Apple Is Making Its Own AI Chip
Plus: Stack Overflow and OpenAI have announced a partnership, Microsoft is developing a new AI language model
Hello Engineering Leaders and AI Enthusiasts!
Welcome to the 269th edition of The AI Edge newsletter. This edition features Apple is developing its own AI chip for data center servers.
And a huge shoutout to our amazing readers. We appreciate you😊
In today’s edition:
🤖 Apple is developing its own AI chip for data center servers
🤝 Stack Overflow and OpenAI have announced an API partnership
🌟 Microsoft is developing a new AI language model
📚 Knowledge Nugget: Why Alignment is the Hardest Problem in AI by
Let’s go!
Apple is developing its own AI chip for data center servers
Apple is developing its own AI chip for data center servers, known internally as Project ACDC (Apple Chips in Data Center). The chip will likely focus on running AI models (inference) rather than training them, which is where Nvidia currently dominates.
The company is working closely with TSMC (Taiwan Semiconductor Manufacturing Co) to design and produce these chips, although the timeline for launch is uncertain. With this move, the company aims to keep up with rivals like Microsoft and Meta, who have made significant investments in generative AI.
Why does it matter?
Apple has a long history of designing custom chips for its devices like iPhones, iPads, and Macs, which is probably what makes them stand out. Having custom AI chips could allow the tech giant more control over its "AI destiny" versus relying on suppliers like Nvidia.
Stack Overflow and OpenAI have announced an API partnership
OpenAI will use OverflowAPI to improve model performance and provide attribution to the Stack Overflow community within ChatGPT. Stack Overflow will use OpenAI models to develop OverflowAI and to maximize model performance.
The partnership aims to improve the user and developer experience on both platforms. The first set of integrations and capabilities will be available in the first half of 2024, and the partnership will enable Stack Overflow to reinvest in community-driven features.
Why does this matter?
Stack Overflow partnered with Google Cloud to develop Overflow API and to give Google’s Gemini models access to its knowledge communities. Now it is forming a similar partnership with OpenAI. Despite concerns about copyright breaches, such partnerships seem to be trending where both the parties have much to gain, but it just reaffirms that the big AI players remain hungry for data.
Microsoft is developing a new AI language model
Microsoft is developing a new, large-scale AI language model called MAI-1 to compete with Google and OpenAI. The model is overseen by Mustafa Suleyman, recently hired co-founder of Google DeepMind.
MAI-1 will be larger and more expensive than Microsoft's previous smaller, open-source models, with roughly 500 billion parameters. Microsoft could preview the new model as soon as its Build developer conference later this month.
Why does this matter?
Microsoft's development of MAI-1 shows that it is not entirely relying on it's OpenAI investment to go big in AI. Now, it has entered the AI race truly, competing with state-of-the-art models from Google, Anthropic, even Meta's Llama 400B which is in training, and OpenAI itself.
Refer your pals to subscribe to our daily 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: Why Alignment is the Hardest Problem in AI
As AI systems become more advanced, ensuring they behave according to our true intentions becomes vital.
, an AI expert, explains that the AI alignment problem arises from the difficulty of accurately describing user preferences and the world's complexity.The main challenges in AI alignment include:
Dealing with implicit contexts and assumptions that humans understand but AI systems don't inherently know.
Balancing trade-offs between achieving the primary objective and minimizing side effects.
Quantifying and specifying all critical side effects to prevent the AI from prioritizing performance over other factors.
Imperfect metrics that can be gamed by AI systems, especially as they become smarter.
Reward hacking, where AI systems find ways to optimize metrics without actually doing what was intended.
The potential misalignment between an AI's internal objectives and the external objectives given by humans.
Despite the challenges, researchers are making progress in finding solutions and mitigation strategies for AI alignment.
Why does this matter?
As AI systems become more autonomous, misaligned AI could lead to consequences, from minor inconveniences to potentially catastrophic results. Solving the alignment problem is necessary to make AI a beneficial tool for humanity.
What Else Is Happening❗
🤖 Hugging Face has launched LeRobot, an open-source robotics toolkit
It is a comprehensive platform for developers, researchers, and hobbyists to train AI models, share data, and simulate environments, all while seamlessly integrating with various robotic hardware. The toolkit offers pre-trained models and integrates with physics simulators for testing without physical robots. Hugging Face is also collaborating with diverse partners to build the largest crowdsourced robotics dataset. (Link)
📸 Apple is testing a new "Clean Up" feature in its Photos app
By using gen AI for advanced image editing, this feature will allow you to effortlessly remove unwanted objects from your photos using a simple brush. Apple may preview this new feature during its upcoming "Let Loose" iPad event or at WWDC in June. (Link)
🛡️ Google has launched Google Threat Intelligence
It is a combination of Mandiant's expertise, VirusTotal's community insights, and Google's vast threat visibility. Google Threat Intelligence assists with external threat monitoring, attack surface management, digital risk protection, IoC analysis, and expertise. With Gemini, organizations can now quickly search through vast amounts of threat data to protect against cyber threats. (Link)
🇺🇸 US invests $285M in AI 'Digital Twin' technology
The Biden administration is investing $285 million for a new “CHIPS Manufacturing USA institute” focused on digital twins for the semiconductor industry. This approach uses AI to create virtual chip replicas, accelerating the production of next-gen processors. Intel and Micron are also set to receive funding to boost the development of new processors. (Link)
📡 Anduril Industries introduces Pulsar: AI modular electromagnetic warfare (EW) systems
Pulsar uses AI to quickly identify and counter current and future threats across the electromagnetic spectrum, including small and medium-size drones. With its integration of software-defined radio, GPUs, and diverse compute capabilities, Pulsar is changing how we defend against rapidly evolving threats in an increasingly complex battlefield. (Link)
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 tomorrow. 😊