The journey of artificial intelligence (AI) began in the 1950s with the pioneering work of Alan Turing, who proposed the Turing Test to determine if a machine could mimic human intelligence. While the technology continued to evolve (however quietly) over the next several decades, it went mainstream in late 2022 when OpenAI introduced ChatGPT. Less than a year later, AI-powered tools permeate our daily lives – making many aspects of how we work and play “better, faster, stronger.” (Thank you, Daft Punk.)
At Revelry, our software developers, designers, project managers and support staff are all about exploring AI, so we asked them to share their favorite AI hacks. We’ve shared a few below.
- Regexes are usually complex, always arcane and exponentially more complicated the longer they get. Let the machine do the hard word of translation: `/use an LLM/i`! – Stu, Engineering
- I love Plantjammer for meal planning. I just tell it what ingredients I have on hand and it suggests recipes that don’t require another run to the grocery store. In the office, I like MidJourney for creating images to accompany our blog posts (like this one). – Laura, Marketing
- “Please review the following code in the style of José Valim, creator of Elixir. If appropriate, provide suggestions to improve readability.” It’s all you need. – Stuart, Engineering
- ChatGPT can help you find names for modules in your application. “I’m working on an application involving X and Y and need a word that means Z interaction between them.” Sometimes there is no good word for the concept and at least you can stop thinking about it. – Justin, Engineering
- Don’t Google for ideas on what to do during vacation. Ask ChatGPT to be the best tour guide ever focused on your area of interest, and let it come up with initial options. – Max, Partner Development
- PromptPerfect is an incredibly effective plugin for GPT-4 that delivers on its namesake promise. – Heather, Strategy
- My family uses YNAB (You Need A Budget) as our budgeting tool, and a lot of the initial setup involves categorizing various purchases and determining how much money to allocate to each category per month. By giving ChatGPT a list of categories and a general monthly income, it was able to create category groups that fit in with our intentions and aspirations for our money. – Dan M., Engineering
- You can ask ChatGPT to apply models of thought to help influence how it responds. “I want to do X. Apply the ABC model by AUTHOR to explain the THING YOU WANT TO KNOW. – Max, Partner Development
- Help transforming data. “Given I have a CURL command that returns JSON like [example json] and I want to abstract a piece of information out [insert specific info here] into another specific format, how might I do that using unix commands” and then refine it from there based on what I am trying to accomplish. It’s useful for using the built in command line tools that are powerful if you know how to leverage them. – Nick, Engineering
- To run a large language model (LLM) on your own computer, try the Dalai project (https://github.com/cocktailpeanut/dalai), which makes it easy to run LLaMA and Alpaca thanks to another project called llama.cpp, which drastically reduces the hardware requirements needed to run them. It also has an integrated web server and user interface, and it’s fun to watch the console as you interact with the models to see everything that’s happening behind the scenes! Alpaca is a fine-tuned version of LLaMA, and actually works reasonably well. – Eric, Engineering
Want to chat about AI (or any innovation topic)? Connect with a member of our product development team. We love this stuff!