How to Easily Implement OpenAI's New GPTs For Growth. Essential Insights For Entrepreneurs.

Don't Miss Out on the AI Gold Rush - Boost Profits, Slash Costs, and Transform Your Business with OpenAI's most important releases.

AuthorAuthorAuthor

Candyd Mendoza, Maxi Veloso & Benson Sung
November 12, 2023

In today’s Future Friday…

Welcome to this week’s Future Friday, your all-access pass to OpenAI’s first DevDay.

There’s so much to unpack here, so let's roll up our sleeves and turn today's wow into tomorrow's win.

Listen to today's edition:

Read Online and listen to the audio version.

Are you a non-technical founder feeling overwhelmed with the tech side of your business? ScaleLabs is here to be your trusted partner. 

From crafting a loved MVP to building a dedicated tech team that scales with your vision, we've got you covered. 

Dive into our FREE guide and discover how you can revolutionize your business operations. A proven, simple framework, used to successfully build and launch over 10 SAAS companies.

Don't just build, scale with confidence.

Join 9,000+ founders getting actionable golden nuggets that are tailored to make your business more profitable.

TOPIC OF THE WEEK

The Tipping Point: Mind-blowing Developments at OpenAI’s DevDay

Remember when days used to be a mad dash from one task to the next? Emails, bug fixes, and that never-ending to-do list.

Now, flip that script. You're chilling with your morning brew while GPT-4 Turbo drafts pitches that sound like you, only wittier.

DALL-E 3's churning out Insta-worthy images before you've finished your avocado toast.

And those meeting minutes? Whisper v3's got them down better than you remember.

Your day's gone from 'do it all' to 'AI's got this.' OpenAI's tools haven't only upgraded your startup—they've upgraded you. And guess what? They’re just getting started.

Early this week, OpenAI wowed the whole world during its first-ever DevDay.

It wasn't just another boring tech conference for developers; it was the day OpenAI gave everyone a front-row seat to the future.

And let me tell you, it's looking bright.

A Quick Overview of OpenAI’s Groundbreaking Updates

Don’t worry; we are not here to repeat like a broken record. Just to give you an idea of what could move the needle in your business and what you can do about it.

  1. GPT Turbo: Think of GPT-3.5 and GPT-4 on steroids. Faster, cheaper, and with higher rate limits.

  2. Updated Knowledge Base: with a cutoff in April 2023. Finally!

  3. Assistants API: with new multimodal capabilities such as code interpretation, retrieval, and function calling, simplifying the development of high-quality AI apps. Powerful stuff.

  4. Custom models program: organizations can now train custom GPT-4 models tailored to their specific domain. Game-changer.

  5. Copyright Shield: OpenAI will step in and defend customers and pay the costs incurred if they face legal claims around copyright infringement.

  6. Lower prices, greater affordability: with inputs 3x cheaper and output 2x cheaper for GPT-4 Turbo.

What a big leap, huh? And this is just a surface scratch of all the implications that this brings.

💰 Impact On Your Business

OpenAI's developments represent a shift in how businesses and society interact with technology.

  • The Assistants API could redefine productivity by automating complex tasks

  • The GPT Store may become a new economic platform for AI-driven applications, where users can share their GPTs with others

  • The ability to create custom GPT models could democratize AI, making it accessible to a broader audience and sparking a wave of personalized AI applications

We’re releasing a podcast next week diving deep into these aspects so stay tuned for that!

Now, let’s run some numbers: 🤓 

Comparing Costs For Different Use Cases — Is it worth your buck?

As Reuven Cohen – who’s part of the first group to test ChatGPT – puts it, the 128k context model is “powerful and PRICEY.”

Here’s an idea of how much it could cost for different uses combining input and output costs, combined:

  1. Complex Legal Analysis: $1.88 per request.

  2. Extensive Technical Support: $0.88 per request

  3. Code Review for Security Vulnerabilities (25,600 lines of code): $2.03 per request

Sure, a few bucks per request doesn’t seem like that big of an investment. But many companies are worried these costs could snowball into tens of thousands of dollars per day.

For a 10,000-person company, we’re talking about daily combined costs of around $17,600 and $40,600 depending on the tasks.

So costs may seem like a valid concern.

But is it really?

This 10,000-person company would be paying millions per day, considering wages and time to completion, for the same tasks.

So this is ridiculously cheaper than the traditional non-AI way. Indeed a game-changer.

Savings that are orders of magnitude greater

If you're still hesitant and looking to further slash costs, using the API may not even be the best choice. Your savings and productivity gains could be orders of magnitude greater.

How?

We recently did an interview with the co-founder of a company that turns expensive LLM prompts into fast, cheap fine-tuned models.

Coming soon to our podcast! — Hint: we’ve mentioned this company previously in one of our Future Fridays. 😉 

Stay tuned for that interview, where we’ll show you:

  • How to build your prototype using GPT-4,

  • then turn your prompts into fine-tuned models,

  • get them running as fast as possible,

  • PLUS reduce your costs between 10X and 20X — so you don’t have to worry about costs anymore.

💡 Ideas to Marinate

  • Companies that rely on OpenAI APIs are in for a boost. They could either become more profitable or decide to offer lower prices to their users thanks to these cost savings.

  • Brace yourself for a surge in new apps. Just like when GPT-3 was introduced, we saw a ton of new AI-powered apps called GPT wrappers hit the scene.

  • It's likely that we'll experience a similar explosion due to custom GPTs, we're going to talk about in just a minute.

How you can — and should — leverage GPTs

❝

Eventually you'll just ask a computer for what you need and it'll do all of these tasks for you.

Sam Altman

Another genius move from OpenAI is the launch of GPTs aka agents, which are custom versions of ChatGPT with specific functions – no coding required.

With an easy-to-use interface, natural language creCuation, and multimodality capabilities such as web searching, image creation, data analysis or retrieval, you can create a GPT for a specific purpose, whatever floats your boat.

Of course, you can — and should — leverage them in multiple scenarios:

  • Privately for your personal productivity boost,

  • Within your company if you’re using Enterprise, or

  • Publicly if you or your company aim to profit directly from these GPTs agents.

Then there’s also OpenAI’s new Assistants API, which follows a similar fashion than GPTs, but this is a tool crafted for developers.

The API makes it easier to build complex AI applications by doing much of the hard work.

Key features include:

  • Code Interpreter for writing and running Python code.

  • Retrieval for gathering external data (We have several pieces about Retrieval Augmented Generation or RAG you might want to check out).

  • Function Calling for incorporating custom functions into responses.

You can experiment with Assistants API beta in OpenAI’s Assistants playground.

GPT Store - 100 million WAU are waiting for your GPT

GPT Store was the next obvious move by OpenAI and is rolling out later this month.

If you’re one of the people who wishes they had created an app for the Apple App Store when it came out in 2009, now’s your chance with ChatGPT.

Roughly 100 million people use the tool weekly. 😉 

So if your company has a product/solution that could be turned into a GPT, there’s no better time to put yourself in the storefront.

And what if you just want GPTs for you or your company’s internal use?

Either if you plan to take this opportunity to market a product for the public, or if you want to improve your personal or company’s productivity, here's a high-level framework to consider when building your first GPT:

Here’s what it’ll look like once you’re inside.

⚒️ Actionable Steps

  • Access GPTs: You need ChatGPT Plus or ChatGPT Enterprise to access GPTs.

  • Create your GPT: Use plain language to customize it, providing prompts and instructions for its specific function.

  • Configure your GPT: More advanced settings are available in the 'Configure' tab, where you can further refine your GPT.

  • Upload Custom Data: You can upload your own data (like text files) to train your GPT on specific knowledge.

  • Set Capabilities: Enable functionalities like web browsing, image generation, and advanced data analysis as needed.

  • Test and refine: Use the live preview to test changes and refine your GPT's capabilities and responses.

  • Publish Your GPT: Once satisfied, you can save your GPT for private use, share it with others, or publish it publicly, potentially in the upcoming GPT Store.

CAVEMINDS’ CURATION

Create an AI-Powered App With Zero Coding Knowledge

During DevDay, Altman created a GPT within 4 minutes.

4 freakin’ minutes.

And just a couple of hours after the event, we started to see the main GPTs out there in the wild. Some of them:

Have 4 spare minutes? Some ideas to build your first GPT

⚒️ Actionable Steps

Access the GPT builder, and try it with this Founder Productivity GPT idea:

Concept:

  • To assist founders in managing their time effectively.

  • To automate routine tasks and streamline decision-making processes.

  • To provide instant access to business intelligence and market research.

  • To facilitate strategic planning and goal setting.

Features:

  • Task Automation: Automate scheduling, email responses, and social media management.

    Prompt: "Develop capabilities to automate scheduling, manage email communications, and oversee social media accounts."

  • Decision Support: Offer pros and cons for different business decisions based on data.

    Prompt: "Analyze business scenarios to provide balanced viewpoints, outlining pros and cons based on available data."

  • Market Insights: Provide real-time market trends, competitor analysis, and customer feedback synthesis.

    Prompt: "Continuously gather and synthesize market trends, competitor analysis, and customer feedback for timely business insights."

  • Goal Tracking: Set and monitor progress towards business milestones.

    Prompt: "Implement a system to set, track, and report progress on business milestones and goals."

  • Learning and Development: Suggest articles, courses, and resources for skill development.

    Prompt: "Curate and recommend educational content, such as articles, online courses, and skill development resources tailored to the founder's needs."

  • Networking: Recommend networking opportunities and introductions based on industry relevance.

    Prompt: "Identify and suggest valuable networking opportunities and introductions based on industry relevance and personal business goals."

Want to go deeper?

  • Neatwork AI - The enhanced version has integrated ChatGPT directly into VS Code to keep your development conversations in stride with your coding.

Visit our Cyber Cave and access the most extensive tool database on the internet.

NEEDLE MOVERS

Elon’s not impressed. This week, he’s unleashed his latest brainchild: Grok.

Clearly, he wasn't about to sit back and watch the AI party without crashing it.

“Grok is an AI modeled after the Hitchhiker’s Guide to the Galaxy, so intended to answer almost anything and, far harder, even suggest what questions to ask!” xAI said in its blogpost.

What’s the most interesting about Grok so far? It’s fed with real-time info via the X platform.

It's still in the early testing phase so it’s soon to judge, but you can get early access program. Participation in the early access program is currently limited to verified users.

Definitely something to keep an eye on!

Three days after DevDay, OpenAI unveiled its new Data Partnerships program.

The AI giant wants to team up with different organizations and create a huge range of datasets for AI training.

There are two ways to get involved: Either by contributing to a public open-source dataset or by working on private datasets for specific AI models.

Imagine turning your palm into a mini-display that you can wave your hand or just say the word to control it.

It remembers stuff for you, sorts out your notifications, and even keeps up with missed calls and messages. Plus, it makes playlists, tracks your health stats, dishes out nutrition tips, translates languages on the fly, and lets you snap moments just by tapping or giving a voice command.

That’s what Humane is doing.

Wearable AI is now more real than ever before.

What OpenAI missed in their event: Agents interaction

Adept launched a new series of experiments, starting with Workflows, a tool that allows AI teammates to learn and perform specific tasks, customized for individual users or companies.

Workflows can automate repetitive tasks, extract structured data from unstructured sources, bridge different software tools efficiently, and even assist users in performing tasks they are unfamiliar with.

This is a bold move by Adept, and in our opinion, it offers a strong case for enterprises as agent interactions can virtually automate processes from start to finish.

And guess what? OpenAI did not mention this possibility with GPTs (yet). Adept saw the gap, and addressed it.

That’s all for today!

Continue Reading

We appreciate all of your votesand comments as well! Don't be shy, give us your thoughts, whether it's a 🔥 or a 💩.

We promise we won't hunt you down. 😉

 

🌄 CaveTime is Over! 🌄

Thanks for reading, and until next time. Stay primal!

Reply

or to participate.