Exploring the Vision of AI with Jared Zhao, Founder of AskEdith
Written on
Introduction to Jared Zhao and AskEdith.AI
After penning an article about one of Jared Zhao's innovations, OnceUponAITime.com, I connected with him through Twitter. Today, I’m excited to share insights about Jared, his key project AskEdith.AI, and an interview we conducted with this AI entrepreneur.
Jared pursued computer science at the University of California, Berkeley, and has contributed to data science research at both Lawrence Berkeley National Lab and RISELab. Currently, his primary focus is on AskEdith.AI, a tool designed to help users extract specific information from large datasets using simple English queries. We will explore the capabilities of AskEdith in detail shortly, so stay tuned to discover what this innovative tool can accomplish.
I personally tested AskEdith and utilized its AI features to analyze my statistics on Medium.
The Interview with Jared Zhao
Q: How did you conceive the idea for AskEdith.AI?
A: Before launching AskEdith, our team was developing a data infrastructure product aimed at non-technical users. One of our biggest hurdles was teaching users how to navigate it. We continuously sought ways to make this easier for them and envisioned an English-to-insight feature. We soon realized that this concept could stand alone, leading us to create AskEdith!
Q: What inspired the name of your product?
A: The name E.D.I.T.H. is inspired by the AI assistant glasses that Tony Stark left for Peter Parker in the Marvel Universe. We felt this name added a touch of personality to our product and made its purpose clearer to users.
Q: Is AskEdith built on GPT-3 technology?
A: Yes, GPT-3 is indeed a key component. We also leverage the BERT semantic search model and various NLP/ML techniques to enhance the capabilities of raw GPT-3.
Q: Can you give a real-life example of how someone might use AskEdith?
A: Picture a marketer at a large corporation where all marketing data resides in a central data warehouse like Snowflake. Currently, accessing this data for analysis on a recent A/B test requires a request to the data analytics team, with a wait time of 2–5 business days. With AskEdith, users can retrieve that data directly.
Q: What challenges did you face in developing and launching AskEdith?
A: Our main challenge has been ensuring AskEdith is user-friendly. While we are dedicating significant effort to enhance the English-to-SQL functionality—which is already performing well—our current focus is on releasing a Chrome extension that will allow users to utilize AskEdith alongside platforms like Retool.
Q: How do you envision the future of AI? Will it influence all areas of life? Will certain jobs be replaced by AI?
A: AI will undoubtedly influence many aspects of life, but I don't believe it will replace jobs anytime soon. Most professions require comprehensive critical thinking that AI isn't equipped to manage yet. Rather, I see AI augmenting roles, particularly in areas facing skill shortages like data analytics.
Q: Do you think that in five years, we will interact with computers using only natural language?
A: Absolutely! The key will be the widespread adoption of natural language as an interface across various products and devices.
Q: What do you foresee as the next significant challenge for AI?
A: Language models have successfully tackled language-related tasks, such as content generation. I believe the next frontier will involve connecting natural language with complex tasks that necessitate reasoning and planning. For example, instructing a robot to "get milk" would require it to navigate various steps and constraints of the human environment.
Q: Do you have plans for future projects?
A: Not at the moment! Our focus is solely on realizing our vision for English-driven analytics. There's still much to achieve with AskEdith. However, as we were inspired to create Once Upon AI Time spontaneously, you never know when the next idea might emerge!
I appreciate Jared for sharing his insights, and you can connect with him on Twitter. Now, let’s take a closer look at AskEdith!
Demonstrating the Capabilities of AskEdith
I have begun using this tool, and I’m eager to show you its functionalities. To start, users must provide the data they want AskEdith to analyze. There are several methods for this; you can directly connect existing databases to the tool.
The number of available connectors is continuously increasing. Alternatively, you can upload a CSV file, which is what I’ll do since I currently lack access to large databases. Nevertheless, I have discovered some intriguing use cases.
For my experiment, I will use a CSV file I obtained from a Chrome extension that tracks my Medium stats. Upon uploading the file, AskEdith immediately presents a list of sample questions to ask.
However, I want to explore something more specific.
Once I pose my question, AskEdith translates it into code:
And executes it!
The response is accurate! It indeed identifies my most popular post from July. Manually searching for such data would be nearly impossible, as I could only sort by views or date. Opening the CSV file in Excel to find this information would also be cumbersome, as it looks like this:
Being able to simply ask a question and receive an answer is truly impressive. It feels like a bit of AI magic. Having tried it, I can clearly see how beneficial a tool like this could be for organizations handling larger datasets. In many cases, employees could independently locate the data they need without relying on data analysts.
Explore AskEdith yourself:
AskEdith - Never write SQL from scratch again!
Connect business applications and draw intelligent insights in the blink of an eye. AskEdith empowers teams to make…
www.askedith.ai
Support My Writing on Medium
If you appreciated this read, consider supporting me and other writers on Medium. For just $5, you can gain unlimited access to all content.
Discover more about the journey of becoming a tech entrepreneur with Jared Zhao, as he shares his experiences and insights in this engaging interview.
In this enlightening talk, Sam Altman discusses the vast possibilities of AI, exploring its implications and future developments.