Introducing Profiler: Test Multiple Apps in MindStudio + AI News

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

MindStudio released Profiler, your own personal workbench to test different models against each other. It’s time to say goodbye to generic “industry benchmarks” and test the models for your use case, in-app, whenever you want.

This week, Google reportedly started to consider offering AI search for a fee, DALL-E added editing features, Apple claims its on-device AI system ReaLM outperforms GPT-4, and Claude announced Tool use (function calling), rivaling GPT’s function calling API.

What a week! Keep reading to learn more about the latest in tech and to find out more about our new pricing & packaging.

From our partners: Zilliz is interviewing our CEO on April 11th. If you’re interested in how RAG actually works and want to learn more about our provider, you can join for free here!

🗞️ Industry news

1. Apple claims its on-device model outperforms GPT-4
   Apple's AI researchers have developed a new conversational AI system, Reference Resolution As Language Modeling (ReaLM), that promises to significantly enhance Siri's capabilities.
   
   ReaLM, which focuses on improving reference resolution—Siri's understanding of context, onscreen content, and background activities—has shown to "substantially outperform" GPT-4 according to Apple.
   
   This advancement could lead to Siri remembering conversation history, understanding onscreen data, and recognizing background sounds like music. The system was tested against OpenAI's GPT-3.5 and GPT-4, with even Apple's smallest model rivaling GPT-4's performance.
   

2. Google considering charging for AI searches

   Google is contemplating introducing a subscription model for its AI-enhanced search features due to the high costs of providing AI services.
   
   This move would mark a significant change in Google's revenue strategy, aligning with industry trends where companies adopt subscription models to offset the expenses of AI.
   
   The consideration comes as AI search requires more computational power than traditional search methods, aiming to recoup the substantial costs involved.
   
   If Google were to move forward with the subscription model, it would mark one of the first time in its history they fail to provide a service that’s assisted by ads. Monetizing LLMs with ads is a challenge for other colossuses like Microsoft too.
   

3. You can now edit images in DALL-E
   The updates include new image editing tools within the ChatGPT interface on both web and mobile platforms, allowing users to easily modify their AI-generated images.
   
   Additionally, DALL-E now offers preset style suggestions like woodcut, gothic, synthwave, and hand-drawn to inspire users and assist in starting their creative projects.

4. Claude to rival GPT’s function calling with “Tool Use”
   Claude now features a public beta for its tool use functionality, enabling interaction with external client-side tools through API requests. This update allows users to provide Claude with custom tools for a broader range of tasks.
   
   During the beta phase, streaming support is not available, and while the tool use feature is considered production-ready, further beta iterations are expected before the final release. Currently, this feature is not supported on third-party platforms like Vertex AI or AWS Bedrock, but integration is planned.

🔥 Product Updates

This week, MindStudio released a few quality of life improvements and two major updates:

• Profiler: benchmarks are helpful for comparing the performance of various models, but they offer limited insight into which model best suits your specific needs. The new profiler addresses this by allowing you to compare different models directly. It sends the same prompt to a selection of models, enabling you to review and compare their performance side by side. Watch a quick demo here.

• Pagination: previous to the update, MindStudio used to load all threads within an app when called. This could result in slow loading times or the infamous “forever loading wheel”. Now, all apps will employ pagination. You should feel a noticeable improvement in loading speed for apps with many threads!

Coming soon:

• Implementation partners: we’re compiling a list of certified freelancers and agencies to help you get more done in less time. If you don’t have the time to learn how to use MindStudio or prefer to delegate the build, we suggest starting from here. Join our Discord to stay up to date and get the list as soon as it’s available.

• Desktop app: MindStudio is working hard to deploy the application wrapped as a desktop app. This way, you’ll be able to offer it to your employees when they boot up the work laptop for the first time and use it on the side while working;

• Budgeting tools: more budgeting tools, such as limits by app and by users, are on the roadmap. We’re committed to ensuring all users can always estimate and limit the cost of each MindStudio app.

💡 Tip of The Week

The “resolved” message in the debugger is a treasure trove of information to help you debug, improve your prompt, and save tokens.

In the debugger, each “programmatic” and “background” message includes two values:

• Sending Message: this is your prompt as is, with the variables in curly brackets and the instructions;

• Resolved Message: this is your actual prompt, aka what’s being sent to the AI. This matches the sending message but executes a function to retrieve the values in the variables, so it will expose what’s inside them.

This is useful for many reasons.

First of all, it lets you troubleshoot data sources. If you’re referencing a variable containing snippets from a query, they should appear in the resolved message. If they don’t, the query data source failed. If they do, but they’re wrong, you might want to change your prompt.

Secondly, it lets you save prompt tokens. Now that you know what an example final message looks like, you can copy paste the resolved message and test it with the model without running the whole workflow again. If you want to learn more on cost saving techniques, take a look at our 20m masterclass here.

The debugger is an advanced feature, but you shouldn’t be intimidated by it! It’s actually quite easy to grasp. Here’s a full tutorial.

🤝 Community Events

We’re working on a new set of workshops to offer in the next 2 weeks. We’ve heard your feedback loud and clear, and we’ll focus on:

• Live builds: we know you love to see MindStudio in action, and we’ll host two new live builds where Giorgio will walk through a real application and build it with you;

• Data sources: RAG is essential to building AIs infused with knowledge. Our first workshop was a success, and we’ll run another one to focus on more detailed examples;

• Q&A: how do you go from beginner to maestro? With tips and tricks coming from months of experience and well over 100 AIs built on the platform. Steal Giorgio’s brain for an hour and ask anything that comes to mind.

Also, we’re running a hackathon in our Discord Group (join). It’s a great way to test your skills, connect with the community, and get your projects going.

Date: Friday April 5th - Sunday April 7th (midnight PST)

Theme: Workplace Optimization - We know you all have day jobs, so what's a tedious task you want to solve with AI that can help you and your coworkers...?

1st place - $500 in credits

2nd place - $300 in credits

3rd place - $100 in credits

Once your app is ready, go ahead and submit it using this form.

The MindStudio team will be the only ones testing it, and our accounts do not incur usage costs. You won’t be charged for our tests. Please ensure your apps are NOT publicly available if you talk about it on Discord.

🌯 That’s a wrap!

AI is moving faster than usual, and we’re happy to be part of this groundbreaking revolution. 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