The Founder of Hinge Thinks Your Voice Is the Best Dating Profile You Have
AI is reshaping how we connect, how we search, and how we confide — and today’s stories show just how personal that shift is getting. A new dating app wants to match you based on how you sound. Google wants to know what images you love before you even search for them. And a 60-year-old chatbot experiment keeps explaining something we still find surprising: people tell machines things they won’t tell other people. Here’s what’s happening and why it matters.
A New App Wants to Match You by Your Voice, Not Your Face
Justin McLeod, who built Hinge into one of the most popular dating apps in the world, is back with a new idea. His new venture, Overtone, has raised $18 million to build a dating platform where voice and audio are the main attraction — not photos, not witty bios. According to TechCrunch, the premise is that hearing someone’s voice tells you more about compatibility than a carefully curated photo grid ever could.
The idea has real logic behind it. Voice carries tone, warmth, hesitation, humor — things a photo simply cannot. Overtone uses AI to analyze and match users based on audio profiles, essentially treating the sound of your voice the way other apps treat your face or your interests. Think of it like Spotify’s recommendation engine, but instead of matching you to a song, it’s matching you to a person.
For anyone who has grown weary of swiping through photos and getting nowhere, this is an interesting alternative. Voice-based matching could reduce the pressure of physical first impressions and surface connections that feel more natural from the start. It also raises new questions about how much audio data a dating app should be allowed to collect and store.
Why this matters: Dating apps have barely changed their core formula in a decade. A voice-first approach, backed by real AI analysis, could push the entire category in a genuinely different direction.
“The founder raised $18 million for an AI dating service focused on voice and audio matching.”
Google Turned 25 and Decided You Should Stop Searching for Images
Google is 25 years old this year, and to mark the occasion, the company has overhauled how its image search works. Rather than waiting for you to type a query, the updated experience now surfaces images based on your personal interests and refreshes that content automatically. Ars Technica reports that the shift represents a move from traditional search toward something closer to a personalized visual feed.
This is a meaningful change in philosophy. Traditional search is reactive — you ask, it answers. What Google is building now is proactive: the system learns what you like and brings it to you before you think to look. It works by combining your past search behavior with AI models that predict what images you’d find interesting or useful.
For everyday users, this could feel either delightful or a little unsettling, depending on how you feel about Google knowing your taste. On the practical side, it means spending less time digging through irrelevant results. On the privacy side, it means Google is building a more detailed picture of your preferences with every scroll.
Why this matters: This is a preview of where all search is heading — less typing, more anticipation. The question is whether users will embrace it or push back.
“Google updates image search with personalized, auto-refreshing content for its 25th anniversary.”
Why People Keep Telling Their Secrets to Chatbots
Long before ChatGPT existed, a computer program named ELIZA was already revealing something strange about human nature. Built in the 1960s by MIT professor Joseph Weizenbaum, ELIZA was a simple chatbot — a program designed to mimic conversation without actually understanding it. Weizenbaum was disturbed to find that people poured genuine emotions into their exchanges with it, treating the machine like a trusted confidant. Wired’s analysis draws a straight line from ELIZA’s waiting room to the millions of people now sharing personal struggles with ChatGPT.
The reason this happens isn’t mysterious. Machines don’t judge. They don’t gossip. They’re available at 3 a.m. with infinite patience. The AI doesn’t flinch at what you say, and that absence of reaction turns out to be remarkably comforting for many people.
For anyone who has caught themselves being unusually honest with an AI assistant, this context is reassuring — and a little sobering. These impulses are deeply human. What’s new is the scale, and the fact that the machine on the other end is now far more sophisticated than ELIZA ever was.
Why this matters: Understanding why people confide in AI isn’t just interesting history. It shapes how designers build these tools, and what responsibilities they carry.
“ELIZA showed people were willing to share personal thoughts with machines in the 1960s.”
Also Happening in AI
Google DeepMind CEO Demis Hassabis made headlines at The Verge by calling for a global AI watchdog agency — one he believes should be led by the United States — with the authority to pause dangerous AI development. Separately, Hugging Face quietly shipped a maintenance release of its widely-used Transformers library (version 5.13.1) to fix compatibility issues with vLLM, a tool developers use to run AI models efficiently. Over at TechCrunch, reports emerged that OpenAI researcher Miles Wang is in talks to launch an AI drug discovery startup valued at $2 billion, adding to a growing wave of researchers leaving top labs to build in healthcare. Meanwhile, Spotify is testing a chatbot for Premium users that helps them find and play music through conversation. And Y Combinator-backed Agnost AI is building software that automatically pulls user feedback out of AI agent conversations — a small but telling sign of how many businesses are now managing AI interactions at scale.
What to Watch
The Overtone launch and the ELIZA retrospective are pointing at the same underlying question: how much emotional territory should AI occupy in our personal lives? Watch for how regulators — and the public — respond as AI moves deeper into relationships, mental health, and self-disclosure. Hassabis calling for a global watchdog this week suggests that even insiders believe oversight needs to catch up with where the technology already is.