Auto-Generate MindStudio Workers in 1 click + $200/m ChatGPT

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

A couple of months ago, I demoed the “MindStudio workflow generator”, an app able to take in your use case and generate the whole worker for you. Now, the workflow generator is part of your workspace, on all plans, and it’s capable of building anything you want. More on this below — or register to the final live webinar of the year here to learn more about all the Q4 updates. We shipped TONS of new features that will help you in your next builds.

In other news, OpenAI started their “12d of OpenAI”, with a new demo or update daily. To rival the announcements, Google also shipped major features.

Notably, Gemini 2 was released in AI Studio, including a live screen sharing feature and full multimodality to talk with the AI like a colleague. The next day, OpenAI also announced live screen sharing in ChatGPT, months after the first demo. At the same time, they also released Sora, o1, o1 pro, the Siri integration, and a new $200/m version of ChatGPT, a controversial launch that sets them in a more B2B space.

Continue reading to learn more!

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

🗞️ Industry news

OpenAI days: what we know so far

Last week, OpenAI started the most impactful 12d in the AI space in a long time. Every day, at 10AM Pacific, the company released a new set of features, an entire new model, or a small demo of a cool update.

As a recap, this is what has been announced so far:

• o1, o1 Pro, and the $200/m controversial plan: on the first day, OpenAI announced an expensive monthly subscription for ChatGPT with no limits, access to a new “pro” mode for the o1 model, and priority use of new models like Sora. For now, it’s unclear if many people feel like the upgrade is worth paying 10x more the standard Plus subscription — but it clearly denotes a focus on interfaces versus dev APIs from OpenAI

• Reinforcement fine-tuning for all: this is the same method they use to fine tune GPT models, now available to select researchers

• Sora: on day 3, OpenAI released Sora publicly and had to close sign ups minutes later. The insane interest in the video model made it impossible to test so far. Sora is available on a separate domain and is linked to your ChatGPT account. Model usage is included in both Plus and Pro, but Plus accounts won’t be able to generate many videos and might be limited to 480p

• Canvas: ChatGPT Canvas got a huge upgrade and can now execute code. If you toggle on Canvas Mode while chatting, the chat will turn into a document and give you the option to “run” the code it generates. For now, it seems to only render python code

• ChatGPT in Siri: Siri can now interact with ChatGPT, and Apple users can sync their accounts from the settings panel. It’s possible to use ChatGPT anonymously as well

• Advanced voice with video & Santa mode: you can now screen share with ChatGPT and get help live. The team demoed this when they announced voice mode, but it’s never been possible until today. Rollout started on Thursday but might take over a week to conclude.

And this is just the first half. Many predict we’ll also get GPT-4.5, DALL-E 4, and some other unexpected surprises from OpenAI.

Google releases Gemini 2 and screen sharing

Google finally shipped something truly interesting. While Gemini has always been an interesting set of models, they failed to attract significant attention from developers and casual users due to very confusing interfaces and brand recognition.

Gemini 2 Flash is here to change that.

In a bold move, the team released the model for free in AI Studio, with a 2m tokens context window. Gemini 2 is fully multimodal, able to generate images and audio. It’s also one of the only models that can properly understand video.

For now, the company hasn’t released it via API. If the pricing stays the same as Gemini 1.5 Flash, and the performance is as good as they claim (basically matching GPT-4o and even surpassing it on some evals), this might very well be the most cost-effective smart model out there.

In addition to the classic chat mode, Google also released Voice Mode in AI Studio, together with the option to share your webcam or screen video. Gemini 2 Flash understands you in real time and replies in a natural voice tone.

First impressions:

• I’m shocked by how good the live mode is. I shared my screen and chatted with Gemini 2 while working on a few automation workflows, and the model was perfectly capable of understanding what was going on in real time

• The voice is robotic - not as good as Advanced Mode in ChatGPT. Still, very good and no real issues using it

• Once they integrate computer use (similar to Claude’s release), this will be one of the best assistive technologies for people with visual or physical impairments ever released

• All of the new features are available in the EU as well

• Gemini 2 Flash enables new agentic flows, and focuses on them with integrations such as Live Search on Google and tool use.

Truly awesome and packed release by Google — kudos to the team.

Anthropic reveals what people actually use Claude for

In an attempt to strengthen security and compliance, Anthropic announced “Clio”, a research project to investigate what people use Claude models for by scanning the interactions with the API and bot.

All interactions are anonymized so no personal data is in the chats it trains on.

According to the research:

• Web & Mobile App Development is on top. Pretty obvious reason here: Claude 3.5 Sonnet is the best model at coding according to many devs that use in IDEs like Cursor or no-coders taking advantage of tools like Bolt.new 

• Content creation is a close second. I started my MindStudio journey building an article writer and this is still one of the most common use cases we see in MindStudio as well

• Academia & Education is huge, making up over 14% combined. Students adapt faster to new technologies

You can see the full chart above. Very interesting indeed!

🔥 Product Updates

This week, MindStudio removed the final barrier to entry: time.

Anyone, on any plan, can now use our Automated Workflow Builder to automatically generate all inputs, prompts, custom functions, API setup, and JSON output structures. All it takes is one click.

The Workflow Builder is accessible by clicking on the bottom left corner within the “System Prompt” tab of a new app:

This is probably one of the most impactful features we ever released, and we’re confident you’ll like it. For the time being, the workflow builder doesn’t consume AI compute — feel free to test it out!

Here’s a more complete overview of the tool. In short:

• It will generate a plan to review before creating the actual workflow. You can accept or reject

• The scaffold will include eventual external dependencies (e.g. you need to send some values via API for it to work), a detailed plan on what it intends to do, and a diagram showing the blocks it will generate

• Once accepted, the AI gets to work and builds out the workflow, custom functions, inputs, and more in about 2 minutes

• You can now review the flow. Most of the time, this will include a few custom functions, in JS or Python, fully ready to go and documented in the readme file

Together with the workflow builder, we released:

• Prompt enhancement in ALL prompts, including Generate Text blocks. Just click on “enhance prompt” and the AI gets to work

• Python functions, including library support. This is awesome for data analysis

• The Generate Chart block to create Quickchart graphs

• The Extract Text from File block to use with an API call

• More output types. You can now have any Generate Text Block output in plain text, JSON, or CSV

• New models from Amazon, the Nova Family. Very cost-effective and near SOTA levels

• Slack Integration. You can now send the result of a flow to a Slack Channel or use the “Post to Slack” block to send whatever you want to your colleagues

• Share access. Like a Google Doc, you can now share your AIs to anyone, only to your colleagues, or keep them private. With 50 free seats on the community plan, this benefits anyone using AIs in a team.

• PDF & Image Generation: we now have a block capable of generating and rendering HTML to create PDFs and image templates. Learn more in the demo here

AND we’re building so much more.

For a recap of all current features, new releases, and plans for the future, register for our last webinar of the year. Our Director of Product, Luis, will go over all of it and answer Q&A from the participants.

💡 Tip of The Week

OpenAI and Anthropic are showing how much products like MindStudio are needed in the market for builders.

The biggest model provider in the world, OpenAI, makes more money from the playground they built to test the API than the API itself.

There’s never been a better time to build AI solutions and interfaces for your clients, and MindStudio can help you do it at lightning speed. The market is now:

• Old enough. We’re 2y in, and people are aware of what AI is, what it can do, and the potential to change how they work

• Less scared. We got that AI is very powerful, sometimes scary, but most agree it’s not going to end the world. There are less articles talking about how it’s going to end us all

• Money oriented. Once we got past the VC stage, now companies are trying to find ways to make money. If one of the largest players chose interfaces, that means our focus of giving you the tools to build products around your market is the right choice to increase your revenue.

Look, building another ChatGPT doesn’t make much sense. We’re going to continue building a better chat endpoint, but our end goal is enabling you to build custom solutions for your clients. Things people are happy to pay for.

We want to empower you to partner up with MindStudio and use the tools to build niche-dominating workflows and apps.

If you have any suggestions on how we can do this better, please reply to this email! I read all replies 🙂 

🌯 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