Hey there 👋
In this week’s Product Prompts:
Learn to use AI to help you build AI apps
Deploy AI Agents to do your work
Map out ideas with an AI Copilot
Atlassian embeds OpenAI in JIRA
and more 👇
📣 Announcements
I built 🎧 productprompts.fm - music created with AI to boost focus and creativity. I built this whole site with the help of ChatGPT, and it took me just 2 hours. Try it out and let me know if it helps you get in the zone 🧠.
🛠️ I also built a tool for taking User Interview audio and video recordings, transcribing them, identifying key painpoints, and sending the insights to my Notion. This kind of AI-powered automation is something we can and should all be doing.
I want to show you how to build your own innovative AI apps like this, which brings me to some more news:
🚢 I’m launching a new course / school called ShipGPT. Learn how to use the OpenAI API to code new projects, automate your tasks, and build new products and MVPs. If you’re interested, sign up here 👉 shipgpt.school.
The core course will be free and available for you to do in your own time. Extras such as weekly Office Hours, guest AMAs, mastermind groups etc will be part of a paid plan.
🔥 Hot take on this: Generative AI is going to create a whole new wave of one-person, full-stack teams, changing how we work, how we build products and even how we start companies. And this will be in 2023, not some time way of in the future.
🏃 I published a new e-book, The AI-Powered Design Sprint Playbook. Design Sprints are great for bringing teams together to go from big problems to customer-tested solutions. But this book shows you how to use AI to boost creativity, reduce cognitive bias and move even faster.
Plus I introduce the Super Solo Sprint, ideal as a sense-check for facilitators to use prior to a full sprint, and also for solopreneurs and small teams who want to get in on the design sprint action whilst still benefitting from the ideas and contributions of a full team.
Here’s your rundown of the recent main events and new tools in AI that product people need to know about.
Greg Brockman, co-founder of OpenAI, gave a Ted Talk.
There was lots of tantalising talk about plug-ins (which very few people actually have yet). The very first thing presented during the talk was how to automate the whole workflow of tasks that requires using multiple tools at once.
The workflow presented consisted of:
ChatGPT suggesting post-TED meal
DALL-E visualizing it
Retrieval plugin for saving the chat for future retrieval
Ordering the ingredients with Instacart
Posting about it on Twitter with Zapier.
He also presented how to use the “Code Interpreter” plugin in order to suggest and automate the data analysis process.
A cool feature of the Code Interpreter plugin is that it’s able to parse files in CSV format, which makes it really useful for Data Scientists, Product Managers and Marketers who want to speed up their data analysis.
This workflow consisted of:
Parsing and inferring the ontology of columns in a CSV file.
Proposing charts for data analysis: histograms, trendlines, word clouds.
Modifying the charts if asked to.
Amazon launched AI Bedrock, their new cloud service for AI-powered text and image generation.
It’s a Foundation Model API Service that provides an easy way to build and scale generative AI applications with foundation models (FMs). It allows you to choose FMs from AI21 Labs, Anthropic, Stability AI, and Amazon to find the right FM for your use case. You can privately customize FMs using your organization’s data too.
An “AI Agent” called Auto-GPT has been the talk of the AI town.
The premise of AI Agents is you give them a goal and they figure out how to accomplish it, then execute on those actions. They can interact with other apps, and the web. Here’s a 5-minute audio summary I created (using Eleven Labs to clone my own voice, because, why not?):
Auto-GPT is still super early stage, but if you’d like to try it out without going through the hassle of installing it, try out Godmode, a new site that allows anyone to use Auto-GPT or BabyAGI (another AI Agent) directly on the web. My early take on Auto-GPT is that it’s particularly good with doing research-related tasks. Lots more to come in this space as the tech progresses.
Atlassian Intelligence is launched
If you use JIRA, you’ll be happy to know that Atlassian is launching its new Intelligence tool. Sign up on the link to get early access. It’s a new virtual teammate that provides insights and recommendations to accelerate work. It leverages AI through internal models and its collaboration with OpenAI.
Bloop.ai - semantic code-search engine
GPT-4 powered natural language search for finding code and asking “What if?”-type questions about it.
Superus - Map out your ideas with an AI co-pilot
AI-powered interactive visual storytelling for enhanced comprehension.
Lookup - AI-powered customer feedback analysis
Helps product teams unify customer feedback in one place, extract insights automatically with AI, and deliver better products backed by real customer needs.
Vote to let me know if this newsletter is one you really value (or not!). And feel free to hit reply and let me know any thoughts or questions - I read and reply to every email.
Bye for now,
Martin