I really enjoyed Fortune's Formula about Claude Shannon. It's about betting and information theory. I think successful entrepreneurship requires a nuanced understanding of risk taking. And specifically how you can create and edge. It helps you understand, once you know what kind of edge you have, how much of your bankroll you should deploy. Think about marketing your product, and finding an acquisition channel that has an insanely profitable payback. Are you willing to deploy half of your capital?

I think the mistake most founders make is to assume growth should be measured on a top-line month over month metric and the higher the number the better.

Folks like to say growth solves all problems. And that's generally right. But I think it's important to focus on accretive growth. The story of 2015-2016 was selling $1 for $0.80. We saw a lot of on demand startups fall victim to this fallacy. And the narrative today has shifted towards unit economic profitability. But think about a company like Instacart. It was rumored to be unit economic unprofitable early on, on its deliveries. But at the same time they became so large so quickly that Whole Foods took notice. So the unprofitable growth served a purpose. The me-too delivery apps faltered because the growth at all costs was a road to nowhere - other than bankruptcy. But for Instacart it was to hit a strategic objective of securing major partners. And they were fortunate to be able to fundraise in a frothy market. Fast forward to today, as they've rolled out annual membership fees and layered on tips for delivery personnel and increased service fe...

In general, channels that have clear and easy to measure ROI get quickly saturated and become priced efficiently - meaning you won't find any arbitrage. If you are looking for arbitrage, then you have to play in channels that are harder to measure directly.

A good example of this was influencer marketing 4-5 years ago. Michelle Phan pioneered selling makeup over youtube. But if you were Maybelline and you paid Michelle in her early days $10K to talk about your mascara, chances are you could measure the # of views you'd receive but be very hard to tie that back to revenue, to calculate ROI. I still think influencers are untapped potential for consumer brands. We can see the arbitrage starting to disappear as influencer marketplaces arise which make it easier to find and contact them but at the expense of a matchmaking fee.

The question to ask yourself is where are your target users spending time online and how can you reach them cheaply. We met an very interesting company that realized 1 year ago that they could scrape users' snapchat handles from instagram and create bots to snap...

OMGPOP had flash games on the web and towards the end of the company we started to make mobile games. There weren't many simple analytics tools we could buy so we had to build our own in house. In the beginning we stored all transactions in a mysql database and analytics meant SQL queries outputted to CSV with charts in MSFT excel. It wasn't pretty but it told us high level stats around the players in our games. So our KPIs were things like daily and weekly active users, day 1 and day 7 retention.

The next iteration was to define 200 queries we cared about, and have them update an analytics table overnight and use them to populate an internal dashboard. So we could see things like player retention in certain games. Or paying user conversion and lifetime values. We over rotated from having a few high-level KPIs to too many granular ones - so no one knew what mattered. Once we hit a few million users, this slowed down considerably. So we undertook a year long project to build out a scalable architecture using Vertica - a columnar s...

I can tell you about how we bought companies at IBM - an old tech dinosaur but surprisingly sophisticated on M&A process - as well as the story of how we sold OMGPOP to Zynga.

In the first case, at IBM, a public company, they really cared about the impact of an acquisition to their earnings per share (EPS). So for most early stage tech founders, there was no way to sell their business to IBM, because it wouldn't impact $B EPS and the machine wasn't set up to ingest a small startup and make use of the team or the tech. But one thing I found interesting is that nearly all acquisitions that went through, originated from a business unit executive who had nurtured a relationship over 2-3 years with a founder and who decided to purchase the tech or revenue to increase his/her business unit's top-line revenue. On the financial side, when you acquire a company, you are spending cash to acquire revenue, so your top-line revenue increases overnight without any impact to your expenses or operating cash flow. So if you are a business unit executive at a mature com...

Personally, I get excited by consumer services and "consumerized" enterprise software that put money back into consumers' pockets. That use technology to give more power back to individuals to get more done with less. I think about it from the perspective of the personal balance sheet. Every individual has a net worth which you can calculate as assets minus liabilities. So the question I like to ask is how can you use current trends in tech to impact global net worth? Well you can increase assets - and marketplaces do a good job of this by providing efficiencies to otherwise opaque transactions.

Think about life before OfferUp or Letgo. And I don't know OfferUp's specific metrics. But they gave people the ability to liquidate their used baby clothing. Or their outdoor patio furniture. These are users who may not have used craigslist on the desktop before. And these items were formerly held on their personal balance sheet fully depreciated - basically worthless. But the marketplace - now that there is liquidity and demand for these used products - has given money back to the consumer by creating a residual value for these produc...

I want to name my first born son touissant after watching this

Ashton Kutcher and Mila Kunis named their daughter Wyatt...so it's possible it could be a female character Ashton Kutcher reveals inspiration behind baby daughter Wyatt's name

Why can't the Delos ppl or other staff detect Bernard as a host by holding their tablet up to him, like they do with Maeve or Clementine? If that were the case, Ford would not be able to simply replace humans with hosts without getting caught.