For those who know me, it’s no secret I’ve been obsessed with this technology for the past 4 years or so, when GPT-2 was released. But it wasn’t until the advent of GPT-3 and the chat models in particular, that Generative AI has become accessible for those with no coding skills – what’s more, I’d argue that it is those with no coding skills primed to both gain the most out of this technology and become among the first power users, but I’ll leave that discussion for another time.
Today, I’d like to share how I’ve been using GenAI in my daily routine, what limitations I’ve found, how to overcome some and a few sources I regularly use to keep myself updated.
- Summarization: the classic use case for the lazy reader or for those who want to maximize their time spent on reading articles in detail. There are tons of apps and extensions for this. Personally, I’m using Chrome extensions like YouTube Summary with ChatGPT & Claude, ChatGPT » summarize everything! have become part of my daily life, speeding up research so I can focus on the most relevant pieces. Another powerful tool is to combine this with Obsidian (an open source note-taking app) to quickly summarize and save content I’ve found online. I can then also leverage an Obsidian GPT plugin to retrieve the contents I’m looking for, though I haven’t found a plugin for this I would recommend yet.
- Writing: in most of my articles, I get some support from ChatGPT and a few ChatGPT-powered tools of my own devise through Langchain. However, I don’t expect it to do all the work, rather help me speedup the wording for content I’ve already outlined. I’ve also developed a set of prompts (chain of thoughts w/self-consistency and assisted step-by-step) to greatly speed up writing, without sacrificing my own writing style.
- SM Material: it’s really hard to come up with new and meaningful stuff to share over social media consistently. That’s why I put together a very simple Botpress-powered chatbot available through our Teams chat which users can leverage for content creation. Simply provide it the URL to a blog post and the bot will give you multiple ideas for content bits for different channels (LinkedIn, Facebook, Twitter, Instagram, etc.) all based on the content of the URL you provided.
- Code Assistant: I’m knowledgeable on Python and other languages just enough to be dangerous (both literally and figuratively 😊), but I’m no developer, so it’s safer to have someone do some safety checks. That’s why I’ve been using OpenAI’s Advanced Data Analysis (formerly Code Interpreter) to help me review code I write and provide suggestions to improve it.
- Online Research: if you’re below 40 years old and not doing online research every day at work or in private life, you live in a rare world where this article could not be less meaningful and you’re simply not reading this. For the rest of you here, I’ve learned to love tools like Perplexity.ai, which will give you (mostly) reputable sources and summaries for a particular topic you’re researching. It’s a fantastic starting point for any topic, specially one you’re completely unfamiliar with.
These are but a few use cases I personaly leverage daily. I'm also exploring many others, AI Agents in particular are showing a lot of promise, but the state of the art is not quiet easy to make productive yet
