el producto #372 🚀
How to use custom GPTs, Reddit IPO, OpenAI video generation, Apple Sports app, Creating a good pitch, How YouTube grows, How the biggest apps do onboarding & more
Hey friends,
Welcome to a new edition of el producto!
🎰 The week in figures
$2T: Nvidia reached a $2 trillion market value on Friday, becoming the third U.S. company to pass the milestone after Apple ($2.83T) and Microsoft ($3.06T). The chipmaker's rapid growth from the AI boom boosted its market value from $1T to $2T in just nine months, outpacing even Apple and Microsoft
1T: Adobe says there are over 3 trillion PDFs in the world. And it wants people to use its new AI feature to interact with them
$86B: OpenAI completed a tender offer, giving employees an opportunity to sell their shares at an implied valuation of $86B
$35B: Credit card issuer Capital One will acquire card network and issuer Discover for $35B
$675M: Figure AI Inc., a startup focused on developing human-like robots, is securing approximately $675M in a funding round, with a pre-money valuation estimated at around $2B
$444M: Monzo Bank is closing a £350M ($444) round that would value the British digital bank at £4B ($5B)
$100M: Eigenlayer, a DeFi service that provides developers with access to the Ethereum staked capital base and decentralized validator, raised $100M
$60M: Reddit will supply its content to Google for training AI models. The companies have reportedly signed a deal worth approximately $60M a year
3M: Substack says 3M+ customers subscribe to paid newsletters on its platform, up from 2M announced in 2023, and credits its recommendations feature for the growth
99%: South Korea is the most digitally connected country in the world, with 99% of adults saying they use the internet or own a smartphone. India was last on the same report, with 56%
💰 Q4 results
DoorDash: reported Q4 losses of $156M. The losses exceeded market expectations. DoorDash's adjusted EBITA stood at $363M, beating expectations of $356M. Total orders rose to $574M in Q4, up 23% YoY. DoorDash added 3M members to its DashPass and Wolt+ subscription services
Etsy: revenue $842.3M, up 4% YoY. Ads were a key revenue driver for the e-commerce platform. Gross Merchandise Sales of $3.6B, down 1.4% YoY. Net income fell to $83.3M, down from $109.5M at the same time last year. Active buyers on Etsy's marketplace rose 3% to 92M, an all-time high
Nvidia: adjusted EPS of $5.16, vs expected $4.64. Revenue $22.10B, and anticipates sales of $24B in Q1, surpassing forecasts of $22.17B. The company's "Hopper" chips, which are AI processors for servers, saw robust sales, driving a 265% increase in total revenue YoY. With a 769% increase over the previous year, the company's net income for the quarter was $12.29B. At $18.40B, data center sales — which now account for most of Nvidia's revenue — rose by 409%, with more than half going to major cloud providers
📰 What’s going on
Threads is testing its two “most requested” features: drafts and an in-app camera shortcut, making these available to a small number of people
Uber is testing a new feature that allows users to book a delivery for prepaid-items in local stores. Uber Eats is also rolling out its sidewalk robot food delivery service to Japan, thanks to a partnership with Mitsubish
WhatsApp rolls out support for bulleted and numbered lists, block quotes, and inline code to highlight and organize messages on Android, iOS, web, and macOS
Reddit filed for an IPO on Thursday, setting the stage for this year's first major tech debut. The company's U.S. listing, expected in March, will also mark the first social media IPO since Pinterest in 2019. According to its S-1 document, Reddit earned $804M in revenue, mostly from advertising, last year, up 21% from the year earlier. It still reported a $90.8M net loss, down from a $158M loss the year before. The company reported 73M DAU, over 100,000 active communities, and 1B total posts. Reddit is offering 75,000 of its most loyal users a chance to buy the firm's shares at listing prices
Google released a pair of lightweight open AI models for developers and researchers to experiment with and build apps. The "Gemma" text-generating models are now open for commercial use and redistribution, marking Google's shift toward openness in the AI space
Google is pausing Gemini's people-generation feature after the AI model produced historically inaccurate images. Gemini has faced criticism for portraying historical figures, including the U.S. Founding Fathers and Nazi soldiers, as people of color, in what seemed to be an effort to counteract AI's racial biases
OpenAI introduced Sora, its latest AI model designed for video generation, marking a move beyond text and images. Users can input a descriptive sentence, and Sora produces a high-definition realistic video clip with additional capabilities such as generating videos based on still images, extending existing videos, or filling in missing frames
Microsoft releases PyRIT, a tool its AI Red Team has been using to check for risks in its generative AI systems like Copilot, to the public
Spotify follows Meta, YouTube and others by offering AUX, a service to connect brands and creators
TikTok announces its “Add to Music app” feature, which lets users add a song playing on a clip to services like Apple Music and Spotify, in 163 new countries
Apple launches Apple Sports, a free app offering scores and stats for the NBA, Premier League, and more, in the US, Canada, and the UK, on iOS 17.2 or later
Apple Music debuts a monthly version of its Replay experience, letting users see their monthly music habits and access listening insights
Apple is removing the ability for European users to install web apps on their home screen. It says it’s too difficult to keep the feature and comply with the EU’s latest piece of big tech regulation, the Digital Markets Act. The legislation targets “gatekeepers” of certain technologies, such as operating systems and browsers and forces them to offer viable third party alternatives. In this case, since home screen web apps are often created using Safari, allowing them to exist without an alternative could mean Apple violates the rules
The first human recipient of a Neuralink brain chip can now control a computer mouse using their thoughts, according to founder Elon Musk
📚 Good reads
YouTube’s VP of Product on how to propel product growth. YouTube is now the top streaming service in the US with almost 9% share of all streaming viewers. In this interview, YouTube’s VP of product management Nicky Rettke, explains how the product has grown so big so quickly
How the biggest apps design onboarding: lessons from Figma, Duolingo and more, by Felix Lee and Ben Shih. The guidebook from Figma, Duolingo, and more. From common mistakes in designing onboarding to secrets behind engaging onboarding
You should be playing with GPTs at work, by Lenny Rachitsky. 20 examples of how people are using custom GPTs to make their teams more productive. Including a short guide to create your own GPT
The Continuous Product Improvement Cycle. This article by Jeff Patton breaks down the various aspects of the continuous product improvement cycle. It offers a great overview of the key components, including sensing (continuously listening and learning from our customers), focusing (using vision, strategy, and outcomes to guide our work), discovering (analyzing, designing, and testing our ideas before we fully invest in them), and delivering (actually building and releasing new products)
The pitfalls of TAM (total addressable market) analyses, and how to think more critically about market sizing. According to CJ Gustafson, many TAM analyses lack credibility and are based on lazy top-down approaches, often lacking analytical rigor and real-world applicability. Companies may inflate TAM numbers without understanding the underlying market dynamics, leading to unrealistic assessments. Assessing TAM through lenses such as the initial insertion point, level of competition, and potential for revenue expansion helps in understanding the true market opportunity. It's important to deconstruct TAM analyses to uncover the real revenue potential and competitive positioning of a company
The rules of a good pitch. Jean de La Rochebrochard argues that you’ve got 1 minute to make a good first impression, 5 minutes to build credibility, and 15 minutes total to convince. And how long you talk for drives all other elements you need to nail
That’s all for this week. As usual, feel free to reach out and share your thoughts by replying back to this email or commenting on Substack
Thanks for your support!
Angel