Should your app be more social?

gm legends. It’s Sunday, Sunday, Sunday.

In this edition, we talk about music platforms’ push to add social media elements, introduce you to an AI assistant that’s actually proactive, examine the breaks on AI models, and then tiptoe around anything potentially uncomfortable (perhaps like a certain LLM) until we get to the most popular new products this week.

Get a cup of coffee, and we’ll bring the sugar. Now let’s get reading. 

P.S. Launching soon? We’d love to hear about it → editorial@producthunt.co 🫶

SoundCloud went live 18 years ago as a launchpad for independent artists to share their music. Now, it’s adding more tools for listeners to share what they’re into. 

The music streaming app, which ranks above streaming stalwarts Pandora and Apple Music in terms of monthly average users, introduced new social features on Thursday, including:

• “Liked By Your Crew”: a daily playlist based on what your friends (or the artists you follow) like 
• “Trending Trackwall”: similar to X’s “What’s happening”
• Recommendations of users to follow

Spotify, which once upon a time made noise about buying SoundCloud, amped up its own social features just a few months ago when it began allowing subscribers to direct message each other.

It’s not like either streaming app is turning up the social element to 11. Users are still going there for the music…but the tweaks could make it a little easier for communities to form around shared fandom. 

We’re curious: What are your favorite social media tweaks to predominantly non-social apps? And the ones you find completely inessential?

A month ago, our Head of Product Curation, Gabe Perez, was on the search for a good app and some good content. He tried out Poke, a new AI assistant that he found to be proactive instead of reactive.

The review period is over, and Gabe remains a total convert. He writes:

“It's become my default AI assistant. It proactively tells me where my attention is needed, helps me with my diet, makes sure I'm not late to meetings, and tells me when an important meeting is booked....all..without me... prompting it.”

Musa Molla asks: What’s the one thing that always breaks first when you take an AI project from demo to production?

You know that feeling. You’ve got the PowerPoint loaded and ready, but when it’s time to give the presentation, suddenly there’s no sound, there’s no video, and there’s definitely no applause.

Musa sees a similar thing at play with AI tools. “The demo looks magical…and then everything breaks when real users hit it.”

He wants to know the failure points you’ve encountered so that maybe y’all can get around them together.

Last week, Ashok Nayak asked: Do you think GPT-5 plays it safe compared to earlier LLMs?

Ashok has a hunch the new model is “holding back” when it comes to sensitive topics like politics. He put up a pool to see if anyone was picking up similar vibes. And, well, I don’t know how to say this, but…