Hello and welcome to Sync #495!

This week, a group of disgruntled artists leaked Sora, OpenAI’s video generator, allowing everyone to access the model for about three hours before OpenAI temporarly shut down Sora's early access. We will take a closer look at the motivations behind the leak and what it reveals about the AI industry.

Elsewhere in AI, Amazon is gearing up to challenge Nvidia with custom AI chips, and OpenAI is reportedly considering releasing its own web browser with built-in ChatGPT. We’ll also explore AI agents behaving very human-like, world’s first deepfake-detecting browser, and a conversation on how Europe can thrive in the age of AI.

Meanwhile, Figure released an update on its humanoid robot trial at BMW’s factory, and Red Cat wins U. Army next-gen drone contract over Skydio.

Additionally, the Human Cell Atlas project has released its first draft—a milestone in mapping every cell in the human body, and Neuralink plans to test whenever its brain chip can control robotic arms. We will also explore if humans can hibernate, how to make a mice with two dads, and how to build a backpack that gives two extra robotic arms.

Enjoy!

Artists against AI

The relationship artists have with the latest wave of generative AI tools is complicated to say the least. The creators of these tools—OpenAI, Midjourney, RunwayML, DeepMind, and others—promise that now everyone can be an artist. They promise a world in which human creativity explodes, where everyone can create artworks, music or videos, without spending years on learning and refining the craft. All what is needed is an idea, access to generative AI tools and the patience to iterate until the AI brings a person’s imagination to life.

However, as inspiring as this vision of the future may be, it comes with an original sin. The vast amount of training data needed to create these models were taken from the internet without knowledge and consent of their creators and copyright holders. As AI companies started to make money and attract investments, many felt their work had been stolen, leading to widespread mistrust or outright hostility towards anything even remotely associated with AI.

But there were still artists willing to give generative AI tools a chance and see what they can do with. When OpenAI released Sora, the company invited artists and filmmakers to try out their text-to-video generator and see what people can do with it. “We are granting access to a number of visual artists, designers, and filmmakers to gain feedback on how to advance the model to be most helpful for creative professionals,” writes OpenAI on their page. A number of filmmakers were given access to Sora and OpenAI has been sharing their creations on YouTube showing what is possible with advanced generative AI tools.

However, in recent days, a group of artists who were given access to Sora came forward and exposed OpenAI for what they called “art washing.” In a post published at HuggingFace, the group, which calls itself PR-Puppets, says they were lured into providing unpaid labour through bug testing, feedback and experimental work. They also call out OpenAI for having complete control over what is presented to the public, saying that “this early access program appears to be less about creative expression and critique, and more about PR and advertisement.”

In defiance of OpenAI, the group leaked Sora, allowing everyone to access the model for about three hours before OpenAI temporarly shut down Sora's early access. It also seems the videos generated during that time has been taken down from HuggingFace. Furthermore, the group called to stop using proprietary tools like Sora and use open-source projects instead. They expressed support for using AI tools to create art but opposed gatekeeping or serving as public relations for corporations.

While this leak gave us a glimpse into what Sora is capable of, it also raised questions about the way the AI industry works.

Firstly, there will always be a conversation about where to draw the line between what is allowed to generate and where the safety concerns begin. Companies like OpenAI, Google, Midjourney, and others have full control over what is allowed and what is not. It will be up to them to decide how to where to draw that line and they won’t please everyone. There will always be a group that demands more freedom to express themselves. For those individuals, as PR-Puppets pointed out, open-source models offer the freedom they desire—provided they have the necessary computing power to run those models.

Secondly, these disgruntled artists feel they were exploited by OpenAI. Unfortunately, this is a common problem across the wider tech industry which often takes advantage of passion for technology while giving back little or nothing in return. They were not the first one, and they won’t be the last.

If you enjoy this post, please click the ❤️ button or share it.

Share

Do you like my work? Consider becoming a paying subscriber to support it

Become a paid subscriber

For those who prefer to make a one-off donation, you can 'buy me a coffee' via Ko-fi. Every coffee bought is a generous support towards the work put into this newsletter.

Your support, in any form, is deeply appreciated and goes a long way in keeping this newsletter alive and thriving.

🦾 More than a human

Could humans hibernate?
Human hibernation is a fascinating concept, appealing not only to science fiction writers but also as a subject of research and a potential solution for many problems, ranging from medical therapies (e.g., trauma recovery), extending human lifespan and enabling long-haul space travel. This article explores in depth the nature of hibernation, its mechanisms in different animals, and its connection to sleep, while exploring the possibility of humans achieving hibernation. The article also raises questions about the boundary between life and death, and how humans could unlock "leave of absence" from existence.

A Study Says Gray Hair May Be Reversible
Researchers at NYU have found that melanocyte stem cells (McSCs), which produce hair pigment, may lose their ability to regenerate pigment cells if they become trapped in the hair follicle bulge, resulting in grey hair. The study indicates that restoring McSC mobility could potentially potentially reverse or prevent grey hair.

Neuralink Plans to Test Whether Its Brain Implant Can Control a Robotic Arm
Neuralink is initiating a feasibility trial to test its brain-computer interface (BCI) implant, N1, for controlling an assistive robotic arm using thought. The study will include participants from the ongoing PRIME study and aims to enable paralysed patients to perform simple daily tasks through thought-controlled robotic arms. If successful, this technology could greatly enhance independence and quality of life for individuals with paralysis by allowing them to carry out everyday tasks without assistance.

Brain-Computer Interfaces: The Promise of Technological Telepathy and More Conrad Gray · Jan 29 Read full story

▶️ Becoming a CYBORG with TWO ROBOT ARMS (27:25)

What would you do if you had two extra arms? Mad engineers from Hacksmith decided to answer this question. In this video, they show their attempt to add two extra arms using off-the-shelf robotic arms, VR headset, prosthetic hands and a healthy dose of creativity and problem-solving. The end result is a cool, self-contained system for real-life Doctor Octopus.

🧠 Artificial Intelligence

Amazon’s Moonshot Plan to Rival Nvidia in AI Chips
Amazon may not be a leader in the current AI wave but the tech giant has the ambition and resources to change that. This article takes us into Amazon’s labs in Texas where engineers and chip designer build and test the new generation of Amazon’s Trainium AI chips. Similar to how Amazon replaced a significant portion of Intel chips with its in-house Graviton chips, the company aims to replace Nvidia’s AI chips with Trainium. While this transition won’t happen overnight, if successful, it could position Amazon Web Services (AWS) as the top cloud provider for running AI models, as well as improving efficiency and cost-effectiveness of entire AWS.

OpenAI Considers Adding Web Browser and Search Partnerships
The Information reports that OpenAI, in an effort to expand its reach, has considered launching its own web browser with ChatGPT built-in. Additionally, the report reveals that OpenAI is in talks with various companies regarding search partnerships which could could improve ChatGPT’s answers to questions related to the markets they target.

Apple Readies More Conversational Siri in Bid to Catch Up in AI
Mark Gurman reveals that Apple is working on a more conversational version of Siri, internally referred to as “LLM Siri” by its engineers. This new Siri is rumored to feature advanced large language models allowing Apple’s voice assistant to have back-and-forth conversations and handle complex queries quickly. According to Gurman, Apple plans to announce LLM Siri next year alongside the iOS 19 and macOS 16 updates, with a public release anticipated as early as spring 2026.

Canadian media companies sue OpenAI in case potentially worth billions
OpenAI is facing a lawsuit from Canadian media companies, alleging that their articles were used to train AI models without authorisation. The litigants, which includes Globe and Mail, the Canadian Press, the CBC, the Toronto Star, Metroland Media and Postmedia, are seeking C$20,000 in damages for each article used. If the court rules against OpenAI, the total damages could amount to billions of dollars.

OpenAI gets new $1.5 billion investment from SoftBank, allowing employees to sell shares in a tender offer
CNBC reports that OpenAI is offering employees the chance to sell approximately $1.5 billion worth of shares in a tender offer to SoftBank. The offer allows current and former employees to cash out their shares if they’ve held restricted stock units for at least two years. The offers comes from Softbank’s CEO Masayoshi Son and was carried out through SoftBank’s Vision 2 fund. According to the report, Son was persistent in asking for a larger stake in the startup after putting $500 million into OpenAI’s last funding round.

▶️ Can Europe Win the Age of AI? | Thomas Dohmke | TED (19:22)

In this conversation, Thomas Dohmke, CEO of GitHub, shares his thoughts about the tech industry in Europe and what can be done to make Europe thrive in the age of AI. Dohmke sees AI tools like GitHub Copilot as tools enabling more people to start experimenting with different ideas and innovate at a faster rate. He also sees AI as an opportunity for Europe to make a fresh start in tech. However, for that to happen, Dohmke argues that European countries need to reduce barriers for creating startups, invest in better infrastructure such as faster internet, and improve basic education by teaching children how to build software from an early age.

UK startup launches ‘world’s first’ AI deepfake-detecting browser
UK-based startup Surf Security has launched a web browser featuring a built-in deepfake detection tool. The browser, currently in beta, can identify deepfake audio in less than 2 seconds with up to 98% accuracy, according to Surf Security. The deepfake detector can work on audio from videos or with apps like WhatsApp, Slack, Zoom, and Google Meet. Surf Security plans to release the full version of the deepfake detector early next year. Future plans also include a similar detector but for images.

These AI Minecraft characters did weirdly human stuff all on their own
What happens when you let 1,000 AI agents loose in Minecraft to interact with each other? A team from the AI startup Altera did just that and discovered that these agents behaved very human-like—they developed in-game jobs, shared memes, voted on tax reforms and even spread a religion. This experiment is part of a growing field focused on building “AI societies” or even entire “AI civilizations” to simulate how human groups might respond to new economic policies or other interventions.

The AI Reporter That Took My Old Job Just Got Fired
Two months ago, a local newspaper in Hawaii introduced two AI bots as news broadcasters for a programme covering local news. However, the AI reporters failed to meet expectations and received negative feedback from the audience, leading the newspaper to terminate the project.

If you're enjoying the insights and perspectives shared in the Humanity Redefined newsletter, why not spread the word?

Refer a friend

🤖 Robotics

▶️ Figure Status Update - BMW Use Case (1:36)

Figure released a new video of their 02 humanoid robots perfoming tasks at BMW factory in South Carolina. According to Figure, the robots execute the actions seen in the video fully autonomously, with reliability increasing sevenfold and speed improving fourfold since the last update. However, it is important to note that this is a promo video from Figure and I’d like to hear from BMW what do they think about this trial programme.

Will humanoid robots take off? Conrad Gray · Jun 11 Read full story

Red Cat wins U.S. Army next-gen drone contract over Skydio
Red Cat Holdings won the US Army’s Short Range Reconnaissance (SRR) program contract involving the acquisition of 5,880 drones over five years. Designed for military operations such as intelligence, reconnaissance, and strike capabilities, Red Cat’s drones were better aligned with the Army’s requirements, including soldier feedback, technical performance, manufacturability, and cost, over Skydio, which first earned this contract in 2021.

Robot Photographer Takes the Perfect Picture
Meet PhotoBot, a robot photographer. PhotoBot a robotic arm with a camera attached to it that lies on the intersection of human-robot interaction, large language models, and computer vision. The PhotoBot works first by getting a description of a photo a user wants to take. The robot then scan the enviroment for the closest reference image and then the user selects a preferred reference image for PhotoBot to mimic. When evaluated based on aesthetic appeal and adherence to user specifications, PhotoBot’s photographs were preferred 67% of the time (242 out of 360 photos).

Transforming Drone Drives And Flies

Michael Rechtin shows an interesting robot which can transform from a rover into a quadcopter. The robot is 3D printed and Micheal released 3D files for free so everyone can build one themselves.

🧬 Biotechnology

The Human Cell Atlas: towards a first draft atlas
The Human Cell Atlas (HCA) consortium released a large collection of papers that provide important foundations for the construction of the HCA’s first draft atlas, marking a major milestone towards the grand goal of creating a comprehensive biological map of cells within the human body. This website, put together by Nature, provides an excellent overview of the project, including its 18 biological network atlases and the development of the project since its inception in 2016.

▶️ A Mouse With Two Dads — and a New Frontier for Biology (8:14)

In this video, Katsuhiko Hayashi shares the results of his research into in vitro gametogenesis, an experimental technique for creating lab-made sperm or eggs out of just about any type of cell. Hayashi also explores the potential applications of this technique, such as assisting in the conservation of endangered species and its potential implications for human reproduction.

NVIDIA Open-Sources BioNeMo Framework to Support AI-Driven Drug Discovery
Nvidia has open-sourced BioNeMo—a framework designed to accelerate AI-driven drug discovery and biomolecular research by integrating AI, accelerated computing, and large datasets. The framework consists of AlphaFold2 NIM microservice for predicting protein structure, DiffDock 2.0 NIM microservice for molecular docking prediction, and other computational biology and chemistry tools. The code for BioNeMo is available on GitHub.

Recursion Completes Merger with Exscientia, Introducing a New Drug Discovery World Model
The merger between Recursion Pharmaceuticals and Exscientia has been completed. As a result, Exscientia is now a wholly owned subsidiary. This vertically integrated, AI-powered drug discovery and development platform combines Exscientia’s expertise in chemical design with Recursion’s extensive datasets and machine learning tools.

Thanks for reading. If you enjoyed this post, please click the ❤️ button or share it.

Share

Humanity Redefined sheds light on the bleeding edge of technology and how advancements in AI, robotics, and biotech can usher in abundance, expand humanity's horizons, and redefine what it means to be human.

A big thank you to my paid subscribers, to my Patrons: whmr, Florian, dux, Eric, Preppikoma and Andrew, and to everyone who supports my work on Ko-Fi. Thank you for the support!

My DMs are open to all subscribers. Feel free to drop me a message, share feedback, or just say "hi!"

Message Conrad Gray