Around a year ago, Christie’s, a British auction house, sold its first piece of computer-generated art, titled “Portrait of Edmond Belamy,” for a cool $432,500.

I have always been intrigued by the notion of computer-generated art. However, every time I try to think of the term, “computer-generated art” something seems off. Somehow, it sounds like an oxymoron to me. Let’s look at why. Here’s the definition of art from Google search:

Image for post
Image for post
Portrait of Edmond Belamy

“Art is the expression or application of human creative skill and imagination, typically in a visual form such as painting or sculpture, producing works to be appreciated primarily for their beauty or emotional power.” …


What are adversarial attacks?

In 2014, a group of researchers at Google and NYU found that it was far too easy to fool ConvNets with an imperceivable, but carefully constructed nudge in the input. Up to this point, machine learning algorithms simply didn’t work well enough for anyone to be surprised when it failed to do the right thing. But by 2014, ConvNets had become powerful enough to start surpassing human accuracy on a number of visual recognition tasks.

Image for post
Image for post
Source: Explaining and Harnessing Adversarial Examples, Goodfellow et al, ICLR 2015.

Let’s look at an example. We start with an image of a panda, which our neural network correctly recognizes as a “panda” with 57.7% confidence. Add a little bit of carefully constructed noise and the same neural network now thinks this is an image of a gibbon with 99.3% confidence! This is, clearly, an optical illusion — but for the neural network. You and I can clearly tell that both the images look like pandas — in fact, we can’t even tell that some noise has been added to the original image to construct the adversarial example on the right! …


Books (20):

  • The Great CEO Within: The Tactical Guide to Company Building. Matt Mochary.
  • This Is Water. David Foster Wallace.
  • The Great Mental Models: General Thinking Concepts. Shane Parrish.
  • 21 Lessons for the 21st Century. Yuval Noah Harari.
  • The Dip. Seth Godin.
  • On Intelligence: How a New Understanding of the Brain will Lead to the Creation of Truly Intelligent Machines. Jeff Hawkins and Sandra Blakeslee.
  • This is Marketing: You Can’t be Seen until You Learn to See. Seth Godin.
  • Tools of Titans: The Tactics, Routines, and Habits of Billionaires, Icons, and World-Class Performers. Tim Ferriss.
  • An Elegant Puzzle: Systems of Engineering Management. …


As many of you may know, Deep Neural Networks are highly expressive machine learning networks that have been around for many decades. In 2012, with gains in computing power and improved tooling, a family of these machine learning models called ConvNets started achieving state of the art performance on visual recognition tasks. Up to this point, machine learning algorithms simply didn’t work well enough for anyone to be surprised when it failed to do the right thing.

In 2014, a group of researchers at Google and NYU found that it was far too easy to fool ConvNets with an imperceivable, but carefully constructed nudge in the input. Let’s look at an example. We start with an image of a panda, which our neural network correctly recognizes as a “panda” with 57.7% confidence. Add a little bit of carefully constructed noise and the same neural network now thinks this is an image of a gibbon with 99.3% confidence! This is, clearly, an optical illusion — but for the neural network. You and I can clearly tell that both the images look like pandas — in fact, we can’t even tell that some noise has been added to the original image to construct the adversarial example on the right! …


Image for post
Image for post

In Antifragile, I mentioned that the book is part of Nassim Nicholas Taleb’s Incerto series. In 2018, Taleb added Skin in the Game to this series. I hopped onto the audiobook over the past week and was happy to find that while sticking to his fuck-you writing style, Taleb did a great job expanding on the ideas introduced over his last 3 books. Listening to this book felt like a quick summary, or rather, a re-read of his last three books and although no radically new ideas were presented here, it was still totally worth it.

I’ll keep this one short and pick only five of my favorite ideas from the…


Image for post
Image for post

Never Split the Difference is the first book I’ve read about the art of negotiation — the other book closest to this one was Influence by Robert Cialdini (review). In this book, Chris Voss, a former international hostage negotiator for the FBI, lays out his approach to high-stakes negotiations and offers a few key lessons that you can incorporate in your daily routines. Negotiation is the heart of collaboration, and is what makes conflict potentially meaningful and productive for all parties. …


Image for post
Image for post

The Sprint process was developed at Google Ventures by Jake Knapp and John Zeratsky. It’s a highly effective process for solving problems by building prototypes and user-testing them over a span of five days. Sprint is almost a required reading for everyone in the tech industry — from the executives to the engineers. I’ve recommended this book to more people than I can remember and recently realized that I haven’t published an official review of this book here, or have one place with all the links to my writings on it. I’m rectifying that this week.

First, here’s a small 90-second video summarizing the…


Image for post
Image for post

I read Adam Grant’s first book, Give and Take, in August 2017. It introduced me to his excellent framework for building your professional (and personal) network. It’s probably the most valuable book on “networking” ever written.

The core idea introduced in the book is that in any network, there are three kinds of nodes — the givers, the takers, and the matchers. …


Image for post
Image for post

In this book, Adam Grant breaks down what “being original” means by studying habits of highly creative professionals and dispelling a number of myths related to originality. His other famous book, Give and Take, taught me a lot about what professional networking should be — so I decided to give Originals a shot too. Here’s what I learnt:

How to generate original ideas

The key to generating a few good ideas is to try to generate a lot of them — if you do so, some of them will inevitably turn out to be good ones. It’s better to not try and perfect mediocre ideas, be prolific instead. The author gives examples of Edison (who had thousands of patents) and Mozart (who composed thousands of pieces) — generating a massive body of work seems to be the key to ending up with a few hits in the process. We are a terrible judge of our ideas — continuously test out your own with your target audience and work on the feedback. Seek out experiences that broaden your perspective by engaging with orthogonal fields. Lastly, procrastination is great to avoid premature optimization. …


Image for post
Image for post

Like most people in the tech industry, I’m a huge fan of Jason Fried and David Heinemeier Hansson (DHH) — two programmers who co-founded a small web app development agency called 37Signals in 1999, and over time grew it to multi-million dollars in profit through their Basecamp product. In fact, they have renamed the company to Basecamp now). Moreover, DHH is more popularly known as the inventor of Ruby on Rails, the immensely popular web development framework at the core of products like AirBnb, GitHub, Shopify, and of course, Basecamp itself. The duo have written four books so far — the first three being Getting Real, Remote and Rework. …

About

Anant Jain

Tech, startups, engineering, product design, etc. Now @brexhq . Past: Co-founder @commonlounge . @iitdelhi ’12. Moved from Medium to https://anantja.in

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store