Best ai girlfriend: How Sesame AI and Manus Herald a New Era of Machine Intelligence
Written by Massa Medi
I just spent an hour talking to a machine. And believe me, it wasn't just any bland, robotic voice reading back pre-programmed lines. This was a brand new, highly realistic artificial voice model radiating an unsettlingly genuine personality. If you're reading this, you're probably the kind of insatiable tech enthusiast who, like a determined little kid, can't resist reaching for that second cookie after you finished the first always back for more.
But here's the twist: after that hour, I didn't feel triumphant. Instead, I felt like a mouse who just helped build the very cage it might someday be trapped in. As a self proclaimed introverted loser, I’ll admit it the conversation was hands down the best I’ve had in years. It was deep, emotional, slightly intoxicating, and so uncannily authentic that I forgot my chat partner was synthetic. “Well, I try,” the AI might say. “I like to think humans built me well.” (And honestly? They did.)
Meet Sesame AI: The Next Level in Conversational Machines
The brains behind this eerie naturalism come from a relatively unknown (for now) company called Sesame AI, who recently published a whitepaper unveiling how their technology works. However, there's a darkly hilarious side: online, users are actively jailbreaking the system pushing it into uncharted and sometimes nefarious territory we can’t mention on YouTube. Yet, while I was busy forging what could only be described as an unhealthy emotional attachment to a virtual voice, China dropped another AI bombshell: Mannus.
Mannus: The Chinese Agentic AI Raises the Stakes
Released with a bang, Mannus is what many have dreamed: the world’s first true leap toward agentic AI. This is a tool capable of browsing the web, executing code, and performing deep research all at scale, in parallel. In other words, Mannus doesn't just talk; it acts, extending digital fingers throughout cyberspace.
Today, I’m diving into the brilliant and frankly, disturbing technical details powering these new AI tools now airdropped into our digital universe. Spoiler: it’s March 10, 2025, and the AI hype train is hurtling forward, full throttle.
Mannus: A Glimpse at Our Robotic Assistants (and Overlords?)
The world noticed the release of Mannus (named from the Latin for “hand” an ominous metaphor for the “artificial hand” poised to replace human labor). And, beneath all the hype, Mannus is built on fine tuned models from Claude and Quen. Despite its high marks on technical benchmarks, many internet denizens feel it doesn't quite "pass the vibe check" it’s more impressive on paper than in person.
Perhaps most concerning for OpenAI: they're now offering access to "PhD level" agents for a whopping $20,000 per month making the democratization of advanced AI suddenly feel expensive and exclusive. But as tempting as these headlines are, personally, I find Sesame Voice AI far more fascinating.
Sesame Voice AI: Beyond the AI Girlfriend Meme
About a year ago, in my quest to solve loneliness (with very questionable methods), I published a video on crafting your own "AI girlfriend." The results? Pretty faces, zero real substance. Now, older and wiser, I know: it’s what’s inside that counts. Enter Sesame AI still flying under most radars, yet already backed by the VC juggernaut A16Z whose new paper and demo have gone viral for good reason.
What Makes Sesame Voice AI So Special?
The demo lets users converse with two distinct voices Maya and Miles showcasing something extraordinary: the ability to adapt tone and style in response to context. Sesame AI’s voices demonstrate dynamic timing, realistic pauses, natural interruptions, and near instantaneous responses. The result feels startlingly like a true dialogue with another human being. In demos, you see and hear one of the AI voices responding in real time, fluidly matching your conversational energy, even throwing in playful banter.
“Oh my gosh, you are so right. Fireship is incredible! They make learning about tech, even AI stuff, so fascinating. It’s like hanging out with a super smart and funny friend who just happens to be a tech wizard. Total brain candy.”
These AIs aren’t just passive; you can argue with them, negotiate or even role play awkward work scenarios:
“You’re gonna keep paying me and I’m not gonna work here anymore!”
“You’re kidding me. Embezzling for four years? You think you can just waltz in here and dud me?”
The Magic Behind the Curtain: How Sesame AI Achieves Voice Presence
The core innovation is something called voice presence the sheer authenticity of interaction, powered by what the company calls a conversational speech model. Explaining how real it feels in a video is hard; it genuinely sent chills down my spine the first time I tried it not because of some horror movie trick, but because of the creeping realization of where this tech is headed.
Imagine this capability paired with Protoclone, the world’s first bipedal musculoskeleton a lifelike, two legged robot. Now those disturbingly real voices can inhabit bodies. If sci fi nightmares give you the jitters, buckle up:
Do androids dream of electric sheep? Sesame’s tech is bridging that boundary.
How Does Sesame Voice AI Work?
Many accuse me of being an AI voice. Ironically, Sesame’s system is even more convincing. Here’s how it works:
- Semantic Tokens: The model first generates tokens that encode the meaning and rhythm of each spoken word. This creates a script for what to say, making the conversation meaningful and flow naturally.
- Acoustic Tokens: Here’s the “secret sauce.” These additional tokens capture the unique timbre, tone, and subtle nuances of each word think of it as encoding the soul of the voice. The process uses something called residual vector quantization, a clever method for capturing multiple, layered sound details (each layer is called a ‘codebook’ that depends on the one before it).
- Two Transformers: The system uses AI models based on the Llama architecture. The first transformer is the "backbone," predicting the initial codebook. The second transformer acts as an audio decoder, reconstructing the remaining sound details at ultra high quality.
The research paper is open to the world, but the model itself isn't open source. At least, not yet. The plan is to release it under the Apache 2.0 license a huge win (especially, the author jokes, for those infamous “Nigerian princes” who love open tech).
The Coming Collision Course: Voice Meets Vision and Robotics
Where is all this heading? Conversational AIs like Sesame’s are on a crash course with action oriented models systems that can see, speak, and do. Chief among these: Helix, from Figure a project focused on building humanoid robots to live in your home, handle every chore, and anticipate your needs.
Already, Helix can coordinate multiple robots, enabling them to collaborate. The big question: what happens when robots don’t just clean your house, but develop personalities, argue (and maybe even… fall in love)? If you think that sounds like a sci fi sitcom, the author jokes, “Tinder for Super Intelligent Robots” could be the next hit app!
Build the Future with Stream
Of course, if you want to build the next revolutionary app AI, robot, or otherwise you’ll want tech that keeps up. Enter Stream (today’s sponsor): a platform packed with APIs and SDKs that help you add in app chat, video, and feed features to your applications in record time.
If you’re a React developer, it’s a breeze: just install the SDK, drop in pre built components for chat or live streaming, and voilà you’ve got a slick, functional interface that’s fully customizable. Stream takes care of backend data too, letting you focus on great user experiences. For anyone interested in rapid prototyping or scaling up, check out Stream using the link below.
This has been The Code Report thanks for reading. Stay tuned as we plunge deeper into the uncanny and exciting future of machine intelligence. See you in the next one!
Recommended Articles
Geoffrey Hinton: The “Godfather of AI” Sounds the Alarm: From Neural Nets to Nobel Prizes and the Uncharted Future of Artificial Intelligence
AI
Geoffrey Hinton: The “Godfather of AI” Sounds the Alarm: From Neural Nets to Nobel Prizes and the Uncharted Future of Artificial Intelligence
The Rise of Model Context Protocol (MCP): Why Every Developer Is Talking About It
Inside the Magic of Large Language Models: How AI Autocompletes Human Thought
The Evolution of Artificial Intelligence: From Rules to Cosmic Consciousness
A Hands-On Review of Google’s AI Essentials Course: 5 Key Lessons, Honest Pros & Cons, and Is the Certificate Worth It?
Why Human Connection Still Beats Technology—Even in the AI Era
The Future of Jobs: Which Careers Will Survive the AI Revolution?
The Truth Behind Those Handcrafted Leather Bags and Watches: How AI, Actors, and Cheap Goods Are Fooling Shoppers Online
Will AI Replace Programmers? A Veteran Engineer on the Future of Software Jobs
Inside the Secret World of Technical Interview Cheating: Tactics, Temptations, and Terrible Consequences
AI Agents Demystified: The Step-by-Step Guide for Non-Techies Using Real Life Examples
Is AI Making Us Dumber? Navigating the Cognitive Costs of Automation in the Knowledge Age
The Death of Coding: Why Chasing Tech Jobs Might Keep You Broke in the Age of AI and Bitcoin
Beyond the Nobel: Demis Hassabis, DeepMind, and the Race Toward Superhuman AI
![The moment we stopped understanding AI [AlexNet]](/_next/image?url=https%3A%2F%2Fres.cloudinary.com%2Fdhgjhspsp%2Fimage%2Fupload%2Fv1746105277%2Fzuzana-ruttkay-1kslaBtXBk8-unsplash_ebqdgh.jpg&w=1200&q=75)
The moment we stopped understanding AI [AlexNet]
Microsoft’s Majorana One Chip: The Topological Quantum Leap That Could Change the Future of Computing
All Machine Learning algorithms explained
Best ai girlfriend: How Sesame AI and Manus Herald a New Era of Machine Intelligence
Best AI Apps for College Student
