We’re at the dawn of a new, exciting, confusing era of software. This wave of generative AI models can both delight, and disappoint. There are near-infinite possibilities, and sharply-constrained realities. It’s clear AI is going to change the way we work, but it’s hard to tell exactly how.
What’s going to happen in data and analytics?
It definitely isn’t the end of data teams, or data practitioners. These models feel like magic… but they’re far from being able to replace human insight. AI can’t fully understand business context, or explain its answers, or even reliably not make things up – all important features in a data practitioner!
Perhaps someday we’ll have products where a Business User™ can submit a natural-language question, and get a reliably-accurate, fully-reproducible, context-aware report with built-out pipelines, properly-formatted charts, and well-reasoned explanations that they fully trust and take action on… but that doesn’t seem super imminent? We’ll see.
At Hex, we’re focused on human-computer symbiosis: how we and our machines can partner together, each of us doing the things our respective neural nets are uniquely well-suited to.
After all, Data work is a creative endeavor, well-suited to human ingenuity. Refining raw data into insights and models requires exploring ideas, building narratives, and taking risks. It’s an art, and a science, and not something AI will learn overnight.
Data work can also be really tedious. It requires writing boilerplate, fixing dependencies, and hunting for misplaced parentheses. Humans can get good at this kind of thing, but it’s not a great use of our time; in fact, it would be awesome if our software could do it for us, so we can spend our brainwaves on the valuable, fun, creative stuff.
Today we’re launching a private beta of Hex Magic: a suite of powerful AI features meant to augment – not replace! – data people. Now, humans can do what they do best, and let the computer do the rest.
Let’s start by writing some new SQL or Python code:
Oops! An error. Let’s see if Hex can help:
Hex can even explain and document your code:
Hex ’s Magic features do more than just pass a prompt – they know about database schemas, past operations, and the project’s execution graph, so they can make deeper, more insightful recommendations.
You can see more – and sign up for the waitlist – over here.
Our early Magic users (magicians?!) tell us this lets them focus on the things humans are uniquely great at: asking questions, working together, and telling stories.
First, it’s definitely not “instant insights in a box”. These models don’t know the right questions to ask, or the how to present the answers – that’s your job! And we think it should stay that way.
This also isn’t going to suddenly democratize advanced data work. Magic can make it easier to do certain things – but you still need to know what you’re doing. The underlying models make errors, both obvious and subtle, so it’s important you understand the code you’re executing. We designed our Magic features to explicitly keep a human-in-the-loop, and, at least for now, every code run still has to be manually triggered by a person.
Will Magic replace some human tasks? Yes! But the goal is forming a partnership, where each party can focus on the things they truly excel at.
We are launching a beta so we can get feedback from more humans before opening this up to everyone. If you’re interested in trying this yourself, you can sign up on the waitlist on our product page, which also has an FAQ and some funny puns.
This is just the start, and we’re already hard at work on a next set of features to accelerate and augment human insight. If you’re interested in coming along on this journey with us, please get in touch: we’re hiring for ML and software engineers, designers, and other roles here.