👨‍🚀 The S.H.I.P. vibe coding framework

Hello AInauts,

Welcome to the latest issue of your favorite newsletter!

We've discovered an absolute game-changer tool for presentations that makes PowerPoint look like a tool from the Stone Age. We also dove back into Vibe Coding and introduce you to the S.H.I.P. framework and how to use OpenAIs latest image API.

Ready? Here's what we have in store for you today:

• 🖖 Use the S.H.I.P. vibe coding framework and this new image API

• ✨ Will this tool soon retire PowerPoint?

• 🎥 Runway Gen:48 - the ultimate test for your creativity

Here we go!

🖖 Use the S.H.I.P. vibe coding framework and this new image API

Do you sometimes have the thought: "This app should exist!" ... and then reality gets in the way: not enough technical know-how, no dev team, no six-figure budget, no time.

But there is a solution: vibe code an app with AI!

Vibe coding means that you can create apps with natural language in no time at all - even as a non-programmer. We have already reported on this in detail here.

Many solopreneurs are already using Vibe Coding to build a profitable SaaS business. Want some examples? Here you go ... that will definitely increase your appetite for more.

The advantages of vibe coding

The classic path from idea to app used to be very time-consuming: Market analysis, validation, MVP planning, wireframes, sprints, development teams. All expensive, slow, risky. We remember these projects (too) well...

Vibe Coding with AI turns that on its head. You start with a clear problem that is really relevant. You validate it and then use AI to quickly create an initial, usable app specification. You don't even have to formulate it perfectly. The AI absorbs the essence of your input and builds you a complete app!

And as a vibe coder, you are spoiled for choice! Bolt, Lovable, Replit, Firebase, Cursor, Windsurf, V0 etc. are all valid options with which you can build really cool apps for free or for little money.

We have experimented a bit with all of them, and mostly use Lovable or Replit. The tools are easy to use. Lovable has a Supabase database integration, Replit also takes care of the entire tech stack for you. In other words, you can publish the apps directly after creating them. Disadvantage: the free version is limited.

Pro tips: Use Cline.bot and the free API credits from OpenAI and Google

If you are looking for a free solution, here are a few tips:

• In tools like Cursor, Windsurf, VSCode you can install the ingenious Cline extension. This has the advantage that you can bring your own API keys and benefit from the generous "freebies" offered by the model providers.

• For example, OpenAI offers you millions of free tokens every day - as long as you letb them train the models on your data.

• All you need is an OpenAI Developer Account and you can activate it here. Google Gemini also offers you free access to the API.

Use the new GPT Image 1 model to vibecode an app!

Ok, are you ready to code your first app? We've got an idea for you...

You remember the release of the latest OpenAI image model in ChatGPT - it made the GPUs melt. But Sam has kept his word and released the model for the API a few days ago under the name GPT Image 1!

This means that any app can integrate this functionality and generate extremely high-quality and detailed images with text. And the new image model is the perfect starting point for building something cool!

To be able to use it, you simply need to have a developer account and verify your organization with an ID card.

Pro tip: Unfortunately, this did not work for us in the first attempt we were not allowed to try again. We then added another person as admin for the organization and had them do the verification - that worked!

By the way, you can now also create images directly on the developer platform - even up to 10 versions at once!

Vibe coding with the S.H.I.P. framework

Follow this process, and keep in mind: Shipping > Perfecting!

S - Specify

• Clearly state the problem you are addressing. Develop your idea or use an existing app that can serve as inspiration.

• To create the detailed specification, you can collect the notes, screenshots, API documentation, etc.

• Put it all into a language model and have a specification created - Software Requirements Specification (SRS) according to IEEE 830/ISO 29148.

H - Hack

• As soon as you have the detailed specification, create a proof of concept.

• Select the Vibe coding tool of your choice, put in the spec and get started - after a few minutes you will have a first draft.

• Important: No perfection - the main thing is that it works! Don't get lost in the details, but make sure you build a simple, functional solution.

I - Iterate

• Once you have an initial status, you can optimize specific features.

• Sometimes the apps get worse and no longer work...

• So test extensively, collect feedback and optimize continuously.

• Only update one feature at a time, not everything at once.

P - Publish

• And of course you should also make your app release-ready!

• Ideally, you should offer a paid feature right away.

• If someone is willing to pay you money, you have validated the market,.

• Build in authentication (Google, Apple ID) and payment connection (Stripe, PayPal) - and then go live with the app.

Our take: The future of vibe coding

We've been experimenting quite a lot with these tools over the last few days, and honestly: it's brilliant what's possible with them.

But not everything works great. We also got stuck when an app was already working at 80% functionality and we then had to spend hours tinkering with the remaining 20% ... Nevertheless, we would encourage you to experiment with this yourself.

This is just the beginning. The next big shift is coming soon: AI agents that continuously work on your code base. Instead of generating code selectively, you will soon be able to say: "Build me feature X" and have it delivered overnight.

Exciting times!

✨ Will this tool soon retire PowerPoint?

In the middle of the jungle of 1000+ AI tools, Genspark.ai emerged a few weeks ago. It's not just an ordinary chatbot, but an entire AI agent team!

The coolest features at a glance:

• Mixture-of-agents system: Genspark uses an intelligent model mix to find the optimal one for each task. All important models are available to you.

• Practical agents: Deep Research researches for you, creates in-depth articles with verified facts, does market research, analyzes trends, etc.

• Image generation: Simply enter your prompt and it will select the best model - choose from GPT Image, FLUX.1, Ideogram, Recraft, Gemini Imagen 3, ....

• Image editing: You can also use tools such as Background Remover, Magic Eraser, Redraw, Unblur, Expand and AI Try-On (for clothes).

• Video generation: You can create a video with a prompt or image, with Kling, MiniMax, DreamMachine, Veo, RunwayML, Hunyuan, ...

• Makes calls for you: The AI can make simple calls (USA, Canada & Japan).

• Mobile app: And with the app you always have all the features in your pocket!

Our tests show: Genspark beats all other slide generators

While everyone talks about Genspark's research features (which are impressive), most overlook the standout feature: automatic presentation creation!

We pitted Genspark against the most popular AI presentation tools. The result? A knockout in the first round! You can watch the result right here - it's really strong, with only one prompt.

If you want to make adjustments, you can do this via the Genspark chat - or you can have the html code generated (for example, to upload it to your web server) and customize it there ("Can you create a scrollable HTML page for me that contains the code of all slides?"). The export to a PDF is usable, PPTX rather suboptimal - you will have to work on that.

The prompt for creating the presentation was super simple:

Create me a comprehensive slide deck with all the features and use cases for https://www.genspark.ai. Make sure to add screenshots / animated gifs.

We have also tested a number of other tools with the same prompt.

But to be honest, we like Genspark's output the best by far - and it's also the only tool that delivers substantial content!

You can test it here for free.

Many of the other providers have either done superficial research or no research at all. (Of course, you can address this by doing the research first and then giving this as input to the tool).

Here are the alternatives:

• Gamma.app 

• Canva.com (with the new "Canva AI" feature)

• ChatLLM's new slide feature

• Prezi.com - disappointing, didn't do any live research

• Popai.pro - dug up the slide templates from the 90s 😵‍💫

• PowerPoint - you can create slides with Copilot, images based on GPT image

• In Google Slides, you can create individual slides with Gemini or use (mostly paid) app extensions from the marketplace.

Our take: PowerPoint skills will be as relevant as fax knowledge in 2025 ...

Genspark generates not just ANY presentation, but slides with substance. Anyone still creating slides manually in 2025 is wasting valuable (life) time.

Stop doing things that AI should be doing for you!

Instead, start reorganizing your presentation workflows while others are still dragging pixelated stock photos into PowerPoint.

🎥 Runway Gen:48 - the ultimate test for your creativity

Runway is one of the very best AI video generators and impresses with its realistic presentation and cinematic effects. The latest edition of the Gen:48 from Runway just closed its doors. The competition is tough - but super inspiring, and we can’t wait to see the winners! ↴

We made it! But no need to be sad. The AInauts will be back soon, with new stuff for you.

Reto & Fabian from the AInauts

P.S.: Follow us on social media - that motivates us to keep going 😁!
X, LinkedIn, Facebook, Insta, YouTube, TikTok

Your feedback is essential for us. We read EVERY comment and feedback, just respond to this email. Tell us what was (not) good and what is interesting for YOU.