A black and white figure's thought-hive
Natural language is hard!
This whole shebang about AI and AI-enhanced products has led me to one of the products that have become part of my daily usage: Perplexity.AI. It’s a search engine where you ask questions and it gets your answers from the internet.
Let me take you back a couple of decades before we go back to it.
How a simple concept gave rise to so much more
This year has witnessed a significant transformation in the world, all thanks to a singular technological advancement: improved text prediction.
It is not a joke nor an exaggeration. The blossoming of AI that we’ve seen can be traced back to OpenAI’s text completion model, GPT. It is not the first, nor the last, nor the best, but it made itself available to everyone in a way that allowed people to build over it. Applications quickly proliferated, finding new uses and capabilities that were previously unimaginable.
How is this possible?
It all stems from a concept known as “emergent behaviour.”
Postman with Enhanced JSON Visualizations
For a while (a few years) I’ve used Insomnia as my favorite REST Tool. I’ve also looked with good eyes at Hoppscotch, but it being browser-based means that you need to go through a lot of hurdles for local development (and I’ve just never had the patience).
But being that everyone still uses Postman, I decided to switch back. It’s not about just following the majority, but ensuring that there’s an active community for the tools that I use. Also, I initially was scared of Postman becoming too close to force to sell paid plans, but the years proved me wrong. Finally, all the extra features I needed eventually made it there.
Using OpenAI's for SEO optimization
I wanted an excuse to try something with OpenAI’s GPT-3 models, and while writing on this blog I realized that most of the pages could have a description autogenerated for them. This seemed like a good fit for a GitHub Action too, so here I explain how I set up my new project Jekyll SEO AI Action.
Taking notes of everything, but quickly
I have stopped reviewing tools as part of my blog for a long time, but I did want to point out the benefits of the second-brain approach of some applications. I have personally found my sweet spot with Obsidian, so I’d like to talk about that.
You shouldn't have to stop coding
I came across this question, which is a fairly common concern I have come across with people around me. Here’s the gist of it:
After I became a Tech Lead my work has become more like managing than anything else, I don’t do more code and just a little technical stuff
I’d like to offer an alternative approach.
Rendering Blender Scenes in AWS Lambda
I recently published TAS-PC, a producer-queue-consumer lambda terraform module. This was an intermediate step for creating a distributed rendering engine.
In this post I’d like to discuss some of the technical aspects of that work.
Terraform-AWS Serverless Producer-Consumer Module
I needed to setup some infrastructure in AWS but did not want to make modifications each time I needed them. What a better opportunity than this one to learn Terraform.
So I created a little module that sets up a small producer-consumer serverless (lambda-based) architecture, powered by a SQS queue.
Not currently supported out of the box
A few weeks ago I migrated worklogger to TypeScript and yarn 2.
But this broke some things along the way.
Not really but still a breakthrough
There’s an article published in TheNewStack.io that is titled Machine Learning Algorithm Sidesteps the scientific method. While I do consider this to be clickbait, I still consider some truth behind it worth discussing.