New AI Features in Windows, Canva, Power Automate, and Google Ads + MindStudio Updates

You're receiving this email because you registered to one of our workshops. You can unsubscribe at the bottom of each email at any time.

MindStudio redesigned some of its core menus and introduced dark mode. In the meantime, our developers are working on audio models to bring you AI assistants that can speak.

Microsoft, Google, and Canva all held announcements this week. Microsoft released Copilot+ PCs, a new Windows experience with on-device AI and a very controversial “Recall” feature, Google is facing backlash for AI overviews suggesting to eat rocks and glue, and Canva refreshed their UI while deciding to focus on B2B more.

As a reminder, we’re now welcoming partners that want to build AIs for their clients. Sign up for extra support, training resources, and more here.

🗞️ Industry news

1. Microsoft announced Copilot+ PCs, their new vision for on-device AI features
   

   

   Copyright Microsoft

   
   On Monday, Microsoft unveiled Copilot+ PCs, a new category of AI-powered Windows PCs. Microsoft claims they can be up to 58% faster than a MacBook Air.
   
   These devices, featuring advanced silicon capable of 40+ TOPS, all-day battery life, and cutting-edge AI models, offer unique capabilities like Recall for easy content retrieval, Cocreator for real-time image generation, and Live Captions for language translation.
   
   Recall is a new and somewhat creepy feature that screenshots all your moves on-device, uses AI to analyze the images, and creates an index. This lets you relive all previous interactions on the machine. While this is impressive, it can also pose one of the most significant privacy risks in tech. The UK Privacy Watchdog is investigating, but Microsoft claims storage and AI processing happens on-device.
   
   Cocreator offers a Krea-like experience inside Microsoft Paint. Sketch what you’d like to draw, and AI will draw it for you. This is not a new concept, Google has been working on AutoDraw for years, but it’s nice to have it inside such an OG app.
   
   Live Captions are moving away from app-specific layouts, like Skype or Zoom, and moving into your OS. Captions have the potential to bring remote work and cross-language work to the next level worldwide.
   

2. Scarlett Johansson didn’t like GPT-4o’s voice and asked OpenAI to remove it. They did.
   
   Following OpenAI’s surprise spring event last week, Altman tweeted “Her,” referencing the popular movie where a man falls in love with an AI assistant.
   
   The AI assistant in the movie is voiced by Scarlett Johansson, who claimed the new “Sky” voice in OpenAI’s assistant is too similar to her real voice. The actress has been approached by OpenAI in the past but refused to give her voice to train an AI model.
   
   OpenAI swiftly removed the voice and changed the default to a much more neutral-sounding one.
   
   The real question remains: Did OpenAI intend to make it sound like Scarlett? Can other people claim the voice “sounds like” them and ask for it to be removed? This is all TBD.
   

3. Google’s AI overviews are telling people to glue pizza and have a healthy breakfast with rocks
   
   

   

   The rocks meme also became a featured snippet now

   
   
   Look, this is not a surprise for anyone who has been working in AI for the past two years, but the new AI Overviews launched by Google last week are offering some weird results.
   
   Reportedly, AI Overviews aren’t currently able to distinguish between a meme publication like The Onion and a real news source. They also can’t seem to properly understand if a Reddit comment is supposed to be funny or a real piece of information.
   
   It’s important to note these are edge cases that drive lots of clicks to media publications and X profiles, and aren’t necessarily a representation of a large-scale problem. I don’t struggle to believe this is happening on a smaller scale (hallucinations, irrelevant results) on a significant percentage of edge search queries, though.
   
   Simply put, AI isn’t good enough to understand what’s real and what’s not. Even though Google has enough information to decipher it, they can’t realistically deploy the most capable AI models for every single search on the platform. Otherwise, even they wouldn't be able to keep up with the insane cost. Some responses are probably cached as well.

This week also saw Canva redesign its entire interface and switch to a B2B approach. Their Enterprise Plan is now generally available.

Microsoft announced many more updates to their core business products, such as Microsoft Power Apps, Microsoft Copilot, and others. Similar to Google, the focus was on AI everywhere, specifically AI that runs invisibly in the background while you work.

Google announced a few underwhelming updates to their Ads product at Google Marketing Live 2024, with most of the focus on the “Power Pair” PPC campaign type and a few ways to enhance images and videos in display ads.

🔥 Product Updates

This week, MindStudio released dark mode for all users. You can access it from the profile dropdown in the workspace page, or in the resources dropdown in the editor. This is what it looks like:

We have introduced two new formats to our webinar offerings: Hybrid and Automated Webinars.

Hybrid webinars are typically workshops. In these sessions, you'll watch a recorded video while our team members are present in the room to assist you. This setup combines the advantages of recorded videos, such as streamlined demos and fewer technical issues, with the live support you can only get in person. Our users love this format, and we're confident you will too.

Our team will respond to your questions in real time if they’re online or as soon as possible if they’re not.

💡 Tip of The Week

You can use send message blocks to extract portions of a larger prompt and output the response.

For example, I have a workflow where a custom function scrapes a lot of data. We’re talking 50k+ tokens or more.

To save cost, I’m using Haiku to go through the scrape only one time and performing all the operations I need:

• Summarize key data

• Outline the structure of the content

• and more

Then, a series of send message blocks cuts and slices the refined output, now in a separate variable. Here’s how it looks in the editor:

By doing so, I’m saving tons of tokens. Instead of resending the 50k output every time, I’m only processing the refined output (more like 1-2k tokens) and getting an even smaller output every subsequent interaction. They’re then all saved in variables I can re-use in my final prompts to the LLM.

If you want to learn more about techniques to save up on token cost, take a look at this masterclass video on our YouTube channel.

🤝 Community Events

If you want to hangout with our team, we usually host a Discord event every Friday @ 3PM Eastern. Join our Discord channel to keep up to date with the hangouts - our entire team is active there.

I also worked on a set of new workshops we will publish next week. We will start with a new batch from next week and go from there.

There’s a kicker for these new workshops. We took the learnings from the previous ones to deliver a better, more streamlined experience optimized to make you learn more in less time. I’m confident you will enjoy the new format.

The first two editions will focus on data sources and one live build. We uploaded the recording of most of the previous workshops on YouTube - make sure to check them here out if you haven’t already.

Thank you for being an invaluable member of our community, it’s always great to see many of you join multiple workshops 🔥

🌯 That’s a wrap!

Stay tuned to learn more about what’s next and get tips & tricks for your MindStudio build.

You saw it here first,

Giorgio Barilla
MindStudio Developer & Project Manager @ MindStudio