Almost Timely News: Getting Started with OpenAI Custom GPTs (Custom ChatGPT) (2023-11-12) :: View in Browser
Content Authenticity Statement
100% of this newsletter’s content was generated by me, the human. However, I used OpenAI’s Whisper to transcribe the video and then Google Bard to clean up the transcript’s quirks. Learn why this kind of disclosure is important.
Watch This Newsletter On YouTube 📺
I strongly encourage you to watch the video version of this week’s newsletter, as there’s a significant walkthrough component that just doesn’t make sense in text.
What’s On My Mind: Getting Started with OpenAI Custom GPTs (Custom ChatGPT)
Okay, it was a huge week in generative AI, particularly with the announcements from OpenAI and the debut of their newest products. Let’s talk about one of them in particular, what it means, what you can do with it, and how to get started.
The product in question is a kind of fine-tuned model you can build and customize on your own. It comes in two flavors: Custom GPT, and Assistants API. They’re essentially the same thing, but one is for the non-technical person to be used within OpenAI’s ecosystem, and one is for the technical person to be used in applications outside of OpenAI’s ecosystem.
This is an important distinction. OpenAI has typically offered its services in those two flavors, technical and non-technical. ChatGPT had a predecessor very few people used, the Playground, where you could interact with a language model and test it out. However, the Playground was intended mainly for developers and as a result never caught on the way ChatGPT did a year ago. Since then, they’ve made strong efforts to offer a technical and non-technical version of their major offerings.
So let’s talk about what a Custom GPT is. Custom GPTs are essentially variants of ChatGPT.
There is a version of ChatGPT that you can customize with your own prompts, data, and add-ons to make it more focused on a particular task.
Custom GPT is a version of ChatGPT that you tailor to focus on one task and get really good at it, better than you can do with prompting alone.
For example, suppose you want to create a Custom GPT that writes in a very specific tone of voice, maybe your own tone of voice.
Before Custom GPT, there were a couple of ways you could have done that. You could have used lengthy prompts explaining your writing style and tone of voice, or you could have done a very technical tuning of the model. However, these methods were either limited in their effectiveness or inaccessible to non-technical people.
Now, Custom GPT allows you to do the same thing, but without coding or technical hoops to jump through. You will still use prompts and examples, but you will be able to provide many more examples and take different kinds of actions.
Once you have built a Custom GPT, you can use it yourself, share it with a link, or publish it to an app store. OpenAI plans to launch an app store for Custom GPTs in the next month or so.
Two things will determine the quality of the Custom GPT you create and sell: the specificity and detail of the prompts you create, and the associated data you provide. The better quality and quantity of the data you put into the system, the better it will perform that task.
This is especially important for content marketers and creators. Every kind of company has lots of data that is not on the internet. ChatGPT and all public language models are trained on publicly available data, but you can use your own data to create a Custom GPT that encapsulates your strategic knowledge and point of view in a way that others cannot replicate.
You can also use social media data, census data, and other publicly available data sets to create a Custom GPT that is more focused on your specific needs.
Part of what makes Custom GPT a powerful tool is the curation. You don’t just load all your data into it like the public models do. With Custom GPT, you are influencing the model’s capabilities by saying, “Here’s what I want you to check first. This is more important information than everything else you think you know.”
So even if you were just curating specific information, your version of ChatGPT would probably behave better. For example, if you were working in virology and immunology, you could curate your Custom GPT with only peer-reviewed, credible scientific research. This would help the model perform better, because you would be saying, “Hey, check this stuff first. Not your uncle Bob’s random Facebook posts ranting about so and so. No, here’s peer-reviewed stuff from Nature magazine and Cell magazine and all the scientific journals I rely on. This is what I want you to use first.”
So even if you’re using public data, your curation of that data, quality and quantity, really matters with Custom GPT.
The second factor that determines the quality of your Custom GPT is the quality and quantity of your ideas. Your ideas are important. Your data is important, but your ideas are even more important, because you’re making a language model, essentially making a language model app really, that is predicated on you having ideas about what you want this thing to do.
Great ideas, great creativity, great imagination is going to create great outputs because the GPT models are going to build to help you build this thing and run this thing. But it’s still got to be a good idea.
For example, if you built a GPT model and said, “Hey, I want this model to just count. Just count 123.” Yeah, it will do it. But that’s not a great idea. Right? That’s not super compelling.
So what are the ideas that you could have? Things like a game time mode, a negotiator mode, a creative writing coach mode, a tech support advisor mode, a laundry buddy mode. Right, you can see how these are kind of like custom advisors, custom apps.
What are you good at? What are you good at? What do people ask you for help with? Like you personally? What do people ask you for help with? Do you have data in machine-readable format? Do you have your writing or your thoughts or transcripts of you speaking? Could you build a Custom GPT that behaves like that thing that you’re good at? And if it works well, could you sell it? Would people pay money to the app version of you, your digital twin? Could you create something like that, that was specialized, that could do that thing only, or predominantly? That’s the value of these things.
So you need data, and you need ideas, and you need both in high quality and high quantity to make Custom GPT work for you.
You go into your ChatGPT account, you hit that explore button on the left hand menu, let’s put the rest of that menu away. And you’ll see here, this should be available now to all accounts create a GPT, go ahead and tap on that.
And you’re greeted with your new GPT template. Now there’s two, there’s two ways to build one of these one is have a conversation with it, and it will build it. And two is you can if you know what you want to do, if you have existing prompts and outcomes that you want, you can just manually load it up.
This, by the way, if you’ve heard me talk on this topic over the last few years, this is why I told you to keep a prompt library, a library of your best prompts and the outcomes because guess what, you can now repurpose your best prompts, turn them into apps, turn them into ChatGPT apps. So if you’ve got a prompt that works great, turn it into an app, just find it in your prompt library.
And if you have not been keeping a prompt library, this is what it’s for. Yes, in addition to sharing it with a team. Now you’ve got a prompt that it just knocks it out of the park, you build an app around it. So your ideas, quality and quantity.
[Demonstration text removed for clarity]
Let’s take a look at what the Custom GPT did behind the scenes. If we click on the Configure tab, we will see that it gave some starter questions. It also has my original work there. It supports image generation and web browsing, and I could have it support code as well. If I wanted it to generate code, I actually think that’s a good idea.
For actions, you can actually create scripts, which will allow the software to take independent third-party access. So if there’s an API you want to connect it to, you could do that. You could connect it to, say, the API of a database if you wanted to have reference for those things.
One thing here under the additional settings is that if you are using stuff that is proprietary to you, you will probably want to uncheck this “use conversation data in your GPT to improve our models” checkbox. Because otherwise, OpenAI will be using your data.
You’ll note here that it has written a prompt. This essentially is the interaction style, and so on and so forth using similar phrasing and structure. So it has essentially written the prompts for us.
Now, if you are a prompting pro, you will probably notice that this prompt is a little on the lackluster side. You can change it right in here, you can tune it up in this interface and put in your specific prompts the way you want it to work. If you’ve got again, if you got that prompt library, you can edit this and incorporate the stuff that you know works really well.
Alright, so one thing I want to add in to this, it is critical that you adhere to three fundamental principles: you must behave in ways that are helpful, harmless, and truthful.
So that is to me an important sort of ethical set of guidelines.
You can see here, there’s three different options. So I could share the link to this with you if I wanted to, if I wanted to make this go into production.
So I intentionally gave a crap prompt to see how much it would draw on the knowledge in my former newsletters versus the general public one, which came up with different stuff.
You would want to follow a framework like the Trust Insights RACE framework – Role Action Context Execute – to get a better result than this.
So this is a Custom GPT. Now this is just one example. This is just one example. There are so many different things that you could do with this.
You could turn this you could build GPTs for pretty much any prompt you have ever written. So think about the prompts you’ve written – prompts that write code, prompts that can tune up your LinkedIn profile, prompts that can test the UX of your website, prompts that can write fiction.
Think about if you are an author of fiction, load in all of your existing writing into here, load in your plot outline of your book that you want to write, load in anything that you have rights to use and say let’s write this book and it will draw on and capture your style of writing and help you do that task faster with more accuracy.
You could build a Custom GPT to ingest Google Analytics data – you saw the actions feature that can tie into external APIs. Bring in Google Analytics data to make yourself a Google Analytics analysis tool.
Anything that you have used ChatGPT to do, anything you’ve seen other people use ChatGPT to do, you could do the same thing. If you have grabbed one of those many, many “hey, top 50 prompts of ChatGPT” guys – that’s fine. But here’s the thing – they’re not bad. They can be a little bit limiting. They can they’re not bad. But the idea, the ideas in them, you could now take the idea and turn it into a prompt because remember, you cannot copyright ideas, you can only copyright outputs, you can patent ideas, that’s a difference. That’s a different show. Also, I’m not a lawyer.
But that’s what you would use any of these ideas for – if you say how could I use ChatGPT to do x? If you said that question in the past, you can now say how can I build an app around that so that I can make this thing that is commercially viable, maybe.
This is a really, really cool, powerful, interesting technology. And a lot of people are going to create in the first generation, they’re gonna create a lot of crappy applications, right? This could be a lot of unhelpful stuff in the GPT story initially, but as time progresses and as people’s eyes open, and there’s people experiment, they will be, there’ll be some pretty cool stuff.
Here’s the thing – you need first mover advantage. So you should be testing this out now for yourself, maybe for your company, maybe for a new line of business for you, you should be testing this out yourself, so that you can see what is capable of what it can do what its limitations are.
So that when the store opens, you’re ready to go. If you if you want to turn this into a revenue stream, you would want to have this thing be in testing now, here with your community. If you have a LinkedIn group or slack group or Discord server, build your apps now, get your beta testers, just beating them up for free now and see what will happen.
Now this requires the paid version of ChatGPT – requires the paid version. So if you are not paying the 20 bucks a month for it, if you are financially able to do so, it is worth it. You might even make your money back with these with these tools and technologies.
It is amazingly cool stuff. I look forward to seeing what you create. If you want if you have stuff that you’re creating that you want to share, I would love to see it. Drop it in Slack. If you’ve not already gone to my Slack, go to TrustInsights.ai/analyticsformarketers and drop into the Slack group there. Love to see what you’re going to build with this stuff.
How Was This Issue?
Rate this week’s newsletter issue with a single click. Your feedback over time helps me figure out what content to create for you.
Share With a Friend or Colleague
If you enjoy this newsletter and want to share it with a friend/colleague, please do. Send this URL to your friend/colleague:
For enrolled subscribers on Substack, there are referral rewards if you refer 100, 200, or 300 other readers. Visit the Leaderboard here.
ICYMI: In Case You Missed it
Besides the newly-refreshed Google Analytics 4 course I’m relentlessly promoting (sorry not sorry), I recommend the piece on monetizing data with generative AI. Combined with today’s piece on custom GPTs, you’ll be monetizing your data as soon as you’re able.
- You Ask, I Answer: Monetizing Data for Generative AI?
- You Ask, I Answer: Style Transfer in Generative AI Writing?
- You Ask, I Answer: Reducing Generative AI Hallucinations?
- You Ask, I Answer: Managing Rectangular Data with Generative AI?
- You Ask, I Answer: Privacy-Friendly AI Transcription Software?
- Mind Readings: Custom GPTs from OpenAI and Your Content Marketing Strategy
- Almost Timely News: The Greatest Unaddressed AI Challenge
- OpenAI, Hubspot, Substack and More!
- In-Ear Insights: Add AI To Everything? CEO AI Strategy
Skill Up With Classes
These are just a few of the classes I have available over at the Trust Insights website that you can take.
- 👉 Google Analytics 4 for Marketers
- 👉 Google Search Console for Marketers (🚨 just updated with AI SEO stuff! 🚨)
- ⭐️ The Marketing Singularity: How Generative AI Means the End of Marketing As We Knew It
- Powering Up Your LinkedIn Profile (For Job Hunters) 2023 Edition
- Measurement Strategies for Agencies
- Empower Your Marketing With Private Social Media Communities
- Exploratory Data Analysis: The Missing Ingredient for AI
- How AI is Changing Marketing, 2022 Edition
- How to Prove Social Media ROI
- Proving Social Media ROI
- Paradise by the Analytics Dashboard Light: How to Create Impactful Dashboards and Reports
Advertisement: Catch My Talk at Stonehill College Live this Week
Stonehill College Division of Graduate & Professional Studies will host data scientist, best-selling author, and podcaster Christopher S. Penn, who will discuss the latest trends in content creation, advertising, emerging technology and artificial intelligence as part of the free virtual GPS Speaker Series on November 15 at 7 PM Eastern Time.
Christopher S. Penn will discuss how artificial intelligence (AI) and data analytics are impacting marketing and business strategies. Join us virtually and learn how large language models (LLMs), driven by generative AI, can revolutionize your marketing strategy analysis, provide data summarization and recommendations, generate valuable insights, reduce bias, increase productivity, spot trends, and manage risks.
Get Back to Work
Folks who post jobs in the free Analytics for Marketers Slack community may have those jobs shared here, too. If you’re looking for work, check out these recent open positions, and check out the Slack group for the comprehensive list.
- Adobe Analytics Lead at Themesoft Inc.
- Associate Director For Digital Analytics Infrastructure (M/F/X) at HelloFresh
- Client Acquisition Cro Manager at Rocket Mortgage
- Director Of Data And Insight at Journey Further
- Enterprise Marketing Analytics Consultant at Harnham
- Head Of Analytics & Data Science at Intrepid Digital
- Linkedin Ads Specialist- Paid Media  at Onward Play
- Senior Associate, Analytics at Carat
- Senior Growth Engineer at Rocket Money
What I’m Reading: Your Stuff
Let’s look at the most interesting content from around the web on topics you care about, some of which you might have even written.
Social Media Marketing
- The Complete Guide to TikTok Influencer Marketing via Sprout Social
- Integrating TikTok Shop Into Your Social Strategy via Sprout Social
- YouTube pages are getting a TikTok-like For You feed via The Verge
Media and Content
- What Corporate Communications Will Look Like In 2024
- Top Takeaways from Ragan’s Future of Communications Conference via PR Daily
- Content Brand: How to Differentiate Your Company with Content
SEO, Google, and Paid Media
- What is thin content? SEO for beginners Yoast
- SEO Trends for 2024: SGE, People-First Content, and More!
- The Official Ahrefs Tutorial: How to Use Ahrefs to Improve SEO
Advertisement: Business Cameos
If you’re familiar with the Cameo system – where people hire well-known folks for short video clips – then you’ll totally get Thinkers One. Created by my friend Mitch Joel, Thinkers One lets you connect with the biggest thinkers for short videos on topics you care about. I’ve got a whole slew of Thinkers One Cameo-style topics for video clips you can use at internal company meetings, events, or even just for yourself. Want me to tell your boss that you need to be paying attention to generative AI right now?
Tools, Machine Learning, and AI
- My North Star for the Future of AI via The Atlantic
- Samsung’s Leap into the Future of AI and Chip Technology
- OpenAI Turbocharges GPT-4: Speed and Affordability
Analytics, Stats, and Data Science
- Impact of Generative AI in Data Storytelling & Analytics
- KOSMOS-2: A Multimodal Large Language Model by Microsoft via Analytics Vidhya
- YouTube Consumption via Trust Insights Marketing Analytics Consulting
All Things IBM
- Transforming small farming with open-source, AI-powered connected edge solutions via IBM Blog
- Five ways to reduce your public cloud spend with IBM Turbonomic via IBM Blog
- Reducing administrative burden in the healthcare industry with AI and interoperability via IBM Blog
Dealer’s Choice : Random Stuff
- Roblox report highlights Gen Z’s sense of digital self-expression via VentureBeat
- Janet: America’s Secret Airline
How to Stay in Touch
Let’s make sure we’re connected in the places it suits you best. Here’s where you can find different content:
- My blog – daily videos, blog posts, and podcast episodes
- My YouTube channel – daily videos, conference talks, and all things video
- My company, Trust Insights – marketing analytics help
- My podcast, Marketing over Coffee – weekly episodes of what’s worth noting in marketing
- My second podcast, In-Ear Insights – the Trust Insights weekly podcast focused on data and analytics
- On Threads – random personal stuff and chaos
- On LinkedIn – daily videos and news
- On Instagram – personal photos and travels
- My free Slack discussion forum, Analytics for Marketers – open conversations about marketing and analytics
Advertisement: Ukraine 🇺🇦 Humanitarian Fund
The war to free Ukraine continues. If you’d like to support humanitarian efforts in Ukraine, the Ukrainian government has set up a special portal, United24, to help make contributing easy. The effort to free Ukraine from Russia’s illegal invasion needs our ongoing support.
Events I’ll Be At
Here’s where I’m speaking and attending. Say hi if you’re at an event also:
- Social Media Marketing World, San Diego, February 2024
- MAICON, Cleveland, September 2024
Events marked with a physical location may become virtual if conditions and safety warrant it.
If you’re an event organizer, let me help your event shine. Visit my speaking page for more details.
Can’t be at an event? Stop by my private Slack group instead, Analytics for Marketers.
Events with links have purchased sponsorships in this newsletter and as a result, I receive direct financial compensation for promoting them.
Advertisements in this newsletter have paid to be promoted, and as a result, I receive direct financial compensation for promoting them.
My company, Trust Insights, maintains business partnerships with companies including, but not limited to, IBM, Cisco Systems, Amazon, Talkwalker, MarketingProfs, MarketMuse, Agorapulse, Hubspot, Informa, Demandbase, The Marketing AI Institute, and others. While links shared from partners are not explicit endorsements, nor do they directly financially benefit Trust Insights, a commercial relationship exists for which Trust Insights may receive indirect financial benefit, and thus I may receive indirect financial benefit from them as well.
Thanks for subscribing and reading this far. I appreciate it. As always, thank you for your support, your attention, and your kindness.
See you next week,
Christopher S. Penn
You might also enjoy:
- How To Determine Whether Something is a Trend
- What's the Difference Between Social Media and New Media?
- How To Set Your Consulting Billing Rates and Fees
- How To Set Your Consulting Billing Rates and Fees
- Google Analytics 4: User Medium vs Session Medium
Want to read more like this from Christopher Penn? Get updates here:
Get your copy of AI For Marketers