Meta’s AI Slop, Spotify ghost music and the end of SaaS pricing as we know it?
Plus: Instacart smart fridges, 2025 forecasts, Tinder returns to growth
Hey everyone 👋, welcome back. I hope you’ve had a great start to the year so far and a big hello to the 200+ new subscribers who have joined us since the last briefing!
This week, we take a look at Meta’s bold product decision to allow AI users onto their social networks and the potential impact of this on the UX of Meta’s products, as well as an in-depth look at Spotify’s practice of adding “ghost music” to playlists to boost profitability.
We’ve also shared a new tool you can use to identify dark patterns UX and a new tool for transforming docs into AI-ready JSON formats along with some personal highlights from 2024.
Enjoy the rest of your week!
More here: Watch on YouTube | Follow on Substack Notes | Upgrade to unlock the full newsletter
Key reads and resources for product teams
New from the Department of Product Substack this week:
Knowledge Series - How to write SQL queries using AI
Writing SQL queries with the help of AI so that you can quickly get the data you need without necessarily being a pro at writing SQL queries. This Knowledge Series explains how to write some of the most useful SQL queries with the help of AI tools like ChatGPT, Perplexity and Claude.
Deep - The UX of Search Explored
The ability to quickly find exactly what you need when you’re using a product can be transformative; get it right and activation and engagement rates may increase. Get it wrong and you risk losing a user who becomes frustrated with not being able to find exactly what they need. In this Deep dive, we explore how the UX of in-product search is evolving with examples from 20+ leading companies including: LinkedIn, GitHub, Linear, Miro, Perplexity, Threads and more.
(Department of Product)
Case study - How Spotify uses “ghost music” to drive revenues
Spotify’s business model relies on paying artists a share of revenues based on the number of streams it gets. However, according to this Harper's Magazine investigation, Spotify developed a new strategy to improve margins by filling popular mood-based playlists with cheaper stock music from production companies, rather than tracks from traditional artists. Spotify's product team created an internal dashboard where playlist editors could track the percentage of PFC music in playlists, and eventually formed a dedicated "Strategic Programming" team to systematically increase PFC's presence across hundreds of playlists. (Harper Magazine)
UX - A Dark Patterns Identifier tool
This mini game will help you spot sneaky design tricks used in various products. Developed by Rohan Dehal, this free interactive experience is designed to help you instantly spot dark UX patterns used by top companies to manipulate user behaviour and understand how these patterns exploit cognitive biases. (Product Artistry)
Interview - Why Salesforce isn’t hiring software engineers this year
Salesforce’s CEO explains how the company’s new Agentforce works - and why they are not hiring any more engineers in 2025. (20VC)
Resource - A 2025 Trends Forecast Pack
This collection includes a summary of hundreds of reports from different sources and summarized the output using ChatGPT. The sources include the likes of Bain, Gartner, Forrester, Goldman, Google, Instagram, HSBC and many more. (Notion Doc)
New product features and innovation this week
Meta has made a clear strategic product decision to allow - and promote - AI generated users across its social platforms. "We expect these Als to actually, over time, exist on our platforms, kind of in the same way that accounts do… They'll have bios and profile pictures and be able to generate and share content powered by AI on the platform... that's where we see all of this going” said Connor Hayes, VP of product for generative Al at Meta.
As part of this series of announcements, Instagram unveiled some powerful new generative AI capabilities for Instagram. Powered by Meta’s AI video model, MovieGen, a video shared by Instagram’s chief Adam Mosseri demonstrated how you can use AI to change your backgrounds, add some new props and even accessorize with a hat or chain. You can watch it in action here.
At this point, nobody really knows what it means when the web - and particularly social products - are flooded with so-called “AI slop”. Early indications from the success of products like character.ai suggest that users are more than willing to engage with AI personas. The only major downside for Meta is that AI slop can’t click on ads. But, if they drive enough engagement, plenty of humans probably will.
In other news
Instacart has announced a multi-year partnership with Samsung that will let users shop for groceries on their screens of their Samsung refrigerators. The integration uses Samsung’s Vision AI food recognition technology to identify what you have in your fridge to determine what items you’re running low on. Samsung says the locally-based AI can recognize “up to” 37 food items including fresh fruits and vegetables. It then uses Instacart’s product matching APIs to suggest items from Instacart that you may need and add them to your next order.
This combination of hardware, AI and traditional open APIs working together is a smart example of how emerging technologies can work together to create entirely new value propositions and could be something we start to see a lot more of in 2025.
Meanwhile…
Microsoft and OpenAI have agreed on a definition of artificial general intelligence. According to reports, the companies have agreed that OpenAI can terminate its partnership with Microsoft when AGI has been reached. The AGI definition it uses is this:
“highly autonomous systems that outperform humans at most economically valuable work.”
This seems to be a pretty solid definition of AGI - but the economic consequences for humans sound grim.
Personal highlights from 2024
Few people could have predicted the rise of products like ChatGPT and it’s possible that the next big thing will appear out of nowhere in a similar Black Swan event. 2024 was quite the year for new entrants who challenged the dominance of Google in Search, Meta in social and more.
Here’s a reflection on some of the products, tech and other things I’ve enjoyed most in 2024:
Tools you can use
Monkt - convert PDFs, Word Docs, Powerpoint, Excel, Web pages and more into JSON formats to help feed AI learning.
Lecca - build a team of AI agents to perform workflows on your behalf.
Norm - ensure your product meets relevant compliance regulations by giving your team access to AI agents with specialised expertise.
📈 Product data and trends to stay informed
Software and new product development remain the top areas for applied generative AI, according to the latest Bain quarterly survey of business executives. 66% of respondents say generative AI is either in production or development for code creation, and 65% report using it for core product development enhancements.
Meta says its Llama model has become the most used open source AI model with more than 650 million downloads - twice as many as the previous 3 months.
Tinder added more paying users for the first time since 2022 last quarter. It now accounts for 57% of Match Group’s total revenue and its turnaround strategy is relying heavily on “advanced AI” that it says will help people find matches more easily. Despite its challenges, over 40% of dates start on apps. Full investor day deck for all product updates
Microsoft's AI segment is forecasted to surpass $10 billion in annual revenue run rate by Q2 fiscal 2025.
Databricks officially grabbed the biggest AI funding round of 2024 with $10 billion raised. It is now the fourth mostly highly valued US startup after OpenAI, SpaceX and Stripe.
Seat-based SaaS pricing models may be nearing their end. As companies reduce human involvement in customer support thanks to AI, fewer seats are needed. As a result, outcome-based pricing could emerge as the dominant SaaS pricing trend in 2025. Companies could charge for outcomes like ticket issue resolution, for example.
Other product news in brief
✍️ Meta has appointed Republican Joel Kaplan to lead its global policy team
🤖2025 could be the year of the humanoid robot
🎵Anthropic’s Claude will no longer use copyrighted song lyrics to train its AI
🧑⚖️California has enacted a law which prevents tech companies from serving addictive feeds to children
Paid subscribers get the full DoP Substack including: DoP Deep dive reports to learn from the world’s top tech companies and The Knowledge Series for sharpening their tech skills, AI tutorials for putting AI into practice at work.
There's a space for AI characters like with Glambase influencers but I don't understand what Meta's trying to achieve with theirs at all. They are literally competing with real creators - people that real people actually want to engage with
Really a lot of interesting stuff! Meta's integration of AI across the platforms, especially in the more everyday uses of Instagram users, could really impact how people think about this type of technology (if the "AI" element is highlighted).
Also, I found the partnership between Instacart and Samsung intriguing. Together with General Electrics I imagine it will be one of the first products that will act en masse on people's grocery purchases, and how it will affect or people will integrate it into such a daily and repetitive activity is certainly something to keep an eye on. Thanks for sharing this curated and great selection.