The Local Maximum

So the overall theme of the show today is applying what we learned last time. Bayesian Probability is good. P-Value hacking - is where we get into trouble. Death to P-hacking! We start by following up on my monologue from last week about Beliefs and Probability. This leads us into an article about the Reverend Thomas Bayes himself, and how his ideas transformed our thinking on probability and belief. And finally, are most of the scientific studies that you hear about wrong? Worry about p-hacking, and think like a Bayesian, and maybe we won't be misled as often; let’s begin to attempt to figure out how to make sense of it all.

I explore the nature of probability, and my approach to thinking about facts and belief. I talk about problem solving a lot; here I go into the fundamentals and work to justify how I think about it. What problem does probability solve? Should we always think in terms of probability? What's the difference between an artificial language (code, math) and natural language? How many water molecules are in the Atlantic Ocean? Sometimes we have vague questions and make vague statements - should that be ok? This episode is a little different, but I hope this is the beginning of a back-and-forth conversation with the Local Maximum listenership to try to get to the bottom of these deep issues.

Aaron and Max have a stream-of-consciousness discussion where we look at some recent discussions on the tech web: Fred Ehrsam on Blockchain AI Marketplaces, and Chris Messina on the effect of AI agents on the economy.

Daniel Kronovet and Max discuss the world of information theory. We go into how it was invented by Claude Shannon at Bell Labs, the meaning of information and entropy, and the difference between measurement and processing. Catch references to a Jim Carrey movie and the Belmont stakes!

An article about MIT research claims that they've built a psychopathic AI. We think the article is a little off base, and try to determine what's actually going on. We cover related examples, of Microsofts Twitter Bot learning from the worst people, and a cool Google Image Program that sees animals everywhere. I also check in with Christian Hubbs, host of Artifically Intelligent, about his podcast. I was interviewed for todays episode, so check that out! I also define a Convolutional Neural Net, but only on the highest of high level! More on that could come in future episodes.

Run Your Mouth Podcaster and Stand-Up Comic Rob Bernstein joins me on the Local Maximum to talk about Comedy, Podcasting, Facebook, AI, Big Data, and of course MS Paint.

Today's concept is underfitting vs. overfitting. Someone who underfits doesn't adapt quickly enough to new data. But if you think that's bad, an overfitter will read TOO MUCH into new data and come up with wild ideas. How do you balance between them and what are some examples? We also talk about Warren Buffet's Recent statements on Bitcoin as Rat Poison Squared, and the response from Union Square Venure's Fred Wilson. How does Bitcoin get it's value? We look at underfitters and overfitters in the crypto space. Finally, the value of the Turing Test. When General AI comes, will it be the fabled machine that can pass as human, or will that idea be sidelined as more profitable uses taken precedent. Also, we learn what an Ornithopter is, and get an update on Yanny-Laurel.

Sarah Spagnolo tells us how we turn data insights into news stories. Does marijuana legalization affect bar-going? What's it like to go on TV? I also have to weight in on #teamyanny and #teamlaurel. Plus an update on the news stories we're tracking on Google and Facebook. How far along is Google's conversational AI demo?

A Washington Post OpEd believes that advances in AI will lead to a new centrally-planned economy. Google's Bots call unsuspecting businesses to save you time - but what are the implications? Bill Gates and Warren Buffet go sour on Bitcoin.

I talk to Mariya Yao about AI and Machine Learning as applied to a variety of business problems and across industries. We cover: - What are some of the misconceptions around AI? - If I have an AI product, how can I talk about it to business leaders? How do I sell it externally? How does it become a business? - Where are various industries on this? - And what are some of the pre-requisites for AI; I’m talking about data collection, and organizational issues.

Aren’t you getting tired of using an internet that’s built to waste YOUR time - I know I am! Product Manager Marissa Chacko and I discuss the promise and perils of these bots and push notifications. We worked on Foursquare's bot Marsbot together, and produced it to help the user first. What are some ways we can focus on building apps that help users, and what are some of the bad habits that have gotten us apps which aren't looking out for our interests?

Engineer and Podcaster Max Sklar shares his story about building software from his time at Yale to the Foursquare Innovation Lab. He’ll cover: What he has learned about working on vast location databases, building large-scale machine learning algorithms, and designing chatbots and apps that people actually want to use. Different ways to approach career and on working in teams His thoughts about where the tech/software industry is going in the next decade and encourage Yalies to follow up on their ideas.

Aaron comes back on the program to discuss the Mark Zuckerberg testimony before congress, particularly the part pertaining to moderation and censorship. In the last episode, I described the current approach that companies like Facebook, Twitter, and Google have towards this problem and why it might be flawed. Today, we see this theory play out in the Facebook CEO's answers. We discuss the case of Diamond and Silk, which I mentioned last week but had no idea would appear in the hearings. Also, Facebook's approach of ensuring that it's tools are used for good, and what this directive means for Facebook's future and how the internet will look in a decade. Also, I finished the 10 episode challenge! This podcast is all grown up now.

Today I answer the question of how I would handle moderation for Facebook and Twitter - and I break down the problems that they're facing and why. Don't hold your breath expecting them to try out my proposal! I also discuss Lindy's Law, which is a helpful rule of thumb in technological and social forecasting. As popularized by @nntaleb, it's another tool in your toolkit for understanding the world.

Today I post-game my discussion with Dennis Crowley. I share his goal of wanting people to execute on their ideas, but what are the common blockers? I talk to Aaron about how it broke his code, and the AI couldn't do it on it's own but only with my help - all because my brain was influenced by the power of suggestion. Plus - answer to a listener question about Marsbot!