The speed at which artificial intelligence (AI) is moving makes tapping into it feel as daunting as trying to go for a swim in whitewater. But I’ve taken the plunge and have returned with some tips and tools for how to begin navigating the raging river of AI.
Am I “the expert”? Not even close. I just want to share some lessons that I’ve learned in the hopes that they help you brave the water too. Because at the end of the day, I believe the best way forward for a bright future with AI is to play, fail, and iterate, together.
On one rather average Monday in November 2021, I opened up my e-mails and did as I normally would: I checked the unread, decided which were urgent, and moved the rest to “Later.”
However, I clicked on one e-mail that mentioned something called OpenAI. Intrigued, I read it, landed on OpenAI’s just-released playground, and that’s the moment things changed forever. I’m not even exaggerating.
Cut to today, and I couldn’t imagine my daily life without AI: I use it to generate recipes, for work, for building workout routines, for thinking through complex problems… I even have a WhatsApp bot connected to ChatGPT that I recently used to give me a guided tour of a city I was visiting in Italy. Like, it’s wild.
My stance on AI is that it’s not AI that will replace humans; it’s humans who use AI that will replace humans who don’t.
So, aligned to that, I’ve learned a bunch of principles from building with AI that I wanted to unpack and share here; because at the heart of successful AI progress is sharing knowledge widely, freely, and kindly.
Quick story on my background
Context always helps, so here’s a little about me (if you really couldn’t care, the principles start below this bit):
My background is in journalism, and I currently work as a graphic and brand designer, illustrator, and hobbyist maker. I know some programming (mainly HTML, CSS, and Ruby), but I mainly use it to design websites and do CSS styling.
At the intersection of my passion for creativity, storytelling, and the transformative power of tech, specifically the potential of AI, I co-founded a non-profit AI community called Gamebreakers. Their mission is to create a space for learning and sharing in AI and equipping people with whatever they need to make a real impact using AI. Within that, I ran public AI art events under Artstronauts, and used to work in a company called recombinary that helps people integrate AI into their businesses and lives.
Basically, this means that I work with AI almost every day.
As a graphic designer, I really want to embrace AI to redefine what it means to be creative in general. And beyond that, what it means to be human.
Principles I’ve learned that help me build with AI
#1 Embrace your inner-n00b
The challenge:
Imposter syndrome and comparing yourself to others can stifle creativity and progress in AI.
This principle helps because:
This is a unique time in history where everyone is still learning, and there is no right or wrong way to build with AI.
We’re all cave people when it comes to AI, so learning to embrace your mistakes and failures and seeing them as opportunities for growth is now more acceptable and more important than ever. We need to see where it breaks in order to know where the risks and the opportunities lie. Don’t let a lack of knowledge hold you back.
#2 Learn together
The challenge:
AI is vast and complex, making it difficult to navigate and learn alone.
This principle helps because:
Finding a community to learn with can bring diverse ideas and perspectives to the table, leading to more powerful and innovative AI solutions. Collaborating can also speed up the learning process and help you build faster.
(For somewhere to start, we’re doing early access at Gamebreakers.com - come say hi!)
#3 Start small
The challenge:
AI can be overwhelming, leading to a fear of starting or the idea that you need to start with large, complex projects in order to “do all the cool things.”
This principle helps because:
Building something simple and small can get your creative juices flowing and help you get comfortable with AI. Plus, starting small allows for experimentation and iteration, leading to stronger final products.
One of my favourite “start small” projects was to build an AI assistant that helped me come up with creative recipes. I created input fields for the weather, my mood, my energy level, and my preferred diet and sent that along with a simple instruction prompt (e.g. “Act like a Michelin chef and use the following inputs to create a unique, delicious recipe for me. Also give it a witty title using puns and humour.”).
This helped me learn simple principles, build up an understanding of prompt design, create something I actually use often, and get a quick win without too much effort!
#4 Build in public & seek feedback
The challenge:
The complexity of AI can make it easy for harmful biases and malicious actors to operate in the shadows.
This principle helps because:
Building in public and seeking feedback can increase transparency and accountability, leading to more responsible and ethical AI. Sharing knowledge and uplifting the community benefits all of humanity and helps improve the impact of AI overall.
It’s also a super useful hack for getting lots of feedback and ideas. Stack Overflow is a great place to learn because of how much people share. There are often many ways of doing the same thing, and seeing how other people solve the same problem can be incredibly valuable (especially in AI, where everyone is still learning).
#5 Ship fast, think slow
The challenge:
AI moves quickly and can leave those who wait behind. But the wake AI leaves has a lot of potential damage in it.
This principle helps because:
The faster you ship, the faster you can learn, iterate, and improve. In the fast-paced world of AI, speed is crucial to staying ahead and creating impactful solutions.
But the counter to that is to think slow. Moving too fast can be like throwing all your money on a horse because it looked fast for two seconds. Speed breeds irrationality, and **we can’t afford irrational decisions being made in AI. **
This video highlighting some of Sam Altman’s points captures why this is important really well, but the TL;DR concerns are: Everyday people have the most to lose (so they should be involved at every step of AI’s development), and it’s very easy for bad actors to operate invisibly in AI.
Tools and resources I use to build with AI
Replit
The challenge:
Quickly shipping AI tools to test
This tool helps because:
It’s free to set up and play with, It’s super easy to share custom web apps via a URL, and there are already lots of cool AI Repls out there that you can fork.
Here are a few I enjoyed that you can get started with:
- GPT-4 Chat UI: Build your own GPT-4 frontend with this open-source Next.js template
- WhatsApp 🤝 ChatGPT with langchain: A simple integration that lets you talk to ChatGPT using WhatsApp.
- YouTube videos AI question answerer: Use 🤖 GPT, 🧮 Chroma and ⛓ langchain to get factual answers about any YouTube video you want.
Playground Beta – or even GPT API
The challenge:
ChatGPT free is often at capacity
This tool helps because:
You are less likely to be locked out due to capacity like ChatGPT. The playground doesn’t have GPT-4 unless you’ve successfully signed up for beta access, but 3.5 turbo is still powerful.
You also get a little more flexibility in responses since it’s not structured as a chat. You can also switch to chat via the model setting on the right-hand side of the window.
Pro tip: Skip the web UI altogether and simply plug into the API! If you’re already Replit, why not create your own UI and just call the OpenAI API? ¯_(ツ)_/¯ Using the API is also cheaper. The Reps linked above have this templated, so you don’t need to start from scratch.
The Brink AI
The challenge:
SO MUCH HAPPENS IN AI EVERY DAY?!
This tool helps because:
Brink is a great newsletter for the biggest news in AI. It’s an easy way to stay up to date with a broad sweep of the biggest events, newest tools, etc., without being completely overwhelmed by news sites.
Often, I follow one or two interesting threads I get from the newsletter online. It’s an easy starting point for me and makes the flood of info online suddenly seem less scary.
TheresAnAIForThat.com
The challenge:
Finding tools for AI that you can use, adapt and share with others
This tool helps because:
This website is a single, regularly updated database for all the main tools being shipped in AI. I find it a great way to have a laugh (there are some WILD apps) and also for inspiration for ideas of things I could build: I might think a project management tool could be valuable, so I go onto this site and see what’s already been done and what features might be missing from them that I can focus on instead.
Conclusion
This will sound like a bit of a cliché, but I still believe that the best AI tool will always be your own curiosity.
The potential opportunity of AI is undoubtedly there, if not hidden behind lots of unknowns, risks, and uncertainty. But I truly think that the best way through all of that… is through it.
If nothing else, go and watch a 5-minute video on AI, read that AI newsletter that’s been sitting in your inbox for weeks, or ask the person next to you what they’ve used AI for. This is a pivotal moment in human history, and being part of it is actually kinda friggin’ cool!
You don’t have to become an expert in AI tomorrow, but just be curious and have fun. That’s all it takes.
Jomiro Eming is a freelance graphic designer based in Cape Town (but works remotely as much as possible). He founded his design studio, Lük Güd Studio, co-founded the AI community Gamebreakers, and also used to work for AI-based company recombinary. Jomiro incorporates AI into his daily life as much as he can, and has made it his mission to blow as many minds with AI as possible. You can read more about freelancing, AI, and everything in between on Jomiro’s substack.
Read More
- Exploring Artificial Intelligence: Understanding the Basics and Its Applications in 2023
- Exploring Artificial Intelligence: Tips to Pivot Into the Industry
- How Developers Use AI to Ramp up in Their Role
- How I Collaborate with AI to Become a More Efficient Developer
