The All Turtles Podcast

Developing a solid pitch takes work, but it’s a necessary process because building any kind of product requires funding. To get that money, you have to talk to people. Mursal Hedayat, cofounder and CEO of Chatterbox, knew she had a good product idea when she noticed that hardly any of the people solving problems for the refugee community were refugees themselves. But to convince others that her product deserved investment, she has had to perfect her pitch.   Show notes Mursal Hedayat, cofounder and CEO Chatterbox (1:42) Chatterbox is an online learning platform where people can learn skills from refugees, starting with language skills Mursal was included in Forbes’ 30 Under 30 list of social entrepreneurs MIT Technology Review covered Mursal and Chatterbox in their Innovators Under 35 series   Conversation with Phil Libin, All Turtles CEO (11:53) Phil’s video on how to pitch investors We want to hear from you Please send us your comments, suggested topics, and listener questions for future All Turtles Podcast episodes. Voicemail: +1 (310) 571-8448 Email: hello@all-turtles.com Twitter: @allturtlesco with hashtag #askAT For more from All Turtles, follow us on Twitter, and subscribe to our newsletter on our website.

An organization’s leadership can make or break the success of the entire team. In entrepreneurship, a CEO has three main responsibilities: setting the vision, building the team, and ensuring that there is enough money to operate. But there is a lot more to managing success. Eugenia Kuyda is the CEO of Replika, an All Turtles product, and through her story of overcoming seemingly insurmountable obstacles, she shares invaluable wisdom for developing a long-lasting leadership philosophy.   Show notes Eugenia Kuyda, cofounder and CEO of Replika (1:42) Replika, an All Turtles product, is an AI chatbot for mental health The origins of Replika (article about Replika’s origin story) Phil Libin’s video on how to be a good CEO A graphic explaining the Japanese concept of ikigai, which loosely translates to “a reason for being”     We want to hear from you Please send us your comments, suggested topics, and listener questions for future All Turtles Podcast episodes. Voicemail: +1 (310) 571-8448 Email: hello@all-turtles.com Twitter: @allturtlesco with hashtag #askAT For more from All Turtles, follow us on Twitter, and subscribe to our newsletter on our website.

What’s the best way for a startup to compete with large companies when hiring top talent? How does someone reach outside their own network to source diverse candidates? To speak on some of these topics and more, Jeremy Vandehey, CEO of Disco, shares advice on how to hire an outstanding team. This is the first episode of Startup Playbook, a new series from the All Turtles podcast about how startup founders and product creators tackle some of entrepreneurship’s biggest challenges.   Show notes Jeremy Vandehey, cofounder and CEO of Disco (1:42) Disco is an All Turtles product that makes it easy to celebrate your company culture and values, all in the messaging tools you already use. Disco is currently hiring engineers. Check out justdisco.com if you’re interested in applying.   Conversation with Phil Libin, All Turtles CEO (8:42) Phil’s video on how to hire the best people We want to hear from you Please send us your comments, suggested topics, and listener questions for future All Turtles Podcast episodes. Voicemail: +1 (310) 571-8448 Email: hello@all-turtles.com Twitter: @allturtlesco with hashtag #askAT For more from All Turtles, follow us on Twitter, and subscribe to our newsletter on our website.

At All Turtles, we build AI products, which means we needed a process for deciding which products to build. We came up with a system for classifying the types of products people pitch us so that we can quickly understand and communicate to each other what we see as opportunities and potential failure points. This classification system is our product taxonomy, and it’s something any product creator or evaluator can apply to their work.   Show notes   Taxonomy discussion with Phil Libin and Jessica Collier (0:48)   1. Flying Shoe — Whoa, is it possible?! Failure modalities: a. DaVinci — Can be imagined, but it can’t be built b. Jetpack — Doesn’t deliver on its promise c. H.W. Bush — Already exists and we just didn’t know about it 2. Costner — Totally possible! But if you build it, will they come? Failure modalities: a. Waterworld — Too long to get to market and costs too much money b. Golf — Boring and we don’t care (e.g. ad tech use case) c. New Coke — There’s already a better version of this d. Timeshare — No plausible exit or sustainable financial path forward e. Segway — People don’t actually want it 3. Play-Doh— What is it? How will it work? What’s the experience like? Failure modalities: a. No particular failure modalities, but the product idea must become a Flying Shoe or a Costner to continue We want to hear from you Please send us your comments, suggested topics, and listener questions for future All Turtles Podcast episodes. Voicemail: +1 (310) 571-8448 Email: hello@all-turtles.com Twitter: @allturtlesco with hashtag #askAT For more from All Turtles, follow us on Twitter, and subscribe to our newsletter on our website.

Austen Allred is the cofounder and CEO of the Lambda School, a coding academy that provides free training for students to become software engineers and data scientists in exchange for a percentage of their income for two years. It’s a model that has moved people from low income to high income salaries. Austen talks through the school’s pedagogy, shares what students need to succeed in the program, and explains why one of the most valuable aspects of the school is the community.   Show notes Conversation with Austen Allred (0:48) The Lambda School provides free training for UX designers and data science in exchange for a percentage of their income for two years (0:56) Austen gathered the stories from a long list of people whose lives have changed because of Lambda School, which he compiled in this Twitter thread Articles from the New York Times and VentureBeat featuring the Lambda School Kwame Yamgnane from 42 Silicon Valley on episode 27 of the All Turtles Podcast (4:58) Sylvain Kalache from the Holberton School on episode 33 of the All Turtles Podcast (5:03)   Advice to entrepreneurs (21:17) The good/fast/cheap triangle of product development (21:20) Phil’s Rule of Three video on the topic of product development (21:25)   Listener question (26:20) From Derek via Twitter: This is a question about the the personalization episode of the Unscaled series: Greater personalization will always be better, but the marginal benefit will keep going down. Bottom line, will 20mgs of x work? Yes, but if 15.52mgs will work, and reduce side effects or chances of dependency, I’d take that. What do you think?   We want to hear from you Please send us your comments, suggested topics, and listener questions for future All Turtles Podcast episodes. Voicemail: +1 (310) 571-8448 Email: hello@all-turtles.com Twitter: @allturtlesco with hashtag #askAT For more from All Turtles, follow us on Twitter, and subscribe to our newsletter on our website.

When controversial headlines flood Twitter and Facebook, these rage machines disseminate disinformation. AI plays a role in perpetuating fake news, but it could also be a part of the solution in detecting and preventing the malignant spread of fake stories. Renée DiResta is a researcher of computational propaganda and disinformation. On this episode, she talks through some recent news stories and how they serve as examples of the ways in which technology facilitates tribalism. Show notes Conversation with Renée DiResta (0:50) Renée DiResta researches the spread of malign information on the internat (1:01) Data for Democracy is a worldwide community working together to find truth through data and technology (1:02) New Knowledge is an information integrity company (1:02) The Covington Catholic news story (1:10) How technology facilitates tribalism (5:54) Using something like AI to intercept disinformation before it spreads widely (9:15) Renée on her Senate report about Russia’s disinformation in the 2016 election (12:20) The sophisticated processes for gaming Amazon’s review systems (19:30)   Eye roll please (31:21) Waiting for the “right moment” to launch a product or company (31:25)   Listener question (34:18) From Jacob via email: Jacob says he is excited about autonomous cars because he hates driving, but in a previous episode, a discussion arose about the fact that not only is it possible that bad actors may interfere with the security of self-driving cars while they're on the road, but that it "will happen." So this listener asks: What can we do about this? Cryptography is not bulletproof, and also has system performance impacts. What is missing in the tech ecosystem to prevent tragedies related to the "hacking" of self-driving cars? We want to hear from you Please send us your comments, suggested topics, and listener questions for future All Turtles Podcast episodes. Voicemail: +1 (310) 571-8448 Email: hello@all-turtles.com Twitter: @allturtlesco with hashtag #askAT For more from All Turtles, follow us on Twitter, and subscribe to our newsletter on our website.

Bärí Williams is the VP of Legal, Policy, and Business Affairs at All Turtles. She has advice for small startups and founders who may not have in-house legal or policy experts on how to navigate things like privacy and contract organization. She also has experience developing diversity programs for companies like Facebook, and has lessons to share about creating and implementing similar programs. As she’s said, tech’s ethical problem is also a diversity problem.   Show notes Conversation with Bärí Williams (0:49) Bärí’s Twitter takedown of Ja Rule (2:46) Bärí’s article “Tech’s ethical problem is also a diversity problem.” (3:47) Facebook’s Supplier Diversity Program, which Bärí implemented (6:58) What is Facebook doing to support the local communities around its offices?  (8:58) Taking a holistic approach to diversity and inclusion (13:22) Making products accessible to all (17:39)   AI use case (29:37) Phil’s Amazon Echo now sets alarms for the correct part of the day without him having to specify every time (30:08)   Listener question (32:02) This question comes in from JK via email: “Any plans to establish a base in London, UK?”   We want to hear from you Please send us your comments, suggested topics, and listener questions for future All Turtles Podcast episodes. Voicemail: +1 (310) 571-8448 Email: hello@all-turtles.com Twitter: @allturtlesco with hashtag #askAT For more from All Turtles, follow us on Twitter, and subscribe to our newsletter on our website.

Max Levchin is the cofounder and former CTO of PayPal. His stories from PayPal’s early days reveal that its methods for addressing security questions from the start can apply to the problems social media companies face today. Levchin started Affirm to do away with amoral loan practices; recently, Affirm allowed government employees who couldn’t work during the shutdown to take their time in paying back loans. Levchin values honesty and transparency, which serve as the foundations for his products.   Show notes   Conversation with Max Levchin (0:47) Max is a cofounder and former CTO of PayPal (1:07) The human-computer loop (7:06) PRNG is a pseudorandom number generator (9:53) All Turtles Podcast episode with Renée DiResta (13:13) Affirm is a better way to borrow money (17:20)   Advice to entrepreneurs (34:10) Phil’s “Rule of Three” video series (34:38) The rock-paper-scissors model of dealing with competition (35:10)   Listener question (42:28) From Steve via email: I saw an article that Google is giving away $25 million to fund humane AI projects to solve seemingly intractable problems. My question to the All Turtles luminaries is what they think of this effort, which coincides with Google’s decision to not pursue $10 billion military contract for use of AI (Microsoft is now pursuing this).   We want to hear from you Please send us your comments, suggested topics, and listener questions for future All Turtles Podcast episodes. Voicemail: +1 (310) 571-8448 Email: hello@all-turtles.com Twitter: @allturtlesco with hashtag #askAT For more from All Turtles, follow us on Twitter, and subscribe to our newsletter on our website.

Fans of the movie Her will remember “Samantha” as the AI who connected with a human because of her ability to express deep empathy. That’s what Michelle Zhou, cofounder and CEO of Juji, references when explaining what her company aims to do: create empathetic and responsible artificial intelligence agents. By focusing on human-machine symbiosis, Zhou has worked toward creating AI chatbots for different types of use cases, from personalized healthcare to conducting interviews.   Show notes Conversation with Michelle Zhou (1:35) Juji builds empathetic and responsible AI chatbots (0:45) Humans feeling more comfortable with machines (3:48) Where did the name Juji come from? (4:55) Books by Cilfford Nass, a renowned Stanford researcher on human-computer interaction (8:02) Kim Scott speaking about ruinous empathy (15:29)   Eye roll, please (21:51) Target the biggest market possible (21:54) We should be able to explain how we design and train AI systems (29:41)   Listener question (26:48) From Ari via email: I’d like to hear All Turtles’ leadership team’s reaction to a quote in the Wired interview of Geoff Hinton: “[There’s a question of whether] regulators should insist that you can explain how your AI system works. I think that would be a complete disaster.”   We want to hear from you Please send us your comments, suggested topics, and listener questions for future All Turtles Podcast episodes. Voicemail: +1 (310) 571-8448 Email: hello@all-turtles.com Twitter: @allturtlesco with hashtag #askAT For more from All Turtles, follow us on Twitter, and subscribe to our newsletter on our website.

Although Wikipedia is a contributor website like Twitter or Facebook, why doesn't it suffer from the same levels of misinformation and fake news? The Wikimedia Foundation’s executive director Katherine Maher joins this episode to talk about the importance of transparency and sticking to your values at scale. She also shares how machine learning will be increasingly important in Wikipedia’s operations, and what organizations using Wikipedia’s datasets should know about the flaws in its information. Show notes Conversation with Katherine Maher (1:02) The Wikimedia Foundation is a non-profit that provides structure for free knowledge and hosts Wikipedia (1:13) Wikipedia is the world’s fifth most popular website (1:25) What’s important to the Wikimedia Foundation? (3:36) Wikipedia’s asymptotic mission (4:27) The importance of transparency in Wikipedia edits (11:42) The Unscaled Series on the All Turtles Podcast is about how economies of scale have changed and how startups can adapt (12:40) Hemant Taneja’s book, Unscaled: How AI and a New Generation of Upstarts Are Creating the Economy of the Future (12:42) Wikipedia as a moderating force in partisan politics (13:20) The role of machine learning in Wikipedia’s future operations (18:15) Articles about women are only 18% of the total biographical pages in Wikipedia (19:58)   AI use case (26:06) How do we make the internet more compatible with democracy? (27:16) Jon’s experiences with his phone’s call screening functions (30:00)   Listener question (34:05) This one comes in via email from Ari, and is another example of a user rising to the challenge of finding a useful application for blockchian. He sent us a link to an article about using blockchain to track food safety and origins. The article explains how blockchain can prevent fraud in the food industry so a consumer in China buying Australian beef can be sure that the package does, in fact, contain Australian beef. The writer of the article asserts that in food production, the manufacturers could record documentation proving that what they’re packaging is what they say they are, and upload that documentation to blockchain-based databases. Food could be labeled with stickers readable by smartphones that connect to the proper documentation. So, is this actually a useful application of blockchain?   We want to hear from you Please send us your comments, suggested topics, and listener questions for future All Turtles Podcast episodes. Voicemail: +1 (310) 571-8448 Email: hello@all-turtles.com Twitter: @allturtlesco with hashtag #askAT For more from All Turtles, follow us on Twitter, and subscribe to our newsletter on our website.

When it comes to curing Parkinson’s disease, the biggest obstacle is in making use of siloed research. So says Katharina Volz, founder and CEO of OccamzRazor, an AI-neuroscience startup. They’ve built a “super brain” to read everything ever published about Parkinson’s and related diseases, and to make connections to clinical research no human possibly could. Hear Volz’s plans for finding cures for patient subgroups—and why she recommends wearing rollerblades in the lab. Show notes Conversation with Katharina Volz (0:53) OccamzRazor is an AI-neuroscience startup focusing on treatments for Parkinson’s Disease (0:55) What is the connection between AI and Parkinson’s disease? (1:14) Why the “one solution fits all model” is a fallacy (2:31) The problem with the breakthrough mentality (6:12) PubMed is a search engine for scientific research (9:01) Why is Parkinson’s disease OccamzRazor’s first focus? (14:48)   Eyeroll, please (20:10) “Move fast and break things” (20:39) The Unscaled podcast series is about how economies of scale have changed (22:55)   Listener question (26:39) I was elated that my most recent question read on the AT podcast revealed Phil’s 3 design principles. I wonder, however, if there should be another principle added to the three: decision criteria auditability. Examples: an AI was trying to identify which dogs were wolves, and when researchers looked into the AI’s decision-making criteria, they realized the AI’s primary differentiating criterion was whether or not there was snow in the background of the image. So what do you think of auditability as a fourth design principle for AI?   We want to hear from you Please send us your comments, suggested topics, and listener questions for future All Turtles Podcast episodes. Voicemail: +1 (310) 571-8448 Email: hello@all-turtles.com Twitter: @allturtlesco with hashtag #askAT For more from All Turtles, follow us on Twitter, and subscribe to our newsletter on our website.

Virtual reality is a tremendously rich medium for creativity and new ideas, but we’re still waiting for the AR/VR revolution to make the waves it’s been projected to make. What can we expect from this field in the coming years? Adrian Kaehler is a researcher who’s worked in computer vision, machine learning, robotics, and with AR/VR software, and he has predictions. And, as the cofounder of the Silicon Valley Deep Learning Group, he knows what it takes to bring disparate technological groups together.   Show notes Conversation with Adrian Kaehler (0:53) The Silicon Valley Deep Learning Group (1:13) Magic Leap’s main product is a head-mounted virtual retinal display (2:44) The homebrew computer club was an early computer hobby group in Silicon Valley that met from the mid-70s to the mid-80s (3:17) What do the non-technical folks bring to the table in SVDLG? (4:21) What can we learn about VR in the wake of the failure of Google Glass? (12:20) How can VR technology improve social interactions rather than impede them? (14:32) Open CV Computer LIbrary (20:37) Open 3D, a modern library for 3D processing (20:47) CARLA, an open-source simulator for autonomous driving research (20:50) OSVF, the Open Source Vision Foundation, has a mission to build a common foundation for all of these technologies to advance together (21:21)   AI use case (26:10) When an AI-powered car nearly caused Jessica to miss her flight (26:15)   Listener question (30:18) Via Twitter (@allturtlesco): In response to one of your articles that mentions algorithmic bias, is there research happening now on the effect of AI's bias on people?   We want to hear from you Please send us your comments, suggested topics, and listener questions for future All Turtles Podcast episodes. Voicemail: +1 (310) 571-8448 Email: hello@all-turtles.com Twitter: @allturtlesco with hashtag #askAT For more from All Turtles, follow us on Twitter, and subscribe to our newsletter on our website.

Mursal Hedayat launched Chatterbox, an All Turtles product, to revolutionize language learning, and it’s a prime example of technology that enhances humans rather than replaces them. Chatterbox is a language and culture training program that employs refugees as tutors. Mursal was inspired to launch Chatterbox because when her mother came to the UK from Afghanistan, she was underemployed. Today, Mursal wants to help refugees make full use of their potential.     Show notes Conversation with Mursal Hedayat, founder of Chatterbox, an All Turtles Podcast (1:05) Chatterbox is a platform where people can learn languages with refugees (1:05) What is a life worth living? (4:45) Why shelter and food aren’t the only resources refugees need to build a life (5:02) The importance of the relationship-building aspect of Chatterbox (6:51) Why Chatterbox provides not just language training, but cultural training as well (7:48) What is the pedagogical thesis behind chatterbox? (8:17) Blended pedagogy, combining independent study with one-on-one tutoring (9:07) The missing piece in other online language-learning programs (10:01) The algorithm in development to match learners with tutors (12:54)   Advice to entrepreneurs (19:39) The importance of having a strong point of view (21:01) All Turtles’ “Dealing with competition” video (21:46) Omatenashe, the Japanese approach to hospitality involving deeply personal, deeply human interaction (22:17) Rakuten, a Japanese ecommerce company using technology to make drivers more valuable rather than trying to replace them (23:30)   Listener question (26:39) From Ari via email: I recently read two articles (this one and this one) related to AI visual systems that clearly show that even a 5-year-old can perform better than an AI visual system. [For example, an algorithm can identify a picture of a vulture, but if the photo is rotated on its axis a few degrees, the algorithm thinks it’s a picture of an orangutan.] I thought these might be the basis for a discussion that debunks the near-term fear of AI overlords, pushed by people like Musk and Hawking.   We want to hear from you Please send us your comments, suggested topics, and listener questions for future All Turtles Podcast episodes. Voicemail: +1 (310) 571-8448 Email: hello@all-turtles.com Twitter: @allturtlesco with hashtag #askAT For more from All Turtles, follow us on Twitter, and subscribe to our newsletter on our website.

Angel investor Jason Calacanis joins us for the second part of his two-part interview, continuing the conversation about what his early investments in companies like Uber and Robinhood taught about success. Now, he invests in bold ideas, and urges founders to focus on building products people love. Then, he, Jon, and Phil lament over everything that’s wrong with the elevator pitch, epitomized by Jason’s story of the time someone tried to pitch him while at a urinal.   Show notes Conversation with Jason Calacanis, angel investor (0:43) Your brand is your reputation (1:04) Why intentionality matters (2:42) Who is Michael Ovitz? (4:36) Everything that’s wrong with an elevator pitch (6:04) Why Launch is based in Australia (9:47) The bold ideas Jason invests in (17:50) Why are they called angel investors? (18:55) Jason’s first business, Hot Tapes (19:50) Jason’s mayoral plans (23:24)   We want to hear from you Please send us your comments, suggested topics, and listener questions for future All Turtles Podcast episodes. Voicemail: +1 (310) 571-8448 Email: hello@all-turtles.com Twitter: @allturtlesco with hashtag #askAT For more from All Turtles, follow us on Twitter, and subscribe to our newsletter on our website.

Jason Calacanis, angel investor, joins us this week for part one of a two-part interview. As an early investor in companies like Uber and Robinhood, Jason shares his insights for achieving startup success. He explains why a company’s brand is so important from its beginning stages, and delivers a hot take on VC culture today. He also gives his opinion on what the number one killer of startups is, and mulls over the following equation: 1 good idea + 1 good idea = zero good ideas.   Show notes Conversation with Jason Calacanis, angel investor (0:56) Thumbtack, one of Jason’s investments, is an online service that matches customers with local professionals  (1:12) DataStax, another of Jason’s investments, is a cloud database company (1:13) Robinhood, another company Jason invested in early on, a financial tech company (1:14) Jason’s book, Angel: How to Invest in Technology Startups--Timeless Advice from an Angel Investor Who Turned $100,000 into $100,000,000 (1:18) Launch, founded by Jason, has a mission of supporting founders and inspiring innovation (2:20) Silicon Valley Reporter was a publication that focused on New York’s tech scene (2:28) Why is brand so important from a company’s early stages? (2:35) The This Week in Startups podcast (5:05) Chris Sacca’s first episode on This Week in Startups (6:43) Phil Libin on an episode of This Week in Tech (7:29) The Midas List of 2017 (10:57) What is, in Jason’s view, the #1 killer of startups (13:22) One good idea plus another good idea equals zero good ideas (14:28)   AI use case (14:35) How Phil’s watch violated principles of good AI product design (16:32)   Listener question (18:39) Sift, an All Turtles product,f is an experiment in news therapy (18:53) Listener question via the All Turtles Facebook page: “Does Sift remove biased statements and opinions and just report the facts?”   We want to hear from you Please send us your comments, suggested topics, and listener questions for future All Turtles Podcast episodes. Voicemail: +1 (310) 571-8448 Email: hello@all-turtles.com Twitter: @allturtlesco with hashtag #askAT For more from All Turtles, follow us on Twitter, and subscribe to our newsletter on our website.