🗞 What's New: Claude just made it cheaper to build AI apps

ClaudeAI recently introduced a feature called Prompt Caching:

• It essentially lets you "reuse text" across multiple prompts. You can "cache" the examples, saving you time and money.
• The 5 stages of being a creator are linear; you will experience each of them. Determine which one you're in to see how to get to the next level!
• $6K MRR with a startup studio. Best friends Matthias Neumayer and Dima Rubanov are rooting out bias in AI with their kid's story generation app.

Ready to grow a vibrant community and access new revenue streams? Try Membership! Memberful seamlessly integrates into your tech stack, allowing you to offer exclusive content and membership perks. #ad

from the Trendy Software Ideas newsletter by Darko

Imagine you're creating an AI app. You're heavily dependent on OpenAI, Claude, or Google's API, and one of your key competitive advantages is your advanced prompt.

However, your advanced prompt is quite lengthy, and includes many examples. This, of course, helps the AI generate a useful output.

This can quickly drive up your API costs, since you're constantly sending the same long input. Providers like OpenAI charge you based on input tokens, meaning the more words you send for the AI to process, the more it costs.

Things can get expensive fast.

Welcome to the world of Prompt Caching

ClaudeAI, one of the top three AI providers in the world (alongside Google and OpenAI), has recently introduced a feature called Prompt Caching. This feature essentially lets you "reuse text" across multiple prompts.

Imagine you have a large prompt with plenty of examples. The examples make up 90% of the prompt's length.

With this feature, you can "cache" the examples, only sending the remaining 10% as the actual prompt.

The impact of this is dramatic: It's up to 90% reduction in input API costs.

This also means that developers can either lower pricing, or increase profit margins.

What is Prompt Caching useful for?

According to the creators of Claude, this is useful if you provide:

• AI assistants: You expect multiple users to enter the same prompt.

• Code generation: You need to reuse the same prompt, or have multiple users working with the same template.

• Code reviews: When asking AI to review long chunks of code, you don't have to send the same code over and over again. This can save both time and money.

• Processing large documents: An example would be feeding the AI a novel, and asking questions about it.

• Any search tool: You can input data from files, and ask questions.

• Any prompt with plenty of examples: You no longer need to worry about optimizing your prompt for length. You can focus on being thorough and getting the best results possible.

Will OpenAI follow suit and release a similar feature soon? We'll see!

Discuss this story, or subscribe to Trendy Software Ideas for more.

This issue is sponsored by Memberful

Ready to grow a vibrant community and access new revenue streams? Try Membership! Memberful seamlessly integrates into your tech stack, allowing you to offer exclusive content and membership perks. Engage your audience, grow your brand, and enjoy reliable, recurring income with Memberful.

Get started for free!

by Jay Clouse

Every creator's journey is unique. It's helpful to understand where you are on your journey, so you know how to navigate your next steps.

In my experience, there are five typical stages of the creator journey. You don't need to achieve all five stages in order to be successful. But, for a creator to build a comfortable, sustainable living, they will likely need to at least achieve up to stage four. Let's explore!

Stage one: Consideration

People in the consideration stage have not yet decided that they will pursue being a creator, but they're actively considering giving it a shot.

Common challenges:

• Self-confidence.

• Lack of knowledge.

• Fear of time commitment.

• Lack of direction.

Success in this stage is simply making a decision.

Stage two: Finding fit

You need to find product-market fit by presenting a solution to your target audience's problem.

Common challenges:

• Creator identity.

• Determining your ideal customer.

• Differentiation.

Success in this stage is finding clarity on your purpose as a creator.

Stage three: Finding traction

By this time, you should have some signals that the direction of your product is resonating with your target audience. In stage three, you'll need to work to consistently receive positive feedback from your target audience.

Common challenges:

• Audience discovery.

• Audience retention.

Success in this stage is consistent audience growth.

Stage four: Going pro

This stage is where amateur, part-time creators decide that they want to level up and become professional creators. In this stage, creators lean into the traction they've begun to experience, and double down on what is working. This is also the stage where monetization becomes a major focus.

Common challenges:

• Financial stability.

• Audience discovery.

• Sustainable revenue model.

• Business building.

Success in this stage looks like consistently earning enough revenue to support yourself (and potentially a small team).

Stage five: Scaling

This stage is where creators truly become full-fledged companies. Often, the systems built to get to, and through, stage four are not strong enough to support a scaling business.

Common challenges:

• Product expansion.

• Team building.

• Financial growth.

• Back office operations.

• Removing bottlenecks.

Success in this stage looks like successful team and business building, higher output, and accelerating growth.

Wrapping up

These five stages are linear; you will experience each of them on your journey, even if only for a short period of time.

Your journey may end or linger in one of the earlier stages. Not every creator wants to go pro, and not every professional creator has ambitions for scale.

Understanding these stages, and the unique challenges they present, can help you see where you currently sit, and how to move up the ladder!

(This article was originally published here.)

Discuss this story.

from the Trendy Software Ideas newsletter

📱 All the leaks and rumors about the next iPhone.

🤖 Researchers have ranked the nicest and naughtiest AI models.

💲 Link to your product here. Our most affordable ad.

📈 Creative B2B ads perform better as time to purchase grows.

💰 Selling a startup in "acqui-hire" is more lucrative than it seems.

❤️‍🔥 Has Burning Man burned out?

Check out Trendy Software Ideas to discover more software ideas inspired by recent news.

by Katie Hignett

When COVID-19 hit, Matthias Neumayer and Dima Rubanov started coding together for the first time since they were teens. What started as a side hustle soon became a full-time job.

They launched a kid's story generation app, then a brand positioning app, then a PDF chatbot. Here's how they're building their startup studio.

Beginnings

Dima: We've been best friends since high school. We coded as teenagers, and ran an internet forum about filmmaking and politics when we were about 13. That was basically our first startup.

Then, I studied economics and finance, and worked in consulting for seven years.

Matthias: I studied law and filmmaking at university. I spent eight months as a trainee lawyer before realizing I didn't want to do it anymore.

Then, I worked in advertising and media production, and had a successful agency for eight years.

The first product

Dima: We created Oscar Stories, an app that generates bedtime stories for children. We made it in two months, launching a very raw MVP on Product Hunt in March 2023. We got about 500 signups on the first day.

Matthias: The core tech stack is the same now as when it first launched. The frontend is React Native, and we have an Express server for the backend.

We decided to build most things ourselves, which was a good decision for scaling. It started out pretty basic, but it's way more complicated now.

Rooting out bias in AI

Matthias: We are at 60K users. We're doing some fine-tuning to make sure our AI model uses language that's accessible and appropriate for children.

Dima: With this new fine-tuning, we are also launching a second storytelling app called LORA, for which we got a government grant.

It's going to try and reduce gender bias within children's stories. The existing training data has a lot of bias, with boys always saving girls.

Business model and marketing

Dima: In the beginning, the app was free. Now, we've added a one-time purchase coin system to the subscription option, where you can buy a bunch of coins, or subscribe for unlimited stories.

We also introduced parity pricing, which had a major impact on sales.

Matthias: It's really hard to market children's products. You can't track users in the way you normally would, so we don't have a lot of data.

We've been lucky with media coverage. We did a lot of press releases, and I also had some PR contacts from my old work.

After some local coverage, we got invited to startup fairs and other IRL events. We were also featured in a Wired article, which was amazing. We have no idea how they found out about us.

If you don't have the chance to go to a live event, a service like Rapid User Tests, where a person is filmed using an app, can be very helpful.

Building a startup studio

Dima: We worked on Oscar Stories for the next nine months before launching Branding5, an AI branding tool.

Five months after launching Branding5, we launched another app, FragDasPDF, a PDF chatbot for the German market. We quickly realized that a chatbot wasn't enough for our customers, who are largely students and researchers.

So, we're about to launch an add-on product with a word processing and formatting interface. It also has AI commands. You can select a paragraph, rephrase it, then save what you've produced to a Word Doc or PDF.

We want a student to be able to interact with their library of PDFs, and write their whole thesis within our product.

Matthias: Our vision is to become a startup studio, consistently putting out bootstrapped products. We want to be number one, quality-wise, in our niches.

Discuss this story.

by Tweetmaster Flex

I post the tweets indie hackers share the most. Here's today's pick:

Forward it to a friend, and let them know they can subscribe here.

Also, you can submit a section for us to include in a future newsletter.

Special thanks to Jay Avery for editing this issue, to Gabriella Federico for the illustrations, and to Darko, Jay Clouse, and Katie Hignett for contributing posts. —Channing