Adversarial Learning

It's been a couple of years, so Joel and Andrew catch up on what's new in life and in data science.

It turns out Joel wrote another book: Ten Essays on Fizz Buzz (Quarantine does strange things to a person.) In this episode Andrew and Joel discuss Fizz Buzz, what makes the book interesting, the value of deconstructing problems in different ways, Joel's side career in acting, why fluffy talks are easier to write than technical talks, the importance of staying fresh, whether "blocklist admin" is the job of the future, Minecraft classes, and how the quarantine is affecting our kids. Please listen to it.

About a month ago we recorded a podcast with Jowanza Joseph about engineering and COVID-19 and other mundane topics. While I (Joel) was procrastinating on editing it, a lot happened in the world. And then Jowanza wrote a blog post about his experiences with racism and how they've affected his life. Which made it feel really weird to just put out an episode with him that doesn't mention any of these issues. So the three of us went back into the (metaphorical) studio and recorded another episode mostly about his blog post, about racism, and about the state of the world. (As always, we are incapable of staying 100% on-topic, but we did pretty good.) These aren't necessarily easy topics to discuss, but I think we did a respectful job of it, and I hope you find it interesting and worthwhile. The original podcast episode follows the new one, so you get two episodes in one. Please listen to it.

Joel and Andrew break quarantine (metaphorically) to discuss treehouses, remote work, distance schooling, outschool.com, Joel's attempt to teach his daughter Python, old-school text adventures, socially-distanced eating, the Twitter UI, what happens to school in the fall, OKRs, and whether we should keep the economy closed or re-open it and kill people. Please listen to it.

in which Joel and Andrew briefly break quarantine to talk about Covid-19, social distancing, and the data science grift

Our guest this week is Pardis Noorzad (@djpardis), former data science manager at Twitter and now Head of Data Science at Carbon Health. Our conversation spans a wide range of topics: the Tenderloin Carbon Health what Duran Duran has to do with Black History Month the "cold start problem" for becoming a DJ how long it takes to achieve domain expertise how many people there are in Canada how getting that first data science job is like DJ-ing loyalty the Go programming language the Hopper-Causey effect data science as quality control Please listen to it.

Our guest this episode is Carl Gold (@carl24k), Chief Data Scientist at Zuora and author of the forthcoming book Fighting Churn with Data. Use the discount code podadvl19 to save some unspecified amount on it. Topics include * churn * writing a book * churn * running a subscription business * churn * domain expertise * p-hacking * churn * how Joel picks what song to use for the podcast intro * churn

On this episode, Andrew is joined by Joel Grus (@joelgrus), the author of Data Science from Scratch, whose second edition just came out. They discuss spreadsheets, writing a book, Python 3, type annotations, Jupyter notebooks, reproducibility, and what it's like to do standup comedy at a conference that has a code of conduct. Please listen to it.

Our guest this episode is data scientist Peadar Coyle. Topics include Brexit whether Europe is a Python continent or an R continent GDPR how data science is different in Europe than it is in the USA Joel's stock rant against "domain expertise" PyMC3 and Bayesian Analysis teaching online courses whether "primer" is pronounced "primer" or "primer" what Joel and Andrew are most excited about in data science Please listen to it.

Adversarial Learning is back from its extended hiatus! Our guest is famous data scientist Josh Wills. We discuss why Josh is a famous data scientist, what it's like working at Slack, data science conferences, NLP's "imagenet moment", whether Joel should remove the MapReduce chapter from the 2nd edition of Data Science from Scratch, and which is the best Rush album. Please listen to it.

Adversarial Learning is back! In this long-delayed episode (thanks, technical difficulties) we are joined by data scientist Schaun Wheeler to discuss our favorite topic, data ethics. Highlights include: * Schaun's Medium post "An ethical code can’t be about ethics" * Do we need a "Hippocratic Oath" for data science * How to hire data scientists who won't steal people's kidneys * Why Joel has a Values Mug * The Manifesto for Data Practices * Is this all secretly a competency problem? * Skin in the Game * Are data ethics issues really just business ethics issues? Please listen to it!  (More episodes coming soon!)

Our guest this week is data scientist for good Lisa Green.  Topics of discussion include What is ethics  Joel's previous life as a financial analyst and the ethical dilemmas therein whether there's anything incriminating in Joel's Yahoo history "identity theft" as a bullshit concept  Google's corrupt bargain with the NHS what the medical code of ethics actually says  polycentric ethics the difference between unethical and incompetent what good a code of ethics does when the "ethical" problems are emergent from the choices of many people that terrible article about the Seattle Nazi convention neuroticism  the Joel test  vgr's bad tweet Please listen to it.

Our guest this week is flashcard kingpin and former Partially Derivative co-host Chris Albon. Topics of discussion include how good it feels not to have a podcast machine learning flashcards being a "natsec bro" whether Chris would punch a Nazi whether Chris would sexually harass a Nazi whether "magister" is a good woke replacement for "master" whether that Andrew Ng job posting is appropriate and whether any of us would apply for it killing your heroes having a day a week without social media treadmill desks Chris's next podcast and somehow Joel gets going on Harry Potter Please listen to it.

Data scientist Vicki Boykis joins Joel and Andrew to variously debunk and rebunk common Data Science Myths. Is data the new oil? Do data scientists spend 80% of their time munging and cleaning data? What happens if you look in the mirror and say "data science" five times? And many, many, many more. Please listen to it.

Andrew and Joel come out of hiatus to discuss data conferences: when to attend them, how to get your talk accepted, how to network, optimal heckling strategies, where to stay, and so on. Somehow they also end up talking about fidget spinners, the "objective" section on resumes, the right way to use LinkedIn, why Andrew doesn't think much of data science bootcamps, and why Joel can't convince any data science bootcamps to sponsor the podcast. Please listen to it.

Friend of the podcast Tim Hopper joins us as we share stories of Very Bad Interviews we've been on. (As you probably expect, Joel has the most humiliating stories.) If you've ever gone on a terrible interview, listen and commiserate. If you've never gone on a terrible interview, listen and live vicariously. Halfway through, Andrew's Internet flakes out and his part stops getting recorded. Thanks to the magic of editing, you'll hardly even notice!

Our guests this week are Curtis Yarvin and Galen Wolfe-Pauly, which means that our topic is Urbit ("a virtual city of general-purpose personal servers"). What is it?  Why is it? Is is a political project? And does it have anything to offer data science types? Curtis and Galen try to explain what Urbit is and answer Joel's objections, while Andrew keeps trying to tie everything back to ham radio.

Our guest this week is Trevor Grant (@rawkintrevo), an Open Source Technology Developer Evangelist (or similar) at IBM. We discuss how to be a high-energy public speaker, all sorts of weirdly-named Apache projects, the "my name is" meme, why Joel and Andrew don't like Jupyter-style notebooks, Voltron, and how to talk to your kid about "normies". Please listen to it.

Our guest this week is Juliet Hougland (@j_houg), data scientist and engineer at Cloudera. We discuss that bad Wired article about physics and software engineering, why Juliet knows so much about Urbit, censorship (Twitter and otherwise), dark patterns (LinkedIn and otherwise), and why none of us was savvy enough to start a social network for data people and raise $19M.

Is good data science always political? Our regularly scheduled guest had to cancel at the last minute, so Joel and Andrew decided to talk about politics.  Topics covered include: the one data scientist who’s not on twitter how much Joel hates the xkcd cartoon free speech in Canada (which Joel gets very wrong, please don't @ me) how one (or one's bot) gets banned from twitter Joel getting a phone call from his kid's school #YarvinGate and #YarvinGate2 how good it feels to hold a pitchfork loving one's neighbor vs tolerating one's neighbor why everyone hates Kant Nazi-punching the pros and cons of street violence whether companies should take political stands whether Shopify should power the Breitbart store what the hell do they even sell in the Breitbart store anyway? the difference between refusing to do business with customers because you don't approve them and refusing to business with customers because twitter doesn't approve of them