Thing is, the data gig isn’t what people think it is, including many practitioners. Having spent a bakers dozen years leading data teams, my take is that being really effective as a data leader is mostly about ensuring the data you’re ingesting is solid. But that has nothing to do with technology! It’s all about understanding business and systems processes and working with various stakeholders to improve data quality. Sure, your data pipelines and modeling need to be good, but that’s table stakes. Your highest value is to be an interlocutor between the business and engineering, and to keep devs honest in their concern for emitting good data. Remember, the SR-71 Blackbird never carried a gun. Yet it was the fastest jet plane ever built and was worth operating at a cost of $300 million per year. Why? Because it provided good information. Never underestimate our value.
I think the big mistake is companies rushing to throw this stuff in front of the end user - sometimes while telling their employees NOT to use it for fear of IP leakage or whatever. AI is really in trusted assistant mode at this point. The big problem is in changing habits and changing ideas. I had a hard time retraining myself to turn to an AI before Google and now I get exhausted - when I’m on a roll it’s like pair programming with a really enthusiastic partner, I’m tired before long.
Agree on the age difference but they have the same challenges. My 15 year old knows about ChatGPT and wanted me to show her how to use it, but it was the blank page problem - it’s one thing to solve a problem where all the parameters are given but what do you do when you can do anything?
What about as an entrepreneur? Like you have an idea for an app and conceptualize it and write it? I'm at the point where I have a couple of interns, great. I have a friend who's talking about creating a fleet of AI employees and his job as CEO is to direct them. Then you end up with agents that are managers - all they do is prompt their employees and report status upwards. What's scary is that at this point in my career, 30+ years in, I have the knowledge to do that, at least on the tech side. What if I could have a hundred employees with no payroll?
I worked through the early days of the dotcom boom and all the signs are here again, this is like 1996. I worked through the cloud revolution where all your infrastructure costs came down to your ability to pay your credit card bill. What I worry about this time around is a failure of imagination.
Start with their interests, that's always a good place to start. The more they use it, the more they can adapt it to their needs and preferences.
My daughter is currently using it to generate SAT questions, and to improve her coding skills. I've provided her with my own workflow and she's adapting it to her own needs.
I've shown her how to create custom GPTs, how to code prompts, I'm trying to teach her what I'm learning myself, at that abstract level.
What about as an entrepreneur? Like you have an idea for an app and conceptualize it and write it? I'm at the point where I have a couple of interns, great. I have a friend who's talking about creating a fleet of AI employees and his job as CEO is to direct them. Then you end up with agents that are managers - all they do is prompt their employees and report status upwards. What's scary is that at this point in my career, 30+ years in, I have the knowledge to do that, at least on the tech side. What if I could have a hundred employees with no payroll?
I worked through the early days of the dotcom boom and all the signs are here again, this is like 1996. What I worry about this time around is a failure of imagination.
"What I worry about this time around is a failure of imagination."
Agreed. I've seen some agentic workflows. They're amazing. I get the impression people are sleeping on this, dismissing them out of hand. I think that's a huge mistake.
I heard people say the Web was a fad too, back in the 1990s. That statement looks incredibly stupid now.
Can agree, if your not doing BI well, AI will be a challenge. AI will be much faster to market with a solid data foundation that is delivering BI day in and day out. Most of the business relies on BI platform to make decisions and understand the business, especially when it is intuitive and accurate. That data delivery maturity drives reliable, trusted data for AI usage. Turns months into weeks for delivery, and weeks into days for experiments.
Listening to Joe's most recent podcast - Bridging the gap between data and strategy - which speaks well to the need for a strategy - but would like to add - the companies data foundation has to go above the "data team" (or the "data teams" that are not talking to each as mentioned in the podcast) for a strategy. Data Strategy has to start from the creation of the data, the operational systems, the 3rd parties, the SAAS's perspective. We are too late if we are building a data strategy from a downstream perspective of a "data team". Enterprise Data Architecture Matters, data can't be treated like a byproduct, or a means to implement a system. The expectations should be that every application being implemented will have the need of BI and AI, to be plugged into your AI Engine. Which pivots to practical data modeling, once your planning on the operational systems, engineering and designing for BI and AI is part of your DNA. Talk and plan data needs at project inception.
Thing is, the data gig isn’t what people think it is, including many practitioners. Having spent a bakers dozen years leading data teams, my take is that being really effective as a data leader is mostly about ensuring the data you’re ingesting is solid. But that has nothing to do with technology! It’s all about understanding business and systems processes and working with various stakeholders to improve data quality. Sure, your data pipelines and modeling need to be good, but that’s table stakes. Your highest value is to be an interlocutor between the business and engineering, and to keep devs honest in their concern for emitting good data. Remember, the SR-71 Blackbird never carried a gun. Yet it was the fastest jet plane ever built and was worth operating at a cost of $300 million per year. Why? Because it provided good information. Never underestimate our value.
I think the big mistake is companies rushing to throw this stuff in front of the end user - sometimes while telling their employees NOT to use it for fear of IP leakage or whatever. AI is really in trusted assistant mode at this point. The big problem is in changing habits and changing ideas. I had a hard time retraining myself to turn to an AI before Google and now I get exhausted - when I’m on a roll it’s like pair programming with a really enthusiastic partner, I’m tired before long.
Agree on the age difference but they have the same challenges. My 15 year old knows about ChatGPT and wanted me to show her how to use it, but it was the blank page problem - it’s one thing to solve a problem where all the parameters are given but what do you do when you can do anything?
“… but what do you do when you can do anything?”
Grappling with this as a parent and a teacher. No firm answer
What about as an entrepreneur? Like you have an idea for an app and conceptualize it and write it? I'm at the point where I have a couple of interns, great. I have a friend who's talking about creating a fleet of AI employees and his job as CEO is to direct them. Then you end up with agents that are managers - all they do is prompt their employees and report status upwards. What's scary is that at this point in my career, 30+ years in, I have the knowledge to do that, at least on the tech side. What if I could have a hundred employees with no payroll?
I worked through the early days of the dotcom boom and all the signs are here again, this is like 1996. I worked through the cloud revolution where all your infrastructure costs came down to your ability to pay your credit card bill. What I worry about this time around is a failure of imagination.
Start with their interests, that's always a good place to start. The more they use it, the more they can adapt it to their needs and preferences.
My daughter is currently using it to generate SAT questions, and to improve her coding skills. I've provided her with my own workflow and she's adapting it to her own needs.
I've shown her how to create custom GPTs, how to code prompts, I'm trying to teach her what I'm learning myself, at that abstract level.
What about as an entrepreneur? Like you have an idea for an app and conceptualize it and write it? I'm at the point where I have a couple of interns, great. I have a friend who's talking about creating a fleet of AI employees and his job as CEO is to direct them. Then you end up with agents that are managers - all they do is prompt their employees and report status upwards. What's scary is that at this point in my career, 30+ years in, I have the knowledge to do that, at least on the tech side. What if I could have a hundred employees with no payroll?
I worked through the early days of the dotcom boom and all the signs are here again, this is like 1996. What I worry about this time around is a failure of imagination.
"What I worry about this time around is a failure of imagination."
Agreed. I've seen some agentic workflows. They're amazing. I get the impression people are sleeping on this, dismissing them out of hand. I think that's a huge mistake.
I heard people say the Web was a fad too, back in the 1990s. That statement looks incredibly stupid now.
It's wild how opinionated people can be about something they've obviously not explored.
oh, don't get me triggered...
Can agree, if your not doing BI well, AI will be a challenge. AI will be much faster to market with a solid data foundation that is delivering BI day in and day out. Most of the business relies on BI platform to make decisions and understand the business, especially when it is intuitive and accurate. That data delivery maturity drives reliable, trusted data for AI usage. Turns months into weeks for delivery, and weeks into days for experiments.
Listening to Joe's most recent podcast - Bridging the gap between data and strategy - which speaks well to the need for a strategy - but would like to add - the companies data foundation has to go above the "data team" (or the "data teams" that are not talking to each as mentioned in the podcast) for a strategy. Data Strategy has to start from the creation of the data, the operational systems, the 3rd parties, the SAAS's perspective. We are too late if we are building a data strategy from a downstream perspective of a "data team". Enterprise Data Architecture Matters, data can't be treated like a byproduct, or a means to implement a system. The expectations should be that every application being implemented will have the need of BI and AI, to be plugged into your AI Engine. Which pivots to practical data modeling, once your planning on the operational systems, engineering and designing for BI and AI is part of your DNA. Talk and plan data needs at project inception.