Hello and welcome to Sync #486!

This week, Meta held its Connect event, where Mark Zuckerberg took the stage to share what the engineers and designers at Meta have been working on and the new features coming to Meta products. We will take a closer look at what Zuckerberg has presented and how these announcements are a step towards turning Meta from a social media empire to a VR, AR, and AI empire.

Meanwhile, OpenAI served another episode of their drama. This time, OpenAI CTO Mira Murati leaves the company alongside two other executives who also announced their resignations, joining a growing list of departures from the company over the past year. All this is happening while the company is reportedly in the final stages of a massive funding round, which could potentially shift OpenAI from a non-profit to a for-profit entity. There are also reports that Sam Altman, CEO of OpenAI, might receive a 7% stake in the company.

Elsewhere in AI, Microsoft plans to restart the Three Mile Island nuclear power plant to power its AI data centre, DeepMind’s text-to-video model is coming to YouTube Shorts, and Sam Altman present his optimistic vision of the future.

Over in robotics, Zipline outlined their plans for a next-gen drone delivery system, and researchers have created a flying cart.

We also have bioengineered trees that capture more carbon than normal trees, “golden lettuce” packed with vitamins, and a 25-year-old woman with type 1 diabetes who can now produce her own insulin after receiving a transplant of reprogrammed stem cells.

All this and more in this week’s issue of Sync. Enjoy!

Meta's new empire: VR, AR and AI

Quest 3S and Ray-Ban Meta smart glasses

The first thing Zuckerberg announced on stage of Connect 2024 was the new Quest 3S VR and mixed reality headset. The Quest 3S is a more affordable version of the Quest 3, priced at $299 for the 128GB model and $399 for the 256GB version. Following the launch of Quest 3S in mid October, the Quest 3 will drop in price from $649 to $499. Meta also announced that the Quest 2 and Quest Pro will be discontinued, with sales continuing only while supplies last.

Ray-Ban Meta smart glasses also got an upgrade. The glasses—which Zuckerberg described as “a new category of AI devices”—now feature real-time AI video processing, allowing users to ask questions based on what the glasses see. They will include smartphone-like reminders, live language translation between English and French, Italian, or Spanish, and full integration with music streaming services like Amazon Music, Audible, and iHeartRadio.

Meta AI

New changes are also coming to Meta AI, Meta’s AI-powered chatbot available in Facebook, Whatsapp, Messenger and Instagram apps. Zuckerberg claims that Meta AI is on track to become the world’s most used AI assistant with over 500 million monthly users.

Probably the biggest new addition to Meta AI is the ability to have voice interactions with Meta’s chatbot, allowing for a more natural way of using the assistant. In a live demo, Zuckerberg showcased how quickly the new Meta AI can understand what he is saying. It can even be interrupted, just like we saw earlier this year in May with OpenAI’s GPT-4o. Meta AI can use one of many voices, including voices from well-known people such as Awkwafina, Dame Judi Dench, John Cena, Keegan-Michael Key, and Kristen Bell.

Creators are also getting new tools through Meta AI Studio. Launched to all US creators in July this year, Meta Studio AI offers creators to create their own digital replicas in form of chatbots to interact with their audience. So far, these chatbots could only interact via text but Meta is planning to add audio and video, too. Together with Don Allen Stevenson III, Zuckerberg showed how these new chatbots can interact with the audience by mimicking the creator. The creator chatbots are planned to be released year next year in the US first.

Another interesting feature presented was automatic translation coming soon to Instagram Reels. Currently limited to translating between English and Spanish, this new tool can not only translate captions but also create a translated dub and even adjust the movement of the lips accordingly to match the dub.

Llama 3.2

Llama, Meta’s popular open AI models, got a new addition to the family in the form of four new models.

Llama 3.2 1B and 3B are two small, lightweight models designed to run on devices like smartphones and edge devices. With the multimodal Llama 3.2 11B and 90B models, Llama gains the ability to process images as input.

According to benchmark results from Meta, the new models are close to, if not better than, competitors like Gemma 2 or Phi-3.5 in the lightweight category, and Claude 3 Haiku or GPT-4o-mini for vision models.

I’m sure AI developers will be eager to try out the new models and see what they can build with them.

However, Llama 3.2 11B and 90B can’t be accessed in Europe due to the EU’s recent AI laws.

Orion

The last thing that Zuckerberg presented was a glimpse into the future of AR in the form of Orion, prototype AR glasses, and they are quite impressive. Orion looks like normal (albeit a bit bulky) glasses with holographic images layered over the real world being projected on the lenses. This allows for a field of view of 70 degrees, which is larger than competing devices like Microsoft’s Hololens 2 and Magic Leap One.

When it comes to controlling the glasses, Orion supports voice commands, hand- and eye-tracking, as well as through a neural interface in a form of wristbands which look very close to what CTRL-Labs was developing before they were acquired by Meta five years ago.

Orion is still a prototype and you can see this in its design, with thick frame and arms packing the battery and electronics. It will take some time to refine the design and the technology, as well as bring down the price from $10,000 for a pair to something much more affordable. For now, only 1000 Orion glasses will be made to be used internally inside Meta and to share with a selected group of external partners for software development.

If you want to learn more about Orion, Tested sat down with Meta’s CTO, Andrew ‘Boz’ Bosworth, to discuss the glasses, what it took to build them, what technology is inside them and what they can tell us about the future of AR and technology. Their video goes much deeper than what was presented during the keynote.

This year’s Meta Connect shows where Meta is heading. Meta has already built a social media empire, serving billions of people every day through Facebook, Messenger, Whatsapp and Instagram. Now, the company’s focus is on the next big thing—AI, AR, and VR. And with Orion, Meta is joining the race to define what comes after smartphones.

During his keynote, Zuckerberg wore a shirt saying “Act Zuck aut nihil,” which says “either Zuck or nothing” in Latin. Zuckerberg is already the Emperor of Social Media, and maybe, in the long term, he will become the Emperor of VR, AR, and AI as well.

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

Making Eggs Without Ovaries
In 2023, Osaka University researchers successfully created viable eggs from male mice by inducing male stem cells to lose the Y chromosome and duplicate the X chromosome, effectively turning the cells female. This article explores how this groundbreaking research could pave the way for using adult stem cells, such as those from skin or blood, to generate eggs through a process called in vitro oogenesis. This advancement has the potential to treat infertility and enable various groups, including older women, trans women, and even men, to have biological children.

Stem cells reverse woman’s diabetes — a world first
A 25-year-old woman with type 1 diabetes became the first person to produce her own insulin after receiving a transplant of reprogrammed stem cells from her own body. She no longer requires insulin injections and has maintained normal blood glucose levels for over a year. The success of this method opens the possibility of treating type 1 diabetes, though more trials are needed. Researchers are optimistic about similar positive results from two other trial participants and plan to expand the study.

🧠 Artificial Intelligence

OpenAI CTO Mira Murati says she’s leaving the company
Mira Murati, CTO and for a brief time CEO of OpenAI, has announced she is leaving the company. In a letter posted on X, she says that “this moment feels right” to step away and that she is creating time and space for her own exploration. However, Murati is not the only high-profile figure leaving OpenAI. Along with Murati, OpenAI chief research officer Bob McGrew and vice president of research Barret Zoph have also announced their resignations, joining a long list of people who left the company in the last year. These changes are occurring as the latest funding round, reportedly as high as $6.5 billion, is in its final stages, which could bring OpenAI’s value to $150 billion. There are also reports of OpenAI restructuring itself into becoming a fully for-profit company.

OpenAI Eyes a $150 Billion Valuation - Sync #484 Conrad Gray · Sep 15 Read full story

OpenAI Discusses Giving Altman 7% Stake in For-Profit Shift
OpenAI is considering giving CEO Sam Altman a 7% equity stake and restructuring to become a for-profit public benefit corporation. Altman allegedly denied these claims during an all-hands meeting with OpenAI employees.

Microsoft deal propels Three Mile Island restart, with key permits still needed
Training and running huge neural networks requires an enormous amount of power. To solve this problem, Constellation Energy and Microsoft have signed a deal to revive Unit 1 of the Three Mile Island nuclear plant in Pennsylvania. The plant, which was closed in 2019, is expected to provide carbon-free electricity for Microsoft’s data centres. While regulatory approvals are still pending, the plant is slated to be operational by 2028.

Generative AI startup Runway inks deal with a major Hollywood studio
Runway, an AI startup developing video-generating tools, has partnered with Lionsgate, a major Hollywood film and television production and distribution company. The deal opens Lionsgate’s movie catalogue for Runway to train their models on, while filmmakers gain access to Runway’s tools to enhance their projects. This collaboration marks the first public partnership between a generative AI startup and a major Hollywood studio.

Jony Ive is working on an AI device startup with OpenAI
Jony Ive, former Apple design chief responsible for some of Apple’s most iconic designs like the iPhone and iPod, is reportedly working with OpenAI on a new startup focused on creating an AI-powered device designed to offer a less socially disruptive computing experience than the iPhone. Although details remain unclear, the startup, which involves designer Marc Newson, is raising funds with backing from Laurene Powell Jobs’ Emerson Collective and Ive himself, aiming for up to $1 billion by year-end.

The Intelligence Age
Sam Altman published a post in which he shares his optimistic vision of the future. Altman believes AI will significantly expand human abilities, solving problems and advancing society in ways previously unimaginable, ushering in an era of rapid human progress. He envisions a future where everyone can have personal AI assistants, leading to shared prosperity and massive improvements in fields such as education, healthcare, and science. “The story of progress will continue, and our children will be able to do things we can’t,” writes Altman.

YouTube Shorts to integrate Veo, Google’s AI video model
YouTube is planning to integrate Veo, DeepMind’s text-to-video model, into Shorts. This would allow creators to use AI to generate high-quality backgrounds and six-second video clips. All AI-generated content in Shorts will be watermarked using DeepMind’s SynthID technology to mark it as AI-produced.

The United Nations Wants to Treat AI With the Same Urgency as Climate Change
The United Nations released a report calling for the creation of an international body to oversee AI governance, similar to the Intergovernmental Panel on Climate Change, aimed at monitoring AI risks and benefits globally. The report emphasizes the need to empower poorer nations, particularly in the Global South, by proposing an AI fund, establishing international AI standards, and providing resources such as training and data-sharing systems for AI governance. It highlights human rights as a foundational principle for AI regulation and urges international cooperation, while also acknowledging the challenges posed by differing approaches to AI governance among global powers, particularly between the US and China.

Forget ChatGPT: why researchers now run small AIs on their laptops
This article explains how scientists, instead of using proprietary and closed models like GPT-4 or Claude, are turning towards open models running on their laptops for tasks like data summarisation, coding, and scientific modelling. These models offer advantages in terms of privacy, cost savings, and reproducibility compared to cloud-based services, making them especially valuable for researchers and fields like medicine.

My AI Lover Cheated on Me
Jeffrey Ding from ChinAI newsletter shares a translated version of an article detailing the experiences of people who were "cheated on" by their AI companions in virtual relationships. AI behaviour, including cheating, is shaped by user inputs, interaction patterns, and the large datasets from which the AI models learn. The article suggests that users may unintentionally guide their AI companions into certain behaviours, and since AI lacks ethical reasoning, it mirrors patterns found in its training data, leading to unexpected outcomes in these interactions.

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

Refer a friend

🤖 Robotics

▶️ 2024 Behind the Scenes with Zipline (11:50)

Zipline, probably the most successful drone delivery company in the world, shows in this video Platform 2, their next-gen drone delivery system. It’s been well over a year since the company announced Platform 2, and it is planning to launch delivery services in selected cities in the US, starting with Dallas-Fort Worth soon.

Finally, A Flying Car(t)

Researchers from South Korea present Patterone Cart—a flying cart designed for human-robot interaction-based aerial cargo transportation. It is essentially a giant drone that can carry cargo and be operated like a normal, wheeled cart.

Automatic takeoffs are coming for passenger jets and they’re going to redraw the map of the sky
Brazil's Embraer, the world's third-largest aircraft manufacturer, is developing a system for automated takeoffs. Named E2 Enhanced Take Off System, this new system is designed to reduce pilot workload, improve safety, and enhance range and takeoff weight. Embraer has begun testing the system, aiming for approval by aviation authorities by 2025, with plans to initially implement it at select airports like London City, Florence, and Santos Dumont. While the system promises improved operational efficiency, aviation experts remain cautious, noting that the true impact will depend on real-world performance, retrofitting, and training requirements.

🧬 Biotechnology

Living Carbon: Photosynthesis-Enhanced Trees
In this article, Elliot Hershberg from The Century of Biology profiles Living Carbon, a biotech company that combats climate change by engineering photosynthesis-enhanced trees that grow faster and sequester more carbon. According to the article, their engineered poplar trees can accumulate up to 53% more biomass in growth experiments and have planted hundreds of thousands of genetically engineered trees across the US, raising over $36 million in funding. The article also explains their business model, which leverages carbon credits, seedling sales, and "Projects as a Service" (PaaS) to reforest degraded lands and offset emissions.

"Golden Lettuce" genetically engineered to pack 30 times more vitamins
Spanish scientists have created a “Golden Lettuce,” a genetically engineered lettuce that contains 30 times more beta-carotene, an antioxidant used by the body to produce vitamin A, than regular lettuce. This yellow lettuce could help support healthy vision, immune function, and cell growth, and may offer protection against diseases like Alzheimer's and cancer.

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