Tag: AI

  • Getting More Time from AI

    As I get older, I find that time is the most valuable commodity I have. It’s the one thing that I need more of, but I can’t get it. Even if I find more efficient ways of doing things, it seems there’s an endless list of things at work and on the ranch that need to be done.

    It seems to be the same for most of my friends in other jobs, whether in medicine, law, or any other position; they’re often overloaded with more work than they can get done in a week. Arguably, it’s not all important work, as sometimes we might tackle a task, only to have our boss throw the work away or delay the project. That’s annoying, but I also understand priorities change.

    Mostly, I understand people ask for a lot of things they don’t really need, but because it’s easy to ask. Over time, they may forget about their request or realize they don’t need it anymore. So I tend to do the work I’m asked and then not worry about if it’s really needed.

    However, that’s not what this piece is about. Let’s say that you have a useful AI Assistant that can truly produce better code, faster than you can. You trust the coding agent (under your guidance), and it saves you time. Let’s say that this agent reduces your coding load by 30%. Forget the issues with the agent stealing your joy.

    A number of studies show that the average developer only codes around 2-3 hours a day. Some might say 4, but that’s just 50% of your workday (I hope). If you save 30% of that time, then you’re saving 1:20 a day, 6:40 a week. That’s a nice chunk of change, but what will you do with the time? Or maybe more importantly, what will your boss expect you to do?

    A lot of management might see 30% and expect you to get 30% more coding done (of 40 hours), asking for 12 hours more of work. I see that a lot as management somehow things that developers work 8 hours a day and should get 8 hours worth of code written. When that doesn’t happen, and it never does, they often schedule more meetings to check on your progress. Hence to 2-3 hours worth of actual work.

    The smarter group of managers might think that you could get 5 hours more work done (30% of 3/day), and the really smart group of managers might use that time to improve the coding skills of the staff. That’s more than a 10% allocation of time and if focused, could help developers write more efficient code (especially SQL), improve their secure coding knowledge, and perhaps even tackle some technical debt with refactoring of existing problem areas. Maybe you could learn to better judge AI-produced code.

    I know that many people are skeptical of AI assistance in coding. I think it’s a tricky thing, but even if the AI doesn’t produce production code, but allows me to conduct a lot of quick experiments that can prove out a hypothesis, I think it can help developers become more productive.

    If they learn to use it as a tool and management doesn’t expect it will turn every developer into a 10x (or even 2x) engineer.

    If you had more time, what would you do in your work day?

    Steve Jones

    Listen to the podcast at Libsyn, Spotify, or iTunes.

    Note, podcasts are only available for a limited time online.

  • A Prompt AI Experiment

    Prompt AI released recently and I decided to try a few things with the tool that might help me in database work. I’ve had to do this task, but I had a customer recently ask me about this as well. They were wondering where they were short FKs.

    This post showed what I tried with Prompt AI.

    This is part of a series of experiments with AI systems. This is part of a series of posts on SQL Prompt. You can see all my posts on SQL Prompt under that tag.

     

    Help Me Query the Database

    In a recent post, I looked at the MSSQL agent in VS Code to query my database. I decided to try a few things with Prompt AI, working inline. First, I connected to the database and then I asked a question of the AI: who hit the most home runs.

    2025-09_0342

    From the time I shot the video, I had updated Prompt, and possibly the AI got smarter, because this was the result: I had a qood query.

    2025-09_0343

    I wanted more, so I asked for seasons. You can see the prompt at the bottom, but seasons were added.

    2025-09_0344

    Click Accept and then run it and I get the correct answer.

    2025-09_0345

    How nice would this be for some of the BI analysts to use this to get work done?

    Getting Help Fixing the Database

    In this db, there aren’t any FKs. I decided to (poorly) ask about this.

    2025-09_0347

    Note the misspelling. One of the amazing things is the AI recognizes little errors, just as a human would, and corrects them. I get an OK result, but not great.

    2025-09_0349

    I rewrite my prompt to ask which ones “should” have FKs. I get this:

    2025-09_0350

    Helpful, but I don’t want to work through this list.

    2025-09_0351

    I get this. No code, and I feel like I’m talking to a junior dev who is either trying to avoid work or doesn’t have the understanding that I want work done. I’ll try again.

    2025-09_0354

    I get something I can use. I wouldn’t likely create all these at once, I’d test things, I might start to add these as migration scripts I can deploy over time, after I’ve verified this isn’t going to break an app or cause issues. Or there aren’t already data issues.

    2025-09_0355

    Still, this is one  of those nice time savings for me writing a script to try and create these or working through time one by one. Now I can just set up tests.

    Or ask Prompt AI to help Winking smile

    Summary

    The new features in Prompt AI let you keep coding inline, but add some AI help to what you do. I’ve seen some neat things, and some frustrating ones, but overall, it’s helpful to produce code.

    If you haven’t tried SQL Prompt, download the eval and give it a try. I think you’ll find this is one of the best tools to increase your productivity writing SQL.

    Video Walkthrough

    This isn’t exactly what is above, because LLMs aren’t deterministic.

     

  • AI Steals Joy

    Take a moment and think about how much work you do each day. If you write code, how much time do you actually spend writing code around meetings, admin work, and other tasks. If you manage systems and develop scripts, how often do you get to work on a script, change it, upgrade it, or add a new feature? I bet it’s a minority of time for most of you. Studies show that many developers spend only an hour or two writing code each day. I would guess it’s lower for sysadmins or DBAs who incorporate coding into their jobs.

    Now think about how much you enjoy each thing you do. I assume meetings aren’t the most exciting thing you work on, nor is updating a document for others to read. When I wrote code as the main part of my job, solving the problem, trying different algorithms, experimenting with enhancements, those were the fun times.

    Now imagine having a competent AI assistant that works with you. What if the AI does the coding and leaves you with the meetings and writing of specifications? It could happen. While it might be fun to direct a team of agents, you still need to understand software architecture and development, but you might not want to get away from writing code.

    As more developers start to use AI agents to get work done, there is rising dissatisfaction among developers because the AI can do quite a bit of the coding, which is the thing that most developers love. The AI doesn’t help with any of the other busy work, though AI agents might do a better job of ensuring docs match committed code (or released code).

    What I’d really like is an AI that would take my place at meetings. Maybe do a report for me or take a summary back, get some action items for me. An AI agent could handle things that wouldn’t necessarily be fun for me. Not all tedious things, but it could free up more time for coding.

    However, that’s not what the GenAI agents are being positioned to do. They’re stealing the joyous part of development. At least, if managers have their way, that’s the place they’ll use GenAI most.

    Steve Jones

    Listen to the podcast at Libsyn, Spotify, or iTunes.

    Note, podcasts are only available for a limited time online.

  • T-SQL Tuesday #190–Mastering a New Technical Skill

    It’s time for T-SQL Tuesday again and this time Todd Kleinhans has a great invitation that is near and dear to my heart: mastering a new or existing technical skill. That’s been a lot of what I try to inspire people to do at SQL Server Central.

    Make a plan and start learning. And respond to Todd’s invitation and write down your plan and share it. Start a blog, use Linked In, whatever. Spread the word on socials as well.

    If you want to host, I’m always looking for hosts for T-SQL Tuesday. Ping me on Twitter/X, BlueSky, or LinkedIn.

    Mastering a New Tech Skill

    Like Todd, I’m interested in AI and I think it will dramatically change the world in the coming future. I also think it’s a bit of a technical skill that is important to learn. I wrote about this a bit in last month’s post.

    How do I work with a GenAI model and improve my technical skills? The easy answer is more and more. I’ve been having more conversations with Claude, usually looking for ways to help me solve a problem or write code, and then ask the GenAI to explain things.

    However, I wrote awhile ago about an experiment in helping someone else learn something: Can an AI Help Me Find a Job?. For me, I’ve been looking a bit more at DataBricks, as I hear this from clients all the time. I wanted to gain some skill here, so I decided to ask Claude to help me.

    2025-09_0225

    I got a good outline of things to do across a few months. I then asked for references and got some:

    2025-09_0226

    The next stage for me is to start embarking on this journey a few nights a week and learn some things that might help me both in my job, and potentially in a future position if I need one.

    My Complete Outline from Claude

    Here’s the end result, with links.

    Databricks Learning Outline with Resources

    Phase 1: Foundations (1-2 weeks)

    Understanding the Basics

    • What is Databricks and why it’s used
    • Core concepts: clusters, notebooks, workspaces, and Apache Spark
    • Databricks architecture and modern data stack integration

    Key Resources:

    Phase 2: Getting Started (2-3 weeks)

    Hands-on Basics

    • Navigating workspace interface
    • Creating and managing clusters
    • Working with notebooks
    • Basic data import methods

    Key Resources:

    Phase 3: Data Analysis Fundamentals (3-4 weeks)

    Core Analytics Skills

    • Data exploration and cleaning
    • Working with different data formats
    • Basic SQL and PySpark operations

    Key Resources:

    Phase 4: Intermediate Techniques (4-5 weeks)

    Advanced Analytics

    • Complex transformations and ETL
    • Streaming data and machine learning
    • Performance optimization

    Key Resources:

    Phase 5: Production and Best Practices (2-3 weeks)

    Professional Development

    • Job scheduling and workflows
    • Security and monitoring
    • Integration patterns

    Key Resources:

    Additional Learning Platforms:
    Certification Path:
    Community and Support:
    Getting Started Steps:
    1. Sign up for Databricks Free Edition: Visit the Databricks Free Edition signup page and pick your preferred signup method Databricks Free Edition | Databricks on AWS
    2. Access free training: Free unlimited access to training content on Databricks Academy provides on-demand learning experiences for all skill levels Databricks Launches Free Edition and Announces $100 Million Investment to Develop the Next Generation of Data and AI Talent – Databricks
    3. Join the community: Connect with other learners and experts
    4. Start with sample datasets: Practice with built-in datasets before using your own data

    The great news is that all self-paced training across AI, data engineering, and more is now free for learners Databricks Training & Certification Programs | Databricks, making it easier than ever to get started with Databricks!