Apple's iPhone 16 Is All About AI
Plus: Weave is launching new AI robot butler, Google's AlphaProteo AI designs new protein binders, OpenAI hints at AI 100x smarter than GPT-4, and more.
Hello Engineering Leaders and AI Enthusiasts!
This newsletter brings you the latest AI updates in just 4 minutes! Dive in for a quick summary of everything important that happened in AI over the last week.
And a huge shoutout to our amazing readers. We appreciate you😊
In today’s edition:
🆕 Apple launches AI-powered iPhone 16
🤖 Weave's AI robot butler does all your household chores
🔬 Google's AlphaProteo AI designs new protein binders
💰 OpenAI's former co-founder raised $1B for a new AI safety startup
🧠 OpenAI hints at AI 100x smarter than GPT-4
🚀xAI reveals world's most powerful AI supercomputer
📚 Knowledge Nugget: Is the impact of generative AI on elections overblown? by
Let’s go!
Apple launches AI-powered iPhone 16
Apple just launched the iPhone 16, and it's all about AI. The new A18 chip powers Apple Intelligence, a suite of AI features:
Smarter writing tools that can rewrite emails and create custom emojis,
Photo searches that understand natural language, and
Siri finally gets what you're saying (and can actually do something about it).
But it's not just about the iPhone. Apple is bringing AI to the whole ecosystem. The Apple Watch Series 10 is 10% lighter and adds AI features like smart photo curation for watch faces. AirPods 4 now understands head gestures, letting you nod or shake to respond to Siri. While iOS 18 launches September 16, the full suite of Apple Intelligence features will roll out gradually through 2024, starting with a U.S. beta this fall.
Why does it matter?
By developing its own AI models and integrating them deeply into its ecosystem, Apple is not just competing with Google and OpenAI – it's changing what we expect from our devices.
Weave's AI robot butler does all your household chores
Weave Robotics is launching Isaac, a personal robot butler, in the fall of 2025. This AI helper can autonomously organize, care for pets, water plants, find objects, and more. Weave offers remote operation services and privacy features for tasks beyond its current capabilities. It will improve over time through over-the-air updates and can learn household layouts.
Isaac uses AI technologies like VLA, VLM, and ML pipeline. Reservations for this robot are open for a $1000 down payment, with a total price tag of $59,000 (or 48 monthly installments of $1385 if you prefer). However, the robot will be delivered to the first 30 customers in fall 2025, initially in the US only.
Why does it matter?
Isaac showcases how AI can free up hours spent on chores. But the steep price tag raises the question of whether only the wealthy can afford AI-powered convenience. Isaac's success (or failure) could change how we think about AI's role in our homes.
Google's AlphaProteo AI designs new protein binders
Google DeepMind's new AI, AlphaProteo, can design novel proteins that bind to specific targets up to 300 times more effectively than current methods. In lab tests, AlphaProteo's creations successfully designed binders for 7 out of 8 target proteins, including some linked to cancer and viral infections.
It could accelerate drug discovery processes, improve disease diagnostic tools, and contribute to developing pest-resistant crops. While AlphaProteo could not design a binder for one challenging protein (TNFα), Google's research team is actively expanding the system's capabilities to address more complex targets.
Why does it matter?
This shows how AI can accelerate the development of new treatments for diseases like cancer and create more effective diagnostic tools. It also saves years of lab work and billions in research costs, potentially bringing life-saving drugs to market faster.
OpenAI's former co-founder raised $1B for a new AI safety startup
Ilya Sutskever, OpenAI's former chief scientist, has launched a new venture with a bang. His startup, Safe Superintelligence (SSI), raised $1 billion from top-tier investors to tackle one of AI's biggest challenges: building superintelligent systems that won't accidentally erase humanity.
With a $5 billion valuation and only 10 employees (for now), SSI focuses on R&D for a few years before even thinking about a product. They're hiring for "good character" over credentials and taking a fresh approach to AI scaling.
Why does it matter?
Sutskever's move comes after his dramatic exit from OpenAI. This massive investment signals that AI safety concerns are becoming more prevalent. Plus, his focus exclusively on safety can influence how the entire industry approaches superintelligent AI development.
OpenAI hints at AI 100x smarter than GPT-4
OpenAI Japan CEO Tadao Nagasaki said ChatGPT is officially in the 200 million user club. Speaking at a KDDI event, Nagasaki announced this milestone alongside some hints about the future.
Nagasaki mentioned "GPT Next," a future AI model he claims could be 100 times more powerful than GPT-4, thanks to improvements in architecture and learning efficiency. GPT-4 NEXT is expected to use a miniature version of "Strawberry" with computational resources similar to GPT-4. They also mention the release of a new AI model next year: "Orion," which has been trained with 10 times more computational resources than GPT-4.
Why does it matter?
The focus on improved architecture and learning efficiency, rather than just increased computing power, suggests a shift in AI development strategies. This could inspire new, cost-efficient approaches to AI design and training across the industry.
Source (The article is in Japanese)
xAI reveals world's most powerful AI supercomputer
Elon Musk's AI startup, xAI, activated "Colossus," a massive AI training system boasting 100,000 Nvidia H100 GPUs. Built in just 122 days, Musk claims it's the world's most powerful AI training rig. Moreover, it will double power within months, adding 50,000 H200 chips.
It's preparing to train GROK-3, xAI's next-gen language model, aiming to dethrone ChatGPT-4 by December. The price? $2 billion minimum, just for the chips. It consumes enough electricity to power a small city and uses up to a million gallons of water daily for cooling.
Why does it matter?
As AI systems become more powerful, questions about environmental impact, resource concentration, and societal implications will intensify. Colossus's energy and water consumption highlights the physical costs of AI advancement and sparks debates on sustainable AI development.
Enjoying the latest AI updates?
Refer your pals to subscribe to our newsletter and get exclusive access to 400+ game-changing AI tools.
When you use the referral link above or the “Share” button on any post, you'll get the credit for any new subscribers. All you need to do is send the link via text or email or share it on social media with friends.
Knowledge Nugget: Is the impact of generative AI on elections overblown?
In this post,
, a researcher at the Integrity Institute, explains that the impact of AI on elections might be overblown. Despite widespread concerns, Motyl points out that real-world events still pack more punch than AI-generated content in shaping public opinion. He suggests that the ease of creating fake content with AI isn't the main issue – it's the distribution mechanisms on social media platforms that need addressing.
Motyl emphasizes that the real challenge lies in the massive growth of harmful content and social media companies' reduced investment in safeguarding elections. He warns that smaller, non-English-speaking countries face greater risks due to inadequate content moderation.
Why does it matter?
Motyl's perspective reminds us to focus on existing problems in our information ecosystem rather than getting caught up in AI alarmism. It's a wake-up call for tech companies to reinvest in trust and safety measures, especially as the volume of misleading content grows.
What Else Is Happening❗
⚡️ In a new patent, Tesla reveals a wireless charging system capable of charging autonomous vehicles without human intervention.
🎮 Recently, researchers created AI agents that can play procedurally generated video game levels to evaluate human engagement, helping to speed up game design.
💻 Replit launched a Replit AI Agent, which creates software applications from simple text descriptions, making coding easier for everyone.
💰 According to reports, OpenAI considered subscriptions up to $2,000 per month for advanced AI models like Strawberry and Orion.
💬 Groq dropped LLaVA V1.5 7B, a powerful, new multimodal AI model that can understand text and images and engage in conversation. It runs 4x faster than OpenAI’s GPT-4o.
🧠 Researchers have discovered a technique called s RE2 (Re-Reading), which boosts AI problem-solving skills by having models repeat questions before answering.
🧱 Altera's Project Sid has simulated a complex AI society within Minecraft, featuring over 1,000 autonomous agents that develop their own economic, cultural, and political systems.
💼 The U.S. Department of Justice has intensified its antitrust probe into Nvidia's AI chip business by issuing a subpoena, signaling increased scrutiny of its market practices.
🤖 OpenAI is reportedly exploring in-house AI chip production using TSMC's advanced 1.6nm process node.
🧬 Scientists atProfluent Bio have engineered a novel AI-designed gene-editing tool called OpenCRISPR-1. It matches or exceeds the capabilities of existing CRISPR systems while improving safety and accuracy.
New to the newsletter?
The AI Edge keeps engineering leaders & AI enthusiasts like you on the cutting edge of AI. From machine learning to ChatGPT to generative AI and large language models, we break down the latest AI developments and how you can apply them in your work.
Thanks for reading, and see you next week! 😊