Sensemaking: Weather Modification & the Texas Floods

Introduction

We use AI research agents from Claude & Grok to learn about weather modification efforts in Texas and any connection to the Texas floods.

We are left with more questions than answers-- and a clearer understanding of the limits of research agents.

The agents proved helpful in learning basic facts:

1. Texas has the largest state-regulated weather modification network of any state in the US.
2. Weather modification efforts are funded by local farmers and water districts, but regulated at the state level by the Texas Department of Licensing and Regulation (TDLR).
3. All weather modification operations in Texas must be licensed by TDLR.

...but not at all helpful in clarifying:

1. Details of the specific operation Rainmaker Technologies conducted in Central Texas on July 2nd, 2025 and any previous operations in the area. (All information comes directly from their CEO posting on X.)
2. Which regulated entity was responsible for the July 2nd operation. Rainmaker has no TDLR permit itself, but contracts with permitted entities. It's unclear which.
3. Whether or not the cloud seeding operations in the days leading up to the storm had any impact - or could have had any impact - on the storm itself or the resulting flooding.

Much of the discussion of this topic has centered on 1) baseless charges that the cloud seeding "caused" the storms or flooding, or 2) ad hominem attacks of "conspiracy theorist" on anyone suggesting a link between weather modification efforts and the intensity of the floods.

Little to no discussion centers on the actual facts of the situation, the science of cloud seeding, and what uncertainty exists around downstream effects of cloud seeding.

We remain interested to investigate if there is - or could be - any specific impact of weather modification efforts on the intensity of Texas storms. If only so we may avoid or mitigate future catastrophe.

Here is our attempt to add nuance to the discussion while grappling with the frustrating limitations of today's 'state of the art' research agents.

Transcript

Quick detour from our regularly scheduled programming.

I live in central Texas where we had this 'once in a thousand year' storm. The death toll is rising every day. It's now over a hundred. Twenty plus kids killed, girls still missing. A really bad situation.

I'm a Texas taxpayer and concerned for safety and all this. I'm just looking into why and how this happened and preparations and mitigations et cetera.

And obviously a fast moving situation and new information coming out. We don't know what all caused what and such.

But here is where AI should help. We have a lot of institutions like xAI pouring money and resources into creating truth-seeking AIs and such. That should help us make sense of these things.

I posted a tweet yesterday. Just kind of noting a couple of things. And I want to get into like what went into this tweet, research that I've done since, and really highlight the limitations of these research agents like Claude and xAI that I've used.

We did a video like this number of months ago on that drone situation in New Jersey. Concluding that like these research agents pretty much suck. They're good at some things. They're good at largely asking questions or coming up with information, but not synthesizing things into actual truth.

So I just feel like this was important enough of a topic to add a little bit of nuance to the situation because this conversation attracts a bunch of people like throwing ad hominem arounds. Let's get to an actual conversation about the issue.

It was news to me that Texas has the largest state weather modification program of any state in the country. Oh, are my taxpayers dollars paying for that? Just like tell me about it. Tell me about this thing.

And then tell me is there any connection between largest weather modification program of any state in the country and one in a thousand year floods that killed 100 people. Just tell me if there's any connection between that. It seems like a legitimate inquiry to me.

And part of my point here is that anybody suggesting any link between these weather modification efforts - that at least one entity was doing, seeding up until the storm started in that area approximately - maybe this had an impact on the storm - [is attacked as 'conspiracy theorist']. It seems to me like a legitimate inquiry.

But there's a whole bunch of bullshit out there. So hopefully I just want to like filter through some of the bullshit.

I don't come to any conclusions here. Which is sad because hopefully that things like xAI or some of these tools could help sensemaking in these environments.

I give huge props to Augustus for coming in within a minute of my post being like, 'Please dig in and let me know.' Okay. I will. And let's just ask some questions.

Cool. So question number one [for Claude] is tell me about the weather modification efforts in Texas. How much does the Texas state government pay per year for weather modification programs? Any all information about what specifically is paid for and why.

So this goes off and does a report compiling a bunch of sources. With a caveat that you can't take any of this as truth because it's just putting together sources and trying to summarize. And there's things that it isn't able to do like search through government databases. So it's always going to be flawed. But it gives you sources. Then you can look at the sources yourself and maybe piece a few things together.

So let's just go through a little bit of this.

Texas operates the nation's largest state-regulated weather modification network, covering 31 million acres through seven cloud seeding programs that run entirely on local funding after state support ended in 2004.

So the state doesn't directly fund weather modification, but the state maintains a state regulatory agency TDLR that regulates anyone doing weather modification. Cool.

The Texas Department of Licensing and Regulation oversees all operations through a comprehensive regulatory framework.... ensuring safety standards while programs report producing over 1 million acre-feet of additional water annually...

Okay, local water districts foot the bill.

Seven major associations orchestrate cloud seeding across West Texas

Texas Tech University's 2019 evaluation found seeded thunderstorms lived 41% longer, covered 44% more area, and produced 24% more rainfall than untreated storms.

Okay so the broader scientific community is skeptical of the science or how effective it is. But apparently there's been enough research and such that farmers and local districts are happy to fund this. Okay. It's a thing. It's not funded directly by taxpayer dollars, but taxpayer dollars are paying for the regulatory agency that oversees this and apparently anybody modifying the weather needs a permit.

Okay give me a chart of all expenditures on weather modification in Texas in the last five years. I wanted a chart. It did not give me a chart. Some sources blah, blah, blah. Give me a list of all entities with TDLR weather modification licenses. I didn't click the research button, so just did some basic searches there.

Where would I find the licensing database?

So you can go there, you can search the licensing database. I was like, okay, like Rainmaker is modifying the weather, they should have a permit, right? No, they are a contractor of TransPecos WMA, that's out west, the Ward County irrigation district. So apparently they are contracting with multiple different districts. This would not be the one that covers their central Texas one. That's just someone on Twitter surfaced this. Claude couldn't find this.

So I asked it for like a report on Rainmaker specifically, we'll get to that in a second, but just one point on how these things are flawed:

It gave me this research report where it said

A critical compliance gap emerged in federal reporting requirements. Despite conducting weather modification operations, Rainmaker Technology Corporation has not filed any Form NOAA-17-4 reports as required by the Weather Modification Reporting Act of 1972.

This is not true! It's not even that they didn't find it. I gave them the form that they filled out. Here's the form that they filled out, Rainmaker Technology at the bottom.

Now this is a form that goes to NOAA. This is some federal agency you've got to notify. Again, contracting on behalf of Ward County Irrigation, TransPecos WMA.

But limitation: 'Oh, they didn't fill out this form.' I gave you the fucking form that they filled out.

You're absolutely right, and I apologize for that significant error. You did indeed provide me a NOAA Form 17-4... This form clearly demonstrates that Rainmaker Technology Corporation did file NOAA Form 17-4 for their Trans-Pecos operations. My statement about "no NOAA reporting" was incorrect based on the evidence you provided. However, this raises several important questions:

So again, we're filtering this like, are these actual important questions or are you still bullshitting? Let's just read it. Ward County Irrigation District, Rainmaker Safety Team, who cares?

This NOAA form covers Trans-Pecos operations only. The July 2nd operations Doricko mentioned for "South Texas Program" or Central Texas would require separate NOAA filings, which may or may not exist.

So it's like, it doesn't know. It can't find this. It doesn't know. Anyway.

I'm interested specifically in what regulatory authority governed their July 2nd operations? Like, who was this with? Who was the entity that is regulated that governed their July 2nd operation?

Okay. That also raises more questions. But like, okay. They did an operation.

The only real information that we have, I think, about their actual operation on July 2nd comes from statements that Augustus is making on Twitter. His response here:

Cool. Well, maybe Augustus can answer some more questions about what that operation entailed.

And you got some good questions being asked by people who you might otherwise think are bots with this kind of username.

Did it rain from your operations two days prior to the tragedy? Is it possible remnants of your operation converge with the incoming storm?

Fuck. I don't know, man. I don't know.

So here is the thread that originated my tweet there. Research rainmaker and their cloud seeding efforts in central Texas. Yeah. Yeah. Subject number two, uh, limitations of these, these agents, your cloud seeding investigation report, it concluded rainmaker was not involved in Central Texas operations during June to July 2025. The report covers other things.

Central Texas falls within an extensive network of cloud seeding programs that have operated for decades without Rainmaker's involvement.

Uh, uh, excuse me. They were active there according to their CEO. Refine your analysis.

Thank you for sharing these posts from Augustus Doricko, the CEO of Rainmaker. This is significant new information that directly contradicts my previous research findings. The posts clearly show that Rainmaker was indeed conducting cloud seeding operations in South Texas in early July 2025. Let me research this more thoroughly to understand the full picture of Rainmaker's actual operations in Texas during this timeframe

And then it like research is further based on that and you think that would like help it find better information.

Oh, the report finds no evidence of his post about South Central Texas operation confirms rainmakers uses only drones and reveals the company lacks Texas weather modification permits.

That's probably true. They're just contracting with ones that do.

Uh, it also verifies that high moisture conditions on July 2 to 5 would have justified suspending operations before.

Okay. That doesn't help at all. I say, 'Investigate its SPECIFIC CLAIM that it played NO ROLE in the catastrophic flooding. It was actively cloud seeding until the storm hit. How can one assert they played no role when their entire fucking business is intended to create rain?'

I understand your frustration with the apparent contradiction. You're right to question this. If rainmakers actively cloud seeding in the area until July 2, just two days before catastrophic flooding that killed dozens of people, their claim of no role seems suspected best. Let me investigate the specific timeline in science.

Well at this point I don't really trust this fucking thing.

But it cites some sources that have some interesting stuff.

It had a good point here:

The fundamental contradiction in weather modification claims reveals itself starkly: companies cannot simultaneously argue their technology successfully creates rain for drought relief while bearing no responsibility for rain-related disasters 48 hours later.

Okay, I wouldn't go that far. I'd say: your technology is either something that people should fucking pay for because it makes rain or it doesn't.

And if the stuff that you're doing doesn't make rain, then why are you being paid and regulated-- to do what exactly?

So if you're doing stuff that creates rain immediately before a massive fucking once in a thousand years storm, asking questions about that is really actually legitimate.

---

Addendum

Not included in the video, we had similarly frustrating experiences talking with xAI's Grok. It was a little better at surfacing recent posts, but largely reported on 'he-said, she-said' type summaries augmented by basic info from web searches. Not helpful.

Closing thought

These research agents provided no definitive answers and led only to more questions.

We need better research agents.