Microsoft launches a deepfake creator (2024)

Read time: 2.5 minutes

Welcome back, Superhuman

Access to chips has been at the heart of the battle between AI startups and tech giants over the past year. But Microsoft just landed a huge blow to its rivals by announcing its AI own chips, which could give Microsoft and OpenAI an edge in the coming months.

But that was only the tip of the iceberg from yesterday’s announcements. Microsoft also announced new apps, upgraded features, and a cool (creepy?) new deepfake avatar generator.

TODAY’S MENU

TODAY IN AI & TECH

INSIGHT

Microsoft’s new chips could provide an alternative to Nvidia and give OpenAI the edge

If ChatGPT is the Rolls Royce of AI apps, then Nvidia chips are the monstrous V12 engines powering the GPT-4 model under the hood. As the scramble for AI chips broke out this past year, Nvidia’s stock price set new records and the company recently joined the exclusive club of companies valued at over a Trillion dollars.

But short supply of these chips, along with eye-watering prices, have limited the ambitions of many AI companies. According to some reports, OpenAI CEO Sam Altman recently made the acquisition of these chips the top priority for the company.

But if you can’t find em, you gotta make your own. At least that’s what Microsoft seems to be doing. The company announced its first AI chips yesterday, called Cobalt and Maia. The chips will be deployed in Microsoft’s Azure data centers next year, and support services like OpenAI’s GPT models and Microsoft’s own Copilot.

As access to chips shapes up to be the deciding factor in who takes the lead in AI this coming decade, Microsoft’s headstart in specialized chip manufacturing could give it - and OpenAI - a decisive early edge over the rest of its rivals.

Read the official announcement here.

TOGETHER WITH 360LEARNING

Everyone's talking about AI in L&D, but nobody’s talking about how to use it.

Until now.

Introducing , a free course designed to help L&D pros level up AI skills and deliver outcomes that matter, faster.

Become an expert in:

NEXT IN AI

Microsoft unveils new deepfake avatar generator and other major AI updates at Ignite event

Out of all of the things Microsoft announced at its Ignite event yesterday, a new tool that lets you create photorealistic avatars and make them say things was the most unexpected.

Available as a preview today, the Azure AI Speech text to speech avatar feature will let users upload an image of the person they want their avatars to resemble and then feed these avatars a script to read out in a video. Avatars will also be able to speak multiple languages.

Microsoft said that the feature can be used to build conversational agents, virtual assistants, chatbots and more. You can see a live working avatar and understand how it works in the video above.

Other key announcements from Microsoft:

You can find the full list of new announcements on the official blog here.

BUZZWORD BUSTER

GPT

Generative Pre-trained Transformers (GPT) are a subset of AI models that use transformer architecture to understand and generate human language. GPTs are a key breakthrough in how AI models function and they are used in many leading AI models and applications like GPT-4 and ChatGPT.

TTOGETHER WITH ARAGON AI

Transform your selfies into AI headshots

Get stunning professional headshots without the hassle! Never pay for a photographer again.

Aragon AI’s tech is built by leading AI researchers and delivers the best results within90-minutes. Just upload some selfies to see for yourself.

RESEARCH

Researchers create 3D models from 2D images in 5 seconds

Researchers from Adobe Research and the Australian National University have created an AI model that can turn a 2D image into a 3D model in less than 5 seconds. This discovery has far-reaching consequences for gaming, animation, industrial design, AR, and VR.

The model can create high-fidelity 3D models from both real-world and AI-generated photos while preserving accurate geometry and sophisticated texturing. Use cases span from making 3D models easier to create in video games and animations to accelerating prototyping in industrial design, improving AR/VR experiences, and democratizing 3D modelling.

5 AI Tools to Supercharge Your Productivity

Notion AI: Access the power of AI right inside Notion to find answers, generate text and more.

Chat2Design: Turn your text into design - create, customize, and share UI components in a snap.

Speechmatics(sponsored): Get a global audience of 3 billion using AI + LLMs. With Speech Intelligence by Speechmatics, you can transcribe and translate over half the world’s population in real time. Try it for no cost without code.

Audyo: Generate human-like AI voices by typing in a doc.

Framer AI: Build professional sites from simple prompts while delivering best-in-class SEO, performance, and hosting.

AI TUTORIAL

How to create posters in ChatGPT

Prompt: Create a poster for an event called [insert event name]. The location is [insert location]. The date and time is [insert time].

We’ve noticed that ChatGPT can struggle a bit with getting text right. It usually takes a few tries to get the right image with the right text.

AI GENERATED IMAGES

Bart Simpsons’ effigy

bart simpson wax figure, 4k, ultra quality

📈Feature your product in the world’s biggest AI newsletter

Superhuman is the world’s biggest and fastest-growing AI newsletter with 450,000+ readers working at the world’s leading startups and enterprises. Companies like Masterworks, Brave, SEMRush, and 1Password feature their products in Superhuman. Main ads are sold out for the next 4 weeks. You can book future ad spotshere.

🧞 Your wish is my command

Reviews of the day

Thanks for reading.

Until next time!

p.s. if you want to sign up for this newsletter or share it with a friend or colleague, you can find us here