Did a friend send you this newsletter? Be sure to subscribe here to get a weekly post covering the technology behind products we use every day and interesting applications of data science for folks with no tech background.
The New York City heat is becoming bearable, Starbucks started selling their pumpkin spice lattes, the streets are filled with runners training for marathons — summer is coming to a close and fall is almost here! For this week’s Day to Data, we’re doing a short wrap up for the last month of summer — what I’m reading, what’s interesting, and more thoughts.
And, for the first time ever, Day to Data is taking off next week! We’re going to be enjoying the holiday weekend and working on new ideas for the fall. Thanks for following and reading Day to Data, we’ll be back soon!
Like, comment and subscribe
This month I’ve been reading Like, Comment, Subscribe: Inside YouTube's Chaotic Rise to World Domination by Mark Bergen. It’s a dive into the history of the video-player behemoth, their time as Silicon Valley’s star child, the blunders of content moderation, getting bought by Google, and more. I’ve been most fascinated learning about the evolution of their recommendation algorithm and ads business. They’ve utilized machine learning to build their “ads machine” by knowing how ads impact watch time, Youtube’s ultimate KPI. Bergen explains the engineers’ simple logic —
If YouTube played an ad for a viewer as soon as she visited the site, she would likely leave. But wait for her to linger.. then show her an ad, and she was more tolerant of commercial interruptions.
What might seem like a simple deduction was only clear with machine learning. The update to their ad showing algorithm became “a freight train that couldn’t stop”, per one former executive. A little ML can go a long way for a business.
The most precious gift: GPUs
Yep, GPUs are all the rage. GPUs (graphics processing unit) are making headlines every day. Better known by the umbrella term “chips”, the demand for GPUs has skyrocketed with the rise of large language models and generative AI. GPUs are a hot commodity for startups and massive enterprises, with Nvidia at the helm of the ship. Analysts estimated $12.5 billion in revenue for Nvidia this quarter, but their latest update put that number close to $16 billion. My fun fact for you all about GPUs? Companies that have been lucky enough to get their hands on some compute are using it as collateral. Two founders are putting GPUs up as collateral so they can raise $25M. It’s not the first time — Coreweave raised $2.3B collateralized by Nvidia’s chips. The latest Hard Fork gave a great rundown of just how far people are going to get some compute. We’ll dive into more of this dynamic in coming articles.
IPOs are back
In the past month, the SEC saw some big movement with IPOs and S-1s filed. 2022 went down as the worst year for IPOs since 1990, with numbers down 88% from 2021. Some notable IPOs came through the market in August, so let’s talk about them.
Better.com finally went public. Better planned to go public in May 2021, then the great-Zoom-layoff-debacle happened and the SPAC was delayed, and the signs were mixed whether the IPO would ever come. They finally IPO’d this past Thursday, then their stock tanked 93% on its first day.
Klaviyo, the marketing automation company, filed an S-1 on August 25. Founder Andrew Bialecki broke the record for the largest ownership stake of a company at IPO at 38.1%, just passing the Atlassian co-founders who both had 37.7%. Shopify held 11.2%. Klaviyo’s biggest announcement? It’s profitable.
Instacart, lesser known as Maplebear Inc., filed an S-1 on August 25 also. The filing claims orders grew 18% from 2021 to 2022, and remained consistent from 132.3 million for the six months ended June 30, 2022 to 132.9 million for the six months ended June 30, 2023. Gross profit grew 49% from 2021 to 2022. People are definitely still ordering their Pad Thai and bagels.
Keeping today’s edition short and sweet. Thanks for reading and making it through August with me! Lots of exciting stuff on the horizon for the coming year — see you then!
Short and sweet from tall and sweet.