👋🏼 Hey there! Welcome to another roundup edition of Offload, a newsletter for non-technical professionals to learn how to build products and automate work with AI and no-code tools.

In this issue, you'll find:

• How I automated a section of my newsletter with AI and no-code tools

• Veo 3 (Google image model) test and opinion

• What is a Vector database

• And, to turn everything into a cushion with GPT-4o

-Offload This-

Automate formatted content creation with Make, Airtable, and OpenAI

Ever feel like writing the same piece of content more times than it should?

Same here.

That’s why I built a small automation to handle one repetitive part of my newsletter: the “Tool of the Week” section.

Instead of writing from scratch each week, I just jot down rough notes in an Airtable table - what I liked, what I didn’t, and where I used the tool - and then AI does the rest.

With a simple no-code workflow on Make.com, I connected Airtable to OpenAI and automated the writing and formatting.

Now, with one click, my messy bullets are turned into a clean, engaging snippet ready for the newsletter. 🧙🏻‍♂️

Why it’s useful ✅ 

• No more copy-pasting or rewriting: just log your thoughts once.

• Saves 45 minutes per week for each content block.

Why should you take a 👀 

• It is super flexible: you can adapt the same setup to automate meeting summaries, LinkedIn drafts, change logs, and more.

I shared the full guide, step by step, with screenshots and prompt templates so you can replicate it or remix it for your own use case.

Here is the link to the post. 

-Tool of the Week-

🛠️ Tool of the Week: Veo 3 - Google

❗ What it does: generates short AI videos with synced spoken dialogue and sound effects. It’s a model, embedded in Gemini or accessible via Flow.

Why it’s worth checking out:

✅ Automatically syncs audio (voice + ambient sounds)
✅ Super easy to whip up 8-second clips with no cuts

A few drawbacks:

⚠️ Artifacts pop up (extra limbs, glitchy scenes)
⚠️ Costs a ton of credits per run. To make a complete scene, will need lots of credits and runs.

How I used it:
I tried creating a quick app ad, but the final cut was more glitchy than glossy.

My take:
Fun for small creative experiments, but not quite there for serious commercial videos. My best take is this one below:

-Prompt Template-

Transforming ambitious goals into actionable plans

Large projects can often feel overwhelming, leading to procrastination or inaction.

The prompt of the week leverages AI to deconstruct complex objectives into manageable tasks, each with defined milestones and deadlines.

By providing a clear roadmap, it facilitates progress tracking and enhances productivity.

I usually prefer asking to also create a Canvas from it, so I can add or delete steps, but here is the standard:

-Learn AI-

Vector Databases, what?

In the last roundup edition, you learned what Retrieval Augmented Generation (RAG) is and that it is an important process to make AI closer to our day-to-day.

But there's a critical piece behind the scenes in RAG that makes that retrieval work: vector databases.

What is that?

Whenever you upload documents, PDFs, or any type of content to feed into a RAG pipeline, that data doesn’t just sit there in plain text or binary.

First, it gets embedded.

In this context, embedded means that each chunk of text is converted into a high-dimensional vector, which is basically a list ("array” in technical terms) of numbers that captures the semantic meaning of that text.

This is a vector: [0.92, 0.22, 0.43 .... 0.11]

It is called high-dimensional because there are more than three numbers in the vector. In practice, a vector may contain hundreds or thousands of numbers.

What is interesting is that not only can text chunks be converted into vectors, but audio and images can also be "translated" into math so the computer understands and can compare meanings.

How can you convert a chunk of text holding a meaning into a vector?

Well, this is what an embedding model does. There are a couple of models out there that can do the job for each type of media - we don't have to explore them now.

Great, now you know that a chunk of text (or something that has meaning) can be embedded, i.e. converted into a vector, by an embedding model. 

Can you guess where this vector is stored?

You guessed right! A vector database.

These are a special kind of database designed to store and search high-dimensional vectors efficiently. Instead of looking for exact matches like traditional databases, it finds similar items based on meaning, not just words.

Great, now you know what a vector database is. Knowledge acquired!

Here is a visual summary:

What would you like to learn next? Let me know!

-Tools I'm Using- 

1. make.com*: a no-code automation platform that connects different apps and services using visual workflows. Useful when you want to automate complex tasks across tools like Airtable, Gmail, OpenAI, and more. I used this week for the teardown of my content automation workflow; more user-friendly than n8n.
   Using score: ⭐️⭐️
   

2. Airtable*: a spreadsheet-database hybrid that feels familiar like Excel but works like a lightweight backend. Great for organizing data, managing workflows, or even powering full apps when paired with automation tools. Used in the same project as mentioned above and many others! 
   Using score: ⭐️⭐️⭐️
   

3. Bubble.io: a no-code platform for building fully functional web apps using a visual editor. Useful when you want to launch SaaS products, marketplaces, or internal tools without writing code, complete with user authentication, databases, and responsive design. It is still my go to app builder!
   Using score: ⭐️⭐️⭐️⭐️⭐️

_{* Indicates referral links. You help me if you use them.}

-Fun with AI-

Turn everything into a cushion with GPT-4o

Here is the step-by-step:

1. Go to chatGPT, make sure GPT-4o model is selected

2. Type this prompt below

You'll get something like this:

You can do this for anything, substitute the input under brackets or input an image.

For example:

• Create a high-resolution 3D render of [the icon attached] designed as an inflatable, puffy object …

• Create a high-resolution 3D render of [the selfie attached] designed as an inflatable, puffy object …

I wonder if there is a company that can “print” those in real life.

-Any feedback-

Before you go I’d love to know what you thought of today's newsletter to help me improve for you.