Texas Bullpen Podcast
Listen to the weekly Texas Bullpen podcast to stay up to date on all things Texas politics with hosts Brad Johnson, Jonathan Richie, and Cassi Schredder, along with special guests.
Texas Bullpen Podcast
Jerome Greener, Texas Regulatory Efficiency Office - Episode 28
Use Left/Right to seek, Home/End to jump to start or end. Hold shift to jump forward or backward.
The newly created Texas Regulatory Efficiency Office, colloquially known as Texas DOGE, is launching its first project mandated by the 2025 Legislature. Director Jerome Greener details the agency, its purpose, and its new tool designed to more easily navigate the 20 million words in Texas code.
Morning everyone. Welcome to another episode of the Texas Bullpen Podcast. I'm Brad Johnson, co-founder here at Texas Bullpen, and today we have Jerome Greener of the Texas Regulatory Efficiency Office. That's quite a mouthful, but really you can just think of it as Texas Doge. And Jerome, you're uh director of TREO. How's it going?
SPEAKER_00Man, it's good. I I appreciate you having me on the on the program today, Brad. It's gonna be exciting to show you and your audience uh the the initial results of the first seven months of everything Trevor's been working on. So uh excited to give you all uh the sneak peek.
SPEAKER_01Yeah, before we get onto that, give us a bit of a background on on your work history, and I know you were at TPPF before this, but uh um give us the the rundown.
SPEAKER_00Sure, yeah. So um, you know, not an attorney, certainly not an economist. Folks are saying, like, why are you doing this? Uh I believe uh through the grace of God that I've got uh the ability to take a look at complex problems and then build, design, and refine systems that allow individuals and groups to solve said problems. Um and, you know, in c in conjunction with the greatest governor in the entire country, I think we're gonna be leaving Texas uh better than we found her. I I take a lot of the experience that I gained from Americans for Prosperity, you know, grassroots advocacy, and then Texas Public Policy Foundation with the think tank component. Um, a lot of the things that we did there with reverse engineering of how to get these uh complex problems resolved is the same type of mental model and mindset that we brought to this big project.
SPEAKER_01So I don't know if you were uh wholly involved in it, but you certainly saw it happening from the outside and maybe the inside. But what this this law was passed last year, this agency was created. Um what was the effort behind the scenes, if you know, like to get this across the line?
SPEAKER_00Sure. There's two competing narratives when folks talk about the Texas regulatory efficiency office. And if you hear me say TREO, that's that's the you know regulatory efficiency office for short. Um the first one is you know, as you mentioned, like Elon Musk, Doge, y'all are gonna go after waste, fraud, and abuse. And we've looked at Senate Bill 14. Um trying to identify operational efficiencies would be something that we would very much like to do. We don't believe that Senate Bill 14 currently has that part of the scope. Uh we think that uh um, you know, future legislatures, maybe even the 90th legislature, if they said, hey, you know, Treyo is doing such a good job on these regulatory efficiencies, we'd like to expand their scope to look at some operational efficiencies. We would certainly welcome that. But as of right now, we don't believe we have that authority. The second the second narrative is you know, why Texas? You know, the jobs report just came out, I think it was earlier this week, uh, maybe it was last week, where uh you know Texas continues to lead the entire country in job creation. 117,000 new jobs in the last year. Uh Texas has done that for 22 years in a row. You know, just in the month of March, we had 46,000 new jobs created here in Texas. Why do we need to do anything different? Right? And I think folks are shocked when they learn that we currently have 20 million words in our Texas Administrative Code. That's a lot.
SPEAKER_0120 million?
SPEAKER_0020 million words in our Texas Administrative Code. 44,000 regulations currently on the books. And those 44,000 44,000 regulations equate to 274,000 restrictions. Those are all you shall do this, you must not do that. Is that type of regulatory accretion going to perpetuate the Texas miracle? I believe the answer to that question is the genesis of Senate Bill 14, right? I think it's the reason why you saw such an overwhelming bipartisan support in both chambers. I think it was uh almost two-thirds in both chambers supporting this bill. That along with Governor Abbott's vision to keep Texas the best place to live, work, and raise a family. We believe at Trail we've got clear direction to partner with our agencies to permanently shrink Texas government through regulatory efficiency.
SPEAKER_01And when did this um you got hired for this late last year, I believe. That's right. Well when did this officially open um after passage of the bill?
SPEAKER_00So um that was so I was hired October 1st, and uh the first thing that I did obviously is recognize that my weakness is I'm I'm not a lawyer, I'm not an economist, I need to surround myself with uh highly capable people, and that's what I spent the first three months doing is building a team, a high-caliber team. And then beginning in January, we started really looking at okay, how are we going to resolve this problem? Um and we were able to procure, I think on average, procurement takes about six months in the state of Texas. We were able to uh do everything we possibly could to condense that timeline down to about three months, not breaking any laws, staying compliant. But hey, if it took uh you know a a couple days in order for a process to make its way through, uh, we wanted to try to see if we can circumvent that particular bottleneck, uh, remaining legal, remaining compliant, um, and and get things done. We were able to get our vendor procured Vulcan Technologies, which is uh based right here, headquartered in in Austin, Texas. And uh they've been able to uh accelerate everything that we've uh we've put into place. They've allowed us to put into overdrive. And I'd love to be able to talk about some of the uh deliverables we think Cinnabil 14.
SPEAKER_01Yeah, and you before the sh show started, we were talking about the four deliverables, I think. You want to leave those out for? Yeah.
SPEAKER_00Yeah, so we from looking at Cinnabil 14, we believe there's four main deliverables we've got to uh that we're responsible for in our office. The first is to create a website that allows uh for Texans uh to easily navigate the administrative code, that 20 million words as we talked about. Uh, we believe we need to partner with our agencies to identify regulatory efficiencies. We need to provide our agencies with guidance materials, and then fourth and certainly not last, but or certainly not least, we need to provide a biennial report uh due on December 1st of this year that'll talk about everything we've done in 2026 and everything that we hope to accomplish in the 90th legislature. If if you if you want, I'm happy to start with that first one, the website. Yeah, let's do it. You're launching something today. Yeah. 20 million words to navigate uh can be difficult, apparently, for some Texans when they're uh specifically, I'm thinking about these rising juniors in high school or just about to graduate seniors, or even if you've if you've already graduated and you've got your first job and you're thinking about, hey, what's the next career path for me in order to have a more prosperous life here in Texas? Uh that was the primary audience when we created the state administration manager, or Sam, short for Sam Houston. And I'm sure you're familiar with generative AI. I was gonna say, yes, I'm familiar with Sam Houston. God willing, right?
SPEAKER_01If you're you're raised here in Texas, generative AI, less so, but Sam Houston, definitely.
SPEAKER_00But you you've seen these generative AIs. You've got your Google Gemini, you've got your Cloud, you've got your Chat GPT, and these generative AIs, their large language models are configured to solve your problem. They want to help you out. That's that's their whole purpose, that's the whole design uh and and reason for their existence. And the the challenge that comes with that is each one of those generative AIs has strengths and has weaknesses. So if Brad Johnson's problem is part of one of those generative AI's weaknesses, it'll simply fabricate an answer in order to fulfill its design purpose, right? It'll it'll hallucinate. And I think that's the reason why you're seeing all these. That is the technical term for hallucinating. Yeah. Um and I'm learning too as we go along here. Uh but I think that's the reason why you're seeing all these lawyers across the country getting disbarred because they're referencing cases that don't exist, right? I think I read somewhere that over 70% of the responses are from these generative AIs is coming from Reddit or Yahoo answers. That's not what we wanted to do with SAM. SAM, when you take a look at SAM, has ingested Texas case law, has ingested Texas statute, has ingested Texas regulation, has ingested the information on our agency's websites. That's its primary database.
SPEAKER_01So when it's narrower in scope, which makes it more accurate. Correct. That's what you're saying, right?
SPEAKER_00So yeah, we're pulling from a designated database of fact, of objective fact, of things that Texas has done and created and said this is the way that we're going to regulate our state. That's what Sam is pulling from. But then it partners with those generative AIs and it says, hey, Google Gemini, for the for the other part of Brad Johnson's question here, uh, this is your strength, give me the answer here. And this it's an it's an agentic AI, right? So the the agent saying, Hey, Google Gemini, give me, give me the answer here because this is your strength. But Claude, uh, this is Google Gemini's weakness. I don't want Google Gemini to answer this. I want you to answer this because that's actually your strength. So it's literally like a project manager saying, here's the best parts of all the answers in order to get you the best and most accurate information. And if your question is so complex that it can't resolve it with that one agent, then the agent has the ability to clone itself an infinite amount of times in order to resolve whatever it is you're looking for. It's mind-blowing technology, right? But again, what I want to emphasize and stress is its primary source, its primary database is all that information that we've ingested to ensure that we're looking at state law, case law, statute, regulations, et cetera.
SPEAKER_01How long did it take y'all to build this up? I mean, uh building these AI models takes, I imagine, quite a bit of time, and especially if it's having to go through 20 million words, right?
SPEAKER_00Yeah, so uh fortunately for us, uh we have the benefit of Vulcan Technologies, which is the same vendor that Virginia, which is probably the closest in model and design to what we're trying to do with regulatory efficiencies. Virginia picked them up and it really put them into overdrive. And they and they shared with us that hey, they were looking at uh identifying a lot of opportunities, but whenever they were whenever they acquired uh Vulcan, they were able to expand their search to include so much more information. And so um for us, it was really more of a configuration of a system that they've already built, and uh we were able to hit the ground running. We spent the month of March looking at about 11 agencies. It's a um that's a that's a that's a lot of uh uh information that we looked at. I think when you when you see the website, there's gonna be 4,177 rules that we review in the month of March. And the reason why that is is we needed a sample pool size big enough to tell us, like, hey, this system that we need to design in order to identify these operational efficiencies, it it needed to be big enough, right? Um that 4,177 rules equates to about 1.5 million words of the Texas Administrative Code. And I have to take a minute here to brag about my team because we worked weekends, we skipped holidays, we came in early, we stayed late, we did everything we asked, we we had to do in order to get the job done. And I mean it when I say this, we are the busiest bunch of bureaucrats in the history of bureaucracy. Uh but I think you'll see the the fruit of our labor of the first tranche of agency results, these first 11 uh in in collaboration with our agencies, we think we've identified roughly around 437 uh rules to be amended or repealed. And uh if if that does move forward, we think that's probably gonna reduce administrative code by roughly about 70,000 words, maybe a little bit less. And uh uh and that'll equate to probably saving Texans about 120 million dollars. Uh and that's just the first this this first tranche of agencies. And I and I do need to explain like this is like the the the first phase of of our first tranche because um we we had a condensed amount of time to take a look at all of these agencies, and we wanted to do it in the month of March and say, okay, here's our system. What we did is we said, what is the biggest, lowest hanging fruit that we can we can spend our time on with each one of these agencies to say, okay, here's here's how we could potentially remove the Texas administration code, probably not remove, but reduce. Here's how we could potentially save Texans money. We wanted to focus our attention on those first, and what we're gonna do in the second phase is come back and complete their entire rule set, right? Hey, here are the opportunities that we were able to do to optimize your your your regulatory framework.
SPEAKER_01And then that's Did you talk to the agencies themselves to brainstorm this at all? Or yeah, yeah.
SPEAKER_00So uh when when you look at Sam, uh at the top of Sam is gonna be a progress page, the same the same vendor Vulcan Technologies that we procured to make the website is the same same vendor that we have to help us generate these these uh um um regulatory uh recommendations for for efficiency. And so when they generate these these recommendations, they have a whole team of folk that look at them for accuracy. They review for accuracy, they've got lawyers, they've got economists, and then we look at them. TREO, the regulatory efficiency team, looks at them and review reviews for accuracy. And then we take these recommendations to the budget and policy team, and then they review for accuracy. And then we take it to the Office of the Governor General Counsel, and then they review for accuracy. And then we take it to our all-star advisory panel, and then they review for accuracy. And then we actually sit down with the agencies themselves because they are the subject matter experts on all of this, right? They review for accuracy. In fact, uh many of the recommendations that we look at come from the agency's self-evaluation. So it's a combination both of what they think is lowest hanging fruit as well as the recommendations that we've generated. But the bottom line is we try to get as many eyeballs as possible to look at this, to review for accuracy. And it's not even done there at the agencies. Whenever we're done sitting down with them and saying, okay, here's what we think we all should look at, they take those recommendations to their stakeholder community and then they review for accuracy there. Once we've done all of those checkpoints to ensure that everything we've done is as as uh iron tied as possible, uh that then what we do is we provide a red line by provision edit saying, okay, no, we're talking about changing rule X to potentially changing it to Y. Here's what this would look like as a point of reference, not a mandate, not a directive that you have to do it this way. But if to to understand like like here's a jumping off point, here you go, here's that red line. And then a corresponding cost-benefit analysis uh that syncs up with that red line, that's where you get those numbers from is uh when we when we're able to come to a fully uh cooperative agreement of here's what we think y'all should move forward with, each one of these agencies, uh, that's what uh allows us to put these numbers up on the board.
SPEAKER_01So bring this down to a lower level. Um how how does an individual who is going to the new website y'all have launched, how do they use this to say help themselves become an electrician?
SPEAKER_00Great question. Right. So um you literally go to the website here, you type how do I become an electrician, and uh And what is the website? It's efficiency.texas.gov, right? Everyone in Texas is gonna be able to use this. Folks that want to come to Texas because there's better career paths here, less burdensome regulations, um, you know, anyone's gonna be able to use this website. You type in how do I become an electrician, and what used to take an individual hours as they go to different uh agency websites, compile information, collate stuff, remove things that are not necessarily relevant to them. Uh now Sam is gonna provide you with a step-by-step instructional on how you can become an electrician and then provide a cost and time summary so that you can anticipate and expect exactly uh knowing exactly what to expect whenever you go down this particular pathway. Um what used to take individuals hours to do all this is now gonna take roughly 15 seconds. We know that's still too long, and in 2026 we're trying to get that down into single digits and then ultimately, you know, immediately. Uh but as of right now, it takes you roughly 15 seconds uh to get that information. And you know, time being the scarcest of all resources, I think we can say Texas is literally doing everything to accelerate job creation and economic expansion right now, and that's what we're most proud of.
SPEAKER_01Aaron Powell And so when someone gets that slate of step-by-step um procedures in order to become an electrician, what are the the safeguards you all have to prevent the model from hallucinating in that regard?
SPEAKER_00Aaron Powell Yeah. So uh we reviewed it for accuracy. Uh the the uh the vendor that we procured reviews it for accuracy. We've actually shared with a specific example how to become an electrician, we shared it with the Texas Department of License and Regulation. We said, please poke holes in this. Let us know we're wrong, we've got to fix it.
SPEAKER_01So it has been stress tested, is what we've got.
SPEAKER_00We've stress tested every possible way. And we know that it's not gonna be perfect. I'm sure there's probably gonna be some sort of hole uh that we that needs to be poked through. Please let us know that uh please get send us that feedback and we've got an evaluation portal up on that website. We we, you know, to borrow a a um um uh a you know a famous tautology, we don't know what we don't know. And as we do uh glean additional information, we want to be able to incorporate that in the website. We want this to be a highly functional website, as sophisticated as possible. Let us know if there's additional functionality or if there's feedback that's gonna allow us to prove what we're doing here. Um but when we send it to TDLR uh and we ask them to poke holes in it, they said we we can't find it. Everything is everything looks good on our end. So uh we saw that as a as a as a major stamp of approval for for rolling out the website.
SPEAKER_01And so that is launched as this podcast publishes, correct? So you can go to say the URL.gov. Okay. Um let me ask you a couple more questions before we go. So, you know, the inspiration for this was, of course, Doge at the federal level, and you kind of touched on it at the top, but I I'm interested to hear more. They Doge ran into the problem where they found a lot of bloat. They cut a lot of bloat, but then they saw uh some things that at least uh reasonable people could uh surmise maybe should not have been cut so quickly or should not have been cut at all. How do you prevent that from happening and who actually does the cutting? Because there it was all executive.
SPEAKER_00Pardon me. I think um the greatest contribution that Doge gave us was reigniting the spirit of regulatory efficiency, right? I think a lot of folks in a lot of states are leaving the cave on this mentality where uh this this natural evolution. Legislatures provide uh pass new laws every year, and those new laws usually necessitate additional rules, right? Uh it's it's it's a natural accretion for for this to happen and uh an evolution, if you will. Um a lot of folks have just said for for many, many years, you know, that's just that's just what happens, and there's nothing to do about it. And I think uh what Elon Musk and his team has done is said, no, you you can do stuff on it. And I think that spirit is very much alive in Senate Bill 14 and the regulatory efficiency office. And we have an interstate doge council. There's, you know, South Carolina, Virginia, Florida, Indiana, Idaho. There was a lot of states that have come together and said, hey, here are the resources that we're pulling together. You know, use this if you're looking to solve X, Y, and Z. Avoid one, two, three uh landmines if you're going to go down this path. Uh different state legislatures uh have designed their their doge teams to do different things. I think Florida's looking specifically at uh local government, uh, while uh there's some other states that are looking at operational efficiencies. Again, we've we we believe our lane is in that regulatory efficiency, but we're all coming together and we're all kind of like growing off of the movement that Elon has put out there and said, okay, we can we can permanently shrink government and we can we can make government work better for its people.
SPEAKER_01One of the uh the few criticisms of this, because as you know, there was a lot of bipartisan support. One of the uh at least critiques I heard throughout this the passage of the bill was, well, we already have Sunset. It already exists to do this. Um any response to that? How is this different uh and not duplicative to Sunset?
SPEAKER_00Yes, so Sunset is uh a product of the legislature. We are a product of the executive branch, and so I believe the Sunset's doing a little bit more of a comprehensive, maybe there's some performance measurements that they're reviewing, and they're and they're looking at agencies about once every 12 years where we're we're part of the executive branch along with these agencies. We're designed to assist them, and we're looking at doing this rough, this trayo-initiated process uh once uh every three years is what we think is going to be the cadence at this point. And uh we're looking specifically at those regulatory uh regulatory efficiencies. So we met with Sunset early on and and let them know here's what we believe is our role, and and here's what we believe y'all's roles are. And uh like together I think we can we can complement each other in many different ways, and that's that's the lane that we're trying to stay in.
SPEAKER_01So basically to sum it up, you you are making the case to the legislature. Uh you're putting out a report saying here's where some bloat is, here's where we can shrink things down, and then they have to do it, correct?
SPEAKER_00So uh great question. So there's there's two different results that can generate from what we're working on. The first result is anything that an agency can change by role, the permissive authority, right? So that's what we're focused on right now with with the 11 agencies that we looked at in March. What are some of the efficiency opportunities that they could take action on right now during the NRM and then we could have we could actually see the effect here, the one $120 million go into effect, uh, those those cost savings. The the second batch, uh if you will, so I guess third batch. So phase one, phase two, and then the the phase three would be those legislative recommendations. Anything that requires statutory action, that's gonna go into that biennial report that says here's everything we've done in 2026, but also here's a lot of the recommendations that we want y'all to consider for the 90th legislature. Um, that'll be the phase three of everything we're working on because that's gonna require them to take action uh in in 2027.
SPEAKER_01Anything else you want to touch on on this?
SPEAKER_00Man, I I think we're we're we're as I said, we're busting our tails to to deliver uh results to Texans as quickly as possible. And uh and as I said, we've got that evaluation portal. If there's any feedback, anything we can do better, please let us know. Greatly appreciate being on your program today.
SPEAKER_01Yeah, thanks for joining, Jerome Greener.
SPEAKER_00Thank you, sir.