Fearmongers would have you believe that AI will replace us, but I truly believe there has never been a better time in history to be a developer.
With the advance of artificial intelligence into generative AI (GenAI) and enhanced computing power, we stand on the brink of a new era of development, whether it’s software, apps or models we’re developing.
More importantly, this is an era where we can unleash what I consider our number one skill. I’m not talking about coding, debugging or writing documentation.
I am talking about our creativity.
Developers are naturally some of the most creative people in the world. With artificial intelligence as an ally, you can experiment faster and be more productive.
It used to be that productivity was about having the best Vi Improved (VIM) commands or Editing MACroS (EMACS), but we’ve come a long way since the late ‘90s.
Now, we do our computing in the cloud and our editing in a browser, and poof, our terminals are gone. And yes, that hurt, but look where we’ve gotten.
With multimodal AI now on the scene, productivity accelerates data and AI life cycles to put models into production so we can make better decisions. We write the code that accelerates that life cycle and build the models supporting better decisions.
And for anyone who didn’t know by now, our organizations are more productive when developers are more productive.
What if there were a technology that could help you become both more productive and more creative?
Generative AI, anyone?
I know you’re already hearing about GenAI everywhere and skeptics would have you believe it’s overhyped or potentially even a creativity killer, but I’m here to convince you otherwise.
Here are four ways I’m seeing GenAI improve the developer experience today:
- Data exploration
- Modeling suggestions
- Brainstorming
- Data generation
Data exploration with GenAI
Imagine having a copilot for developers feeding an endless stream of ideas – that's the vision behind SAS® Viya® Copilot.
Here’s an example: Let’s assume I have air quality data from which I want to gain insights. I ask my copilot to show me all the different air quality data sets and briefly describe the data in each one.
My copilot says I can see air quality data daily and monthly, so I decided to play with the monthly data. The assistant gives me suggestions that fit the data and automatically generates buttons to execute different types of analyses.
I don’t have to think about or code data preprocessing approaches. It’s done for me, so I can have more time for important matters.
Now that I see my options, I start with a geographical analysis. SAS Viya Copilot executes it, plots the average air quality index across different areas and explains what I’m looking at on the map. I decided to analyze air quality trends and patterns over time and asked for a trend analysis. Right away, it generates insights grouping the data into regular cycles, seasonal variations and unexpected changes.
Some kind of developer dream, you might think? These kinds of copilots are on the way from SAS.
Building models with GenAI
You know how time-consuming building models can be. But what if you were working with someone who gave you model ideas you might never have considered? SAS Viya Copilot accounts for that, too. Just push the button for the most promising model – time series, machine learning, optimization, etc. – and the copilot can train it with your data and give you the results.
Once you choose and execute your model, you might as well show off your work. With some help from the SAS business intelligence tool Visual Analytics, you can build a C-suite-ready dashboard.
Want to look behind the scenes at the code being generated? SAS Viya Copilot is fully transparent: human in the loop, no black boxes. You can have the copilot send your coding interface to SAS Studio, where you’ll see the code and have the chance to accelerate different tasks with it. From there, you can add comments, fix bugs, convert Python code into SAS code or vice versa, a world of possibilities.
Brainstorming with GenAI
My favorite button is brainstorm. If I run out of ideas or strategies, I push the brainstorm button and SAS Viya Copilot recommends even more data possibilities.
Whether you’re a developer, data scientist, or business analyst – in banking, health care, insurance, or any other industry – you can build a copilot to be more productive and solve real-world problems. Just give it a name, a description, and a few starters and upload your data. Your creativity is the only limit.
Not only is SAS Viya Copilot fast, precise and flexible, but it is also large language model (LLM) agnostic and secure because it’s running on your Viya server. It also speaks and works in many languages, including English, Italian, Chinese, German, etc.
Generating data with GenAI
There is no good AI without good data. And sometimes real-world data isn’t good enough.
No matter how much you may wish it weren’t so, sometimes data can show bias. Or it may have sensitive information. Or maybe you simply don’t have enough data.
Let’s say a financial institution wants to detect fraud. The number of fraudulent cases is extremely small, so traditional modeling approaches struggle to effectively train models from what’s available.
Synthetic data may be the answer for all of these examples. That’s why I'm excited about a new SAS product being tested: SAS Data Maker.
Upload your existing data and Data Maker generates high-quality synthetic data in a completely transparent, low-code, no-code way. Data quality and privacy are automatically audited, and you can compare relevant variables between the original and newly generated data.
No more manual data collection or buying costly third-party data sets, so you save time and money. No more showing potentially sensitive information, so you can safeguard privacy.
I don’t know about you, but I am ALL IN on creativity and productivity. I don’t see GenAI as a threat or a danger, but as a tool or partner to make my job even more fun and productive.
