch (Head of Design, Hulu), Tina Koyama (Product Design Manager, Twitter), Jill Nussbaum (Product Design Manager, Instagram)

Tina Koyama: product manager @ Twitter. She’s swiss/Japanese (she pointed this out)
Jill Nussbaum: design manager at Instagram in NY. Worked at R|GA, worked on Nike and has been a teacher for 9 years. She teaches a class called Design in Public Space.
John S Couch: currently at Hulu, previously ebay, Magento, Wired Japan, then startup. SF – LA

Their brands are in the spotlight currently. 

Particular project with strong audience reaction:

Jill talked about the Instagram rebranding project: biggest logo and UI change to date. started with research specifically focusing on the core element of the brand that people connect with (rainbow and camera lens). There will always be lovers, haters, and mime creators. The UI changes improved user experience a ton.

IG did a good job of PREPARING THE COMPANY as a whole for the rebrand. The involved everyone in the company all through the redesign process so everyone was excited and felt a proud ownership when the new UI launched. The truth for how successful the product is is always somewhere in the middle between the SUPER negative and SUPER positive feedback (outliers). Don't get stuck in overhype or in trolls.

Internally socializing.


Tina talked about Twitter's intentional redesign and had a significant shift from a social media app to a news app. This change happened about a year ago. On the tweet, there's a retweet, like, comment and significant change is real time for like and retweet counts (animation). Realtime user feedback. When things break on the internet, things break on TW also, so they have immediate feedback. 

  • No one on the team is ever surprised by feedback on TW because they design with their users along the way and focused on the most important part of the brand

Why did IG change the design?

Instagram hadn't had significant change since launch in 2010, community had grown and products had grown, so they decided they needed a change as well, it was time for the brand to also grow. It's was scary!

Why did TW change the design?

For TW, it was more about focusing on the most important piece of their brand and focusing efforts to communicate. Tina would describe it as a focus rather than a product change.

Why did Hulu change?

Ostensibly to bring live into the product. Also, they hadn't changed fro 10 years. John saw an opportunity to bring OTT (over the top) and also an opportunity to entirely change the user experience and marketing/brand focus. They made a video of people using the mobile product. They got the new product launched in 14 months. Launched with new show Handmaids Tale. He says it’s important for the WHOLE COMPANY to be an advocate for the end user and that this core piece sometimes gets lost.

Narrative, continuous storyline is important for designers to keep sight of this so it doesn't get lost in all the KPIs, data, etc. 

Focus of Tina's class she teaches is the narrative. Her students create what she calls design fiction. It's important to tell the story of the product in the near future to get buy in and excitement for everyone. Can also show a few different paths they can take – create stories and films of people using the product.

3 ways they learn at Instagram:

  • Foundational research  (in the field, in homes, businesses, understand the challenges) or lab research
  • Building quick prototypes and dog fooding it in their internal test app and then run experiment in real app (maybe 1% of users)
  • Run experiment in real app (maybe 1% of users)

^^ go into understand phase quickly so they can understand if they're moving in the correct directly.

Twitter has a similar process to Instagram. Design is always partnered with research. – challenging to design products from end to end in a big company. Shifted entire company to a jobs-to-be-done framework, very user-centric framework. This gets everyone thinking about the user experience as a whole vs. individual products. They ask themselves:

  • When do people hire twitter? And, just as importantly or maybe more importantly
  • When do people fire twitter?

Is there anyone in the company who owns overall user experience end to end?

  • One example of a brand who does this well is Apple: they have continuity throughout their entire user experience from someone passively observing a billboard to someone browsing their website
  • At Twitter, the CEO Jack Dorsey is the visionary who aligns everyone and evangelizes all of the principles really well throughout company
  • At Instagram, they have a slogan that says “nothing is someone else’s problem” so the expectation is that everyone is responsible. They have a series of internal reviews and a weekly design review. Design leadership from across company, not just your product team. It's the responsibility of each design lead to know what is going on across the company. So, even though there are 40 designers at IG, the end user experience feels like there's one designer.

Companies are very design-centric, do you ever have to explain what DESIGN is?

  • There are a lot of different designers at Instagram who have different skillsets. Instagram is a design lead company and one of the most important questions they area always asking is who should we bring in at which point in the project.
  • HULU is a mix of Hollywood + technology + design which makes evident the omnipresent tension between Hollywood and silicone valley. Hollywood is established (100 years) and more static while Silicone Valley is agile. With the different players converging at HULU, the company had to define what design means for them. Hulu defined roles of UX vs. product design vs. information architecture vs. motion designers vs. communication design

Did you anticipate the reactions to the redesign when launched?

Instagram : Knew there would be a strong reaction, prepped the company as a whole and tried to be as transparent as possible (w/things like videos). When feedback came out, they listened,  but gave more weight to the analytics or anecdotal and how customers were actually using the product

Twitter: anticipated reactions and they work so closely with users for the entire process of the redesign. They wanted to bring the brand to life and The big things they did were changed typography for headlines to make them more readable and be able to easily differentiate the content being shared vs. personal

  • Changed reply icon to be more universally recognizable
  • differentiate content vs. person behind it (shapes)
  • Iconography changes; changed reply from back arrow to comment bubble (universally recognized)
  • All changes informed by user feedback, so user responses were quite positive
  • Twitter prepped users/were transparent by launching blog posts about their process
  • They recently launched their design blog, about 6 months ago
  • RADICAL TRANSPARENCY translates well

Hulu got ahead of it with videos and how-to use the new products. "Disruption that doesn't serve the user is just indulgent." - John. So make sure the improvements help the customer and communicate the why and the process.

  • Hulu used videos as teasers ‘this is coming’
  • TV is a more passive medium compared to TW and IG, the mobile experience has an opportunity to bring more interaction than the living room hulu experience

Most Valuable Product (MVP): for each redesign, did you launch what you wanted?

Twitter : design systems project let them launch in a couple months (WHAT). Because they built a strong foundation so it was easier to launch from there.

Instagram : they have a design systems team (SO IMPORTANT) so they can solve problems without having to solve the entire UI for the whole product. They looked at how users in Indonesia were hacking IG as a platform for ecommerce. They were using hashtags to group like products. They hacked search by putting the word ‘pants’, for example,  in as the name so their pants for sale would come up easily when someone searched pants.

Hulu: it was a year-long launch in which they interated and grew as they went.

They all spoke about the importance of getting offsite and starting initial design phase on paper.

Closing remarks, one piece of advice, mantra: 

Jill, Instagram: one of IG's design principles; do the simple thing first. Design all of the things and all of the possible solutions and then figure out (through data, research, rigor) and then, from there, iterate and simplify.

Tina, Twitter: so important for us to understand the people who use the product, especially international and those not around us. Empathy. Designing for 10 people is very easy, designing a universal experience that works across countries is. Soft skills: very important to work in a large organization; how to negotiate, persuade, present well. 

John, Hulu: courage. A lot of designers are introverts, so it can be hard to stand up for design. Being exposed while creating; the data as well as the craft. And then stand with a design through critiques. The greatest designs are when everyone initially thinks you're completely out of your mind, but then they start to tip and see that you're absolutely right. Mantra: be courageous about what you're doing and being ok if it fails.



opening fact: 55% of our understanding of conversation comes from visual cues. 

Kate has a background and interest in both behavioral psychology and UX strategy. She is a former theater nerd and has been in 40 musicals, so she has strong feelings about motion in general. She explains that our brains have evolved to derive great meaning from motion and so it makes a lot of sense that we include animation in our UX strategy.

She advocates for creating animation with a purpose and with personality. In doing so, we can change behaviors and influence KPI. She uses the Fogg Behavior Model to illustrate some points:


Some guidelines for good animations

  • Clear connection : a clear visual that shifts cognitive focus. An example is an animation that activates when clicking 'add to cart' and a little dot floats from the product up to the cart icon to make the connection. 
  • talked about our working memory, as humans, of somewhere between 2-7 items (the jury is still out) and so, based on that concept, visual elements should be combined. Reduce clutter to reduce mental clutter. 

Concept: Perception of Speed

  • Optimistic UI:  This concept is where the product assumes success and keeps the user moving to the next step rather than forcing the user to wait while the system catches up. 
  • Not optimistic UI: picture an icon commonly used for waiting or loading he page (three dots, circle figure, etc)
  • Transition animation: this allows a pause in the experience to be a feature rather than a bug (YAY)

Concept: Create a Narrative

  • Stories are 22 times more memorable than facts alone
  • create your narratives around the user's journey, what makes sense?

Concept: Celebrate Progress

  • Kate used the example of the Apple watch celebrating your exercise rings
  • This taps into our dopamine release (PAVLOV), so be responsible and don't abuse it. It's highly addictive and so validating (that's why we're addicted to social media LIKES). It's powerful enough to consistently change behavior.

Concept: Education Yourself on the Outliers

  • Kate used the example of consideration of what might cause an epileptic episode.
  • Design for the outliers and you'll design a better experience

Concept: Create animations with personality and embody the BRAND TRAITS
"Brand is a decision making tool." - Reed, Brand Strategist

  • Identify 3-5 adjectives that sum up the brand and use these as good guard rails. Mailchimp is a good example. Apple: simple, declarative, clean.
  • Motion design should follow the brand traits. It's amplified when an animation is off brand because of the MOVEMENT
  • Nike is an example of a brand who does motion design well (from their commercials, to apps, to website)
  • Trust based brands (such as credit card companies) use movement that is graceful and evokes security, ballet-like movements
  • Brand trait of HUMAN examples: think of the rhythm of heart beats, breathing. Example: Being John Malkovich, why were the puppets so human like? Because of the fluidity of their movements.
  • Brand trait of RESPECTFUL examples: tea ceremony (respecting the process). This should be contextual (ie: the first time user/experience can be slower, take your time vs a command as in DO IT NOW).
  • Brand trait of DOWN TO EARTH examples: grounded, point of origin, using a sense of gravity or physics. 

FInal notes - she suggests Keynote and After Effects as good animation software to start with.

What User Experience Exposes About Your Culture // Leah Hacker (Accomplice), Olivia Hayes (Accomplice)

FIRST RULE: Always know your audience.
"Internal disconnect always finds it's way into your end user experience ... always." - ACCOMPLICE
^ a mirror
How can you tell? 

They also make a podcast called It'sWorthDoingRight.

Contents : case for culture, issues/symptoms, conclusion


  • constraints should be based on empirical evidence
  • the point of this topic is not to debate financial success, we know that's a dumpster fire
  • Accomplice is arguing for overall sustainability and longevity as measured by employee and customer satisfaction and retainment
  • true biology - finances, karma, business
  • the system; what you put in, you get out


  • the culture of your company impacts outcomes
  • company culture is significantly correlated to customer experience
  • culture influences innovation

What do we know about BAD culture?
The women of Accomplice took to the streets of SXSW and asked the people, here are some summarized responses:

  • there is backstabbing
  • people are selfish
  • individuals feel isolated
  • you can just feel it in your gut

a reasonable conclusion (as proposed by Accomplice): this definitely affects the product
Do most companies realize how important their company culture is? NO (GIVE US PING PONG TABLES PLZ - agc)


  • The company has a single visionary (internal, example is Steve Jobs)
    There can be a failure to propagate the vision
    Single innovator/decision maker
    From a business perspective, it becomes very difficult to scale and innovate

    this creates an inconsistent product
  • User Experience Symptoms after the exit of the visionary - the product degrades over time (DONGLE LYFE)

case study APPLE // clean, simple, intuitive

  • the whole dongle dilemma is a symbol of innovation, but at the cost of degradation of the user experience
  • becomes a parody of its former self


  • design becomes chaotic
  • the internal symptoms become business implications. This is costly in-market missteps, infrastructure failures put the company somewhere between vulnerable and liable
  • the user experience systems experience feature bloat, lack of scaleable infrastructure. This leads to critical product failure.

case study YIK YAK // anonymous

  • a proximity based social media user feedback app largely marketed on college campuses
  • didn't anticipate all the hate speech and harassment that ensued
  • good example of how money doesn't always fix problems, sometimes it kills the product faster
  • many products need TIME to mature
  • Yik Yak couldn't fix the issues and it shut down in 2017


Internal Symptoms

  • fear-based decision making
  • "copy cat" syndrom
  • abrupt shifts in mission, values, focus

Business Symptoms

User Experience Symptoms

  • incomplete products are fear released and they lack value
  • abrupt shifts in offering

case study YAHOO // 
Yahoo wanted to transition from a search engine to a media source, so they acquired a bunch of companies. As they were evolving, consumers were confused about what they were, so they lost the audience and user base.


Internal Symptoms

  • lack of transparency
  • gender/homogeneous leadership
  • group think

Business Symptoms

  • high turn over
  • consumer distrust
  • lack of strong innovation

User Experience Symptoms

  • alienation of potential user types
  • tone-deaf feature sets
  • poor product

case study APPLE HEALTH KIT //
Mostly men worked on the product, it didn't have any capability around the menstrual cycle which effectively alienated 50% of the users. This mess was egregious to offensive.

case study GOOGLE //
In their 2015 image label feature, the product identified groups of black folks as gorillas. This is an example of how a shortcoming in a diverse group of engineers and creators leaves holes. This product was encoded with the same unconscious biases as it's creators.

Referenced Kate Heddleston for her Diversity in Tech expertise. They also referenced her blog post about The Null Process.

Referenced the podcast How I Built This with Guy Raz. The episode is an interview with Yvon Chouinard, the CEO of Patagonia.
Referencing the Growth At Any Cost issue - Patagonia was in danger of falling into this when they decided to slow down and take action as though they would be around for 100 years (long term). It totally worked, Patagonia is a wildly successful company. They are ethical and created sustainability by taking a pause. 


Changing Minds; Behavioral Science for Designers // Steph Habif (Tandem Diabetes Care), David Ngo (Behavioral Delta)

FAV TAKEAWAY FROM THIS SESSION: "You have to fight your tendency to fall in love with your invention, you need to fall in love with your problem"

ABOUT David Ngo

  • created Stanfords first Behavior Design major in 2011
  • works closely with BJ Frogg (Frogg Behavioral Model)
  • finishing a book on behavioral design with Frogg
  • works with non-profits in global malnutrition and education
  • works with for-profits in millennial news engagement and mobile health
  • works with startups to tackle huge challenges in behavioral addiction, technology addiction, and consistent mental health via meditation
  • Behavior Delta is a design firm that leverages neuroscience and behavior design to support purposeful profit companies

David used the Frogg Behavior Model (same as Kate did in her Animation session same day) to illustrate the part of the human experience where designers have the most influence to prompt behavior.

What behavior are we trying to prompt?

  • scale back and mimic psychology of product
  • simplicity changes behavior over time (referenced, a BJ Frogg concept)
  • identify the scarcest resource for that person in that time
  • most likely to prompt action if it matches your routine
  • there is a depravation of human connection in many digital products
  • this is related to the tech addiction
  • we need to design for human and social connection


ABOUT Steph Habif

  • behavioral scientist with 20 years of experience leading healthcare teams on ways to design for consumer engagement
  • was a freelance consultant for 10 years
  • spend much time in academia, currently at the Design Lab at UC San Diego
  • recently joined Tandem Diabetes Care as Senior Director of Behavioral Sciences
  •  at her job, she leads human factors and market research to ensure insights effectively translate to design and communications

She has done a lot of work with the Insulin pump. It's challenging because;

  • the FDA presents an interesting regulated design constraint
  • the product cannot be changed without another submittal once the design has been approved by the FDA

Building Trust

  • building trust in machines takes time
  • it took people 50 years to trust an elevator without a person in it even though the early elevators were capable of effectively operating the same as our current-day elevators
  • how can we design for trust on an expedited timeline, sooner rather than later?
  • there is a concept called Algorithm Aversion: how people feel about algorithm-enabled machines (we DON'T LIKE THEM, we don't trust them)
  • systems must be trusted to do what we expect them to do
  • the developers of technologies care about people
  • humans form trust over time 

control -----------------------------------------------------------------------> trust


Maslow's Hierarchy of Needs 

Maslow's Hierarchy of Needs