Replit Co-Founder Haya Odeh on the Artist and the Machine Podcast

"If something doesn't look beautiful and engaging and intriguing for the eyes, then what's the point?"

Theme

Most talk about AI and software focuses on the technical barrier coming down. In our conversation with Haya Odeh, Replit's Co-Founder and Head of Design, a different barrier came up – whether people see themselves as someone who builds things at all. Replit noticed that users would describe an idea, watch it start to take shape, and then hesitate when a button asked if they were ready to build it. Changing the language in the user interface kept more of them going. For tools aimed at people who don't think of themselves as technical, choices like that end up deciding who sticks with it.

Signal

What happens after people get started is just as interesting. Replit's agent responds in whatever vocabulary you bring, whether that's design or engineering language, and it explains what it's about to do before it does it, keeping the human in the driver’s seat. Odeh hears from users who picked up terms like "information architecture" from the agent and now use them in their own work. For creative teams, that means a well-designed tool can train designers in engineering and engineers in design, but only if it explains its work rather than just returning a result.

Podcast

Dani sat down with Haya Odeh, Co-Founder and Head of Design at Replit, which is also the platform our own website was built with. She talks about designing for people who never planned to make software, why her team spent a full day on a loading animation, and what it was like to watch a dark mode she designed in 2018 get copied widely.

"If something doesn't look beautiful and engaging and intriguing for the eyes," she says, "then what's the point?"

After Hours

Our virtual series After Hours is back this Thursday with special guest, Blake Finucane. Blake is an art historian who covers all things art, technology, and where the two thread together. We’ll be digging into AI, creativity, and so much more.

Join us live at 10am PT / 1pm ET everywhere we are (Instagram, Linkedin, YouTube) — or add the session straight to your calendar now.

LA Summit

We're in Los Angeles on December 2 with UMG's Michael Nash, Meow Wolf's Vincent Kadlubek, 1X's Dar Sleeper, Amazon's Jo Shoesmith, Gossip Goblin and more, followed by an invitation-only Studio Day on December 3.

Pulse

  • Hey, It's Taylor Swift: Swift's rights company has filed to trademark her voice in the UK, the EU and Australia, following its US filings in April. Australia accepted a sound mark of her saying "Hey, it's Taylor Swift" on October 12, six days after filing. Trademarking a spoken voice has never been tested in court.

  • Seventy People and Some AI: Grimes describes her new album Psy Opera as AI-inspired but not AI-generated. Her team has since confirmed that the video for "Accelerando" includes some AI-generated elements, made alongside a human crew of more than 70.

  • Goldman Runs the Numbers: Goldman Sachs' latest Music in the Air report estimates AI licensing, creation tools and new products could add about 10 percent to recorded music revenue by 2035. It also warns that millions of AI-generated tracks could pull royalties away from human-made music.

If you’re creating something aligned with Artist and the Machine or you’d like to recommend work shaping this space, we’d love to see it for the chance to be featured. Reach out to us at [email protected].

Til next time,

Artist and the Machine.

xx