- The Singularity
- Posts
- Microsoft seeks to make the entire company an AI platform
Microsoft seeks to make the entire company an AI platform
Plus: OpenAI has regulation on their mind
![](https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/6eddeb49-d7b5-4ba0-878f-2b8167e92188/MSBUILD.png?t=1685060248)
![](https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/61915e3d-b0b3-45fb-942c-3669afb12542/singularity_thumbnail.png)
The latest AI news for you and your business.
Greetings friends! On Wednesday we were treated to Microsoft’s Build developer conference, just a couple of weeks after Google I/O, and well, there are a lot of similarities. I mean, you can only added AI based chat to productivity apps so many ways, right? There were some good surprises for developers and consumers that will help Microsoft to distinguish itself in the AI race. Let’s dig in.
In this issue
⚒️ Microsoft Build Conference, AI all the things
📜 OpenAI's Democratic Quest for AI Regulation
🌍️ OpenAI's Global Vision: An International Regulatory Body for AI
🏢 Infosys Unveils Topaz AI: A Game-Changer in Generative AI Solutions
Microsoft Build Conference, AI Copilot all the things*
Microsoft didn’t hold back their annual Build developer conference this week, making numerous new product announcements and of course a lot of new toys for developers. Nearly all of which either uses AI or allows developers to build software that has AI integrated. During the keynote, CTO Kevin Scott, implied that Microsoft seeks to make the entire company an AI platform. As Microsoft adds Copilot to nearly the entire product line, it also is encouraging developers to be active participants by augmenting the AI through creation of their own Copilot plugins. It’s clear that Microsoft is all-in on AI and it wants their vast constituency to join them on this mission.
![](https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/8b6c3918-4bd6-47cd-8d5e-b10238ba9a74/image.png)
Build, like Google I/O, is a developer conference and so much of the news centers around developers. However, it would be difficult to announce how developers could use a new feature without announcing the feature for consumers as well. Here are some of the highlights from Build this year:
Azure AI Studio lets developers build their own AI ‘copilots’.
Plug-ins for AI enabled Apps.
Power Pages Copilot to help make designing web pages easier.
Fabric. A new data analytics platform designed for AI
ChatGPT (Bing) comes to Windows
New AI hub in the Windows Store
Semantic Kernel open source framework for easily adding AI flows to apps.
[read more]
*disclaimer: I am a full-time employee of Microsoft Corporation. The views presented here are my own and in no way reflect those of my employer. They are based entirely on publicly available information and I have given every effort to be unbiased in my analysis.
OpenAI's Democratic Quest for AI Regulation
OpenAI has announced a new grant program designed to foster a democratic process for AI regulation, aiming to fill a gap left by the rapid pace of AI innovation. The program will fund projects developing proof-of-concepts for this democratic process, providing valuable insights for a larger and more ambitious global initiative. Through this program, OpenAI hopes to create a democratic system that encourages diversity and transparency in decision-making about AI behavior. This innovative approach places a strong emphasis on ensuring that AI regulations reflect the diverse perspectives of the public interest. Applications for the grant are open until June 24, with selected recipients required to engage at least 500 participants, publish a report, and open source their work.
[read more]
OpenAI's Global Vision: An International Regulatory Body for AI
OpenAI's leaders propose the establishment of an international regulatory body for AI, considering the technology's rapid evolution and potential risks. They draw parallels with the International Atomic Energy Agency, suggesting that AI systems above a certain capability threshold should be subject to international authority for safety compliance and deployment restrictions. The organization also recommends tracking compute power and energy usage in AI research as objective measures. The proposal, while broadly stated, is seen as a conversation starter in the industry and a call for public oversight, even though the specifics of such a mechanism are yet to be designed.
[read more]
Infosys Unveils Topaz AI: A Game-Changer in Generative AI Solutions
Infosys has introduced a new solution called Topaz AI, designed to accelerate business value from generative AI. The offering is aimed at helping businesses maximize value from AI, merging Infosys’ deep learning solutions and generative AI to create new revenue streams, improve operational efficiency, and yield more profitable outcomes. It combines various AI frameworks, pre-trained models, and data pipelines to enable AI at scale. Infosys claims that Topaz AI will be a key component for businesses to realize AI’s true business potential.
[read more]
![](https://media.beehiiv.com/cdn-cgi/image/fit=scale-down,format=auto,onerror=redirect,quality=80/uploads/asset/file/718b99cd-4e4f-4050-8162-d97205c4d4c0/AIExplained.png)
Embeddings & Vectors
In all of the developer focused news lately, you may have heard the terms ‘Vector or Vector Store’ and ‘Embedding’, but what are these things? They play a vital role in making AI useful… and performant.
Let’s start with Vector. A vector is a way to represent information in a computer-friendly format. It's like a list of numbers that tells a computer where something is located. Vectors can represent things like directions, positions, or properties of an object. For example, in a map, a vector can show the direction and distance from one place to another. In AI, vectors are used to represent things like words or images, making it easier for computers to understand and work with them. Vectors are super useful because they let computers compare and analyze different pieces of information to find patterns or similarities.
Now an Embedding, is a special type of vector. It is a way to represent information, like words or images, using numbers that computers can understand. It's like translating data into a special numerical language. Embeddings help computers understand the meaning and connections between different pieces of information. In AI, embeddings are used in tasks like understanding text and recognizing images. For example, in text, embeddings help computers understand the meaning of words and how they relate to each other. In images, embeddings help computers recognize objects and patterns. By using embeddings, AI can make better predictions and do things like understanding emotions in text, recognizing objects in images, and suggesting similar things based on what you like.
When developers create vectors or embeddings, they need to save them so that they can use them later to make the applications they build more useful. A Vector Store is a special type of storage (think of it as a database) that is made specifically for Vectors/Embeddings. There are a few new startups that are specializing in hosted Vector Stores (just like companies host websites), such as Pinecone.io.
So there you have it. You now know all about Vectors, Embeddings and Vector Stores. You can impress your friends or annoy your spouse with your newfound knowledge. 😁
Tools to check out!
You’ve reached the end for today, but fear not! We’ll back again with more news and tips to inspire you.
If you enjoyed this newsletter, please consider sharing it. 😄