All Hands on Data #84
Welcome to the last All Hands on Data for January! Check out our team's favorite articles of the week below.
Trick prompts ChatGPT to leak private data
In perhaps more of a "don't do this" article, the author discusses Google's recent discovery that it can prompt ChatGPT to disclose sensitive data. What I take from this is if you're using ChatGPT in your product, make sure that you're doing some checking and sanitizing of what your users are sending to your app and to OpenAI. - John Forstmeier
AI and Data Science How to Coexist and Thrive in 2024
Sukhadeve talks about how the "collaboration between AI and Data Science [has emerged] as a transformative force" in the industry. The interplay between them in the upcoming year is trending in a few directions but the ones I found most intriguing: AI-generated synthetic data for training models, AI-powered cybersecurity tools, and the heightened focus on data regulation. - Katt Baum
The integral role of data science in navigating deepfakes
We've all seen it - Tom Cruise telling a story, the President congratulating you personally, even some of our favorite TV show cartoon characters are covering pop hits. Deep fakes have arisen exponentially, and have gotten better and better as time has gone by, making it almost impossible to tell what is real and what's not. This article goes over the importance of data science in research and development to develop detection methods in subtle inconsistencies and more. As the technology continues to grow, so does the importance of deepfake detection. - Reed Cowan
Slashing Data Transfer Costs in AWS by 99%
One of the most expensive parts of data work is trying to manage egress costs between systems. For those of us on AWS, every 1TB costs $90. But what if it could cost only $0.08? Daniel walks through a deep dive on a hypothesis to reduce egress costs drastically. It won't work for every situation, but when it does? chef's kiss - Blake Burch
SQL for Google Sheets with DuckDB
I love a good succinct article. This one really hits the mark. Our team at Shipyard announced our partnership with DuckDB earlier this month. I was able to spend time with the product and see why people love it. The biggest thing I took away was the simplicity of using DuckDB. This article shows that with showing how you can load in Google Sheets. - Steven Johnson
How Data Engineers Should Prepare for an AI World
With all the take of AI in the data space, there's been an overwhelming response that AI is coming after data jobs. When in reality, AI is only going to make data more accessible and more easy to manage, letting data practitioners focus on the fun projects as opposed to data administration. - Angel Catalan