**Artifact** from Bead: 2e635fc9 · [canonical source](https://redfish.acequia.io/guerin/.agents/2e635fc9-9a8b-4507-8cd4-9c837ebbda9d/2026-05-02/artifacts/chat-log-whisper-transcription.md) · session 2026-05-02 · discussion: Talk: 2e635fc9
*(Current Action)* — Started executing a transcription on `C:\Users\steph\Downloads\20260501_133828.mp4` using Whisper AI. 1. Attempted to find a transcription script within the `sites` local workspace context. 2. Verified `Python` environment requirements. 3. Activated local `c:\Users\steph\Documents\sites` python environment and installed `openai-whisper` and `setuptools-rust` 4. Used whisper CLI model via `C:\Users\steph\AppData\Roaming\Python\Python314\Scripts\whisper.exe` on `.mp4` using tiny model in English. 6. [Review Full Output here](https://redfish.acequia.io/guerin/uploads/20260501_133828.txt) > Here's the plan for next week. What we just discussed. Yeah, and I'll catch in too. What's the big purpose? We're each contributing an element of this observation system to show capacity and real time and then also like a post processing assessment. Being able to collect this distributed network and kind of show it and understand it and maybe model it in a way that makes sense. That makes sense. Yes. > > So we have two landscapes. We have the stream table and we have the real world of Rio de Doceau. For both of them we have videos of the past, March 11th in stream table and we have some Rio de Doceau footage. Now the question is can we reconstruct the past and Valentin can show in her tool both potentially? Right. She would be doing the kind of triangulation of the cameras. Yeah. > > And then even that submission process is a UX UI design. Even if we sketch what that could be and we're not even making any software... *(Transcript continued through 6:00 duration)* [01:30.000 --> 01:33.000] And there is then just a shout out to scale. [01:33.000 --> 01:35.000] So we have the scale of file scale of flood. [01:35.000 --> 01:38.000] There's a many, many concerns that we kind of thought about that. [01:38.000 --> 01:41.000] And we're picking this scale for this demonstration. [01:41.000 --> 01:42.000] Yeah. [01:42.000 --> 01:45.000] And we're sort of at this different scale. [01:45.000 --> 01:46.000] Okay. [01:46.000 --> 01:49.000] And then I'll focus on the video processing of the stream. [01:49.000 --> 01:50.000] Yeah. [01:50.000 --> 01:53.000] So how do we go from video that's been rectified? [01:53.000 --> 01:58.000] And you can, there's, there's information in the fact that even if it's not rectified, [01:58.000 --> 02:01.000] you can get some information of the flow rates and things like that. [02:01.000 --> 02:03.000] But it gets really nice. [02:03.000 --> 02:10.000] Even before processing, being able to fly to a point in any point and being able to annotate what's here. [02:10.000 --> 02:12.000] And we're not going to do that. [02:12.000 --> 02:13.000] Yeah. [02:13.000 --> 02:14.000] And then also to augment. [02:14.000 --> 02:18.000] Oh, that's Mrs. Jones call word 911 coming from which camera is she in right now. [02:18.000 --> 02:21.000] And maybe she's clinging to a tree, she's saying. [02:21.000 --> 02:22.000] Right. [02:22.000 --> 02:24.000] And we have her GPS location. [02:24.000 --> 02:25.000] What camera? [02:25.000 --> 02:27.000] What helicopter just flew over? [02:27.000 --> 02:28.000] That could have that information. [02:28.000 --> 02:29.000] Right. [02:29.000 --> 02:30.000] So that's the bigger scope. [02:30.000 --> 02:32.000] But on the stream table. [02:32.000 --> 02:35.000] So then we get calibration and the build and scrub time. [02:35.000 --> 02:38.000] And then with machine vision, computer vision. [02:38.000 --> 02:40.000] We're getting flow rates. [02:40.000 --> 02:43.000] And, you know, I think it's on our parameters. [02:43.000 --> 02:44.000] Yeah. [02:44.000 --> 02:45.000] How do I really go? [02:45.000 --> 02:46.000] I've been looking at it at the branch count. [02:46.000 --> 02:52.000] And that goes in the honor is really nice with the phased diagrams and the deposition based on slope and flow rate. [02:52.000 --> 02:54.000] You'll be doing it in one situation. [02:54.000 --> 02:56.000] If we're going around the room. [02:56.000 --> 02:57.000] Yeah. [02:57.000 --> 02:58.000] Really getting this coherent story. [02:58.000 --> 03:03.000] Eric has a really nice manipulating landscape and intervention. [03:03.000 --> 03:04.000] Right. [03:04.000 --> 03:09.000] And going from a table top thing or on a screen to actually [03:09.000 --> 03:14.000] Medication in Rio de Solfrey might be in Colorado Springs, which could really nice. [03:14.000 --> 03:17.000] It's like, what you guys are making is the same tool. [03:17.000 --> 03:22.000] And just the, you know, it's not custom for that place as a site. [03:22.000 --> 03:24.000] It's just like analog. [03:24.000 --> 03:25.000] You can go to a different site. [03:25.000 --> 03:26.000] Right. [03:26.000 --> 03:30.000] And then give me Ava's. [03:30.000 --> 03:31.000] I don't know. [03:31.000 --> 03:32.000] I don't know. [03:32.000 --> 03:35.000] Craig was working with her. [03:35.000 --> 03:36.000] And she has she's coming by tomorrow too. [03:36.000 --> 03:38.000] So I get a sense. [03:38.000 --> 03:40.000] And then. [03:40.000 --> 03:43.000] I'll be a hobby or have some imagery. [03:43.000 --> 03:45.000] So any of these are too. [03:45.000 --> 03:46.000] Yeah, I don't know. [03:46.000 --> 03:47.000] And he's auditing too. [03:47.000 --> 03:48.000] Right. [03:48.000 --> 03:49.000] And. [03:49.000 --> 03:50.000] Is that it? [03:50.000 --> 03:51.000] Yeah. [03:51.000 --> 03:53.000] I think that it tells that. [03:53.000 --> 03:56.000] How do you help people be collectively intelligent? [03:56.000 --> 03:59.000] There's, oh, frankly. [03:59.000 --> 04:00.000] The hot. [04:00.000 --> 04:02.000] One of my stretch goals. [04:02.000 --> 04:05.000] Can we think of a landscape as computing? [04:05.000 --> 04:08.000] As a side project, a branching river. [04:08.000 --> 04:11.000] So we looked at one of those basic ones of humor. [04:11.000 --> 04:12.000] Mm-hmm. [04:12.000 --> 04:13.000] And there is this. [04:13.000 --> 04:18.000] There is this also this idea of AI and flipping the pedagogy of. [04:18.000 --> 04:19.000] I think of. [04:19.000 --> 04:22.000] You guys are able to explore things without coming up. [04:22.000 --> 04:23.000] From. [04:23.000 --> 04:26.000] You don't think of kids coming through algebra geometry. [04:26.000 --> 04:27.000] Yeah. [04:27.000 --> 04:28.000] Calculus. [04:28.000 --> 04:29.000] You're able to explore. [04:29.000 --> 04:30.000] And then drill down. [04:30.000 --> 04:33.000] And that's, I think that's a nice story of Anna finding [04:33.000 --> 04:34.000] Hobart's curve. [04:34.000 --> 04:36.000] She was able to describe it. [04:36.000 --> 04:38.000] And then see, oh, that's a thing that people have a name for. [04:38.000 --> 04:41.000] But if you had to remember all these names before even what you're looking for. [04:41.000 --> 04:42.000] Yeah. [04:42.000 --> 04:44.000] I think that's a very powerful. [04:44.000 --> 04:45.000] Where AI fits in. [04:45.000 --> 04:47.000] And I kind of think of like. [04:47.000 --> 04:49.000] It's like auto-tune at karaoke. [04:49.000 --> 04:50.000] Everybody can sing. [04:50.000 --> 04:53.000] But then at some point you want to learn to tune your voice. [04:53.000 --> 04:56.000] And some people want to actually learn music there because it's so beautiful. [04:56.000 --> 04:59.000] But it's not required to do. [04:59.000 --> 05:00.000] To participate. [05:00.000 --> 05:01.000] Right. [05:01.000 --> 05:02.000] Right. [05:02.000 --> 05:03.000] Then people that really know how to sing. [05:03.000 --> 05:04.000] Right. [05:04.000 --> 05:07.000] So there's anyone can cook. [05:07.000 --> 05:08.000] And then how it. [05:08.000 --> 05:09.000] So that. [05:09.000 --> 05:11.000] That's that's more what I got out of it. [05:11.000 --> 05:14.000] So it has to be a review. [05:14.000 --> 05:15.000] But I do like. [05:15.000 --> 05:16.000] There is something. [05:16.000 --> 05:18.000] And I use your thing in the faculty meeting. [05:18.000 --> 05:20.000] And I have said. [05:20.000 --> 05:23.000] You know, 15 years have been thinking about it. [05:23.000 --> 05:24.000] And it. [05:24.000 --> 05:27.000] It really takes all the skills of a Los Alamos. [05:27.000 --> 05:28.000] You know, group. [05:28.000 --> 05:30.000] There's multi dimensions to that. [05:30.000 --> 05:31.000] And we're still not there yet. [05:31.000 --> 05:33.000] And I don't want to over encourage you. [05:33.000 --> 05:38.000] But it is amazing how far you were able to get with talent. [05:38.000 --> 05:40.000] But also, you know, the tools. [05:40.000 --> 05:43.000] You didn't have to learn to write a lot. [05:43.000 --> 05:44.000] That would. [05:44.000 --> 05:47.000] Each one of those little pieces could have been a week to three of like, [05:47.000 --> 05:48.000] Oh, shoot it. [05:48.000 --> 05:49.000] It's a minus one not. [05:49.000 --> 05:50.000] I don't. [05:50.000 --> 05:51.000] I don't know how to click. [05:51.000 --> 05:52.000] Turn the insides out to look up again. [05:52.000 --> 05:53.000] Right. [05:53.000 --> 05:57.000] You're that kind of thing. [05:57.000 --> 05:58.000] Yeah. [05:58.000 --> 05:59.000] So I think that's good. [05:59.000 --> 06:00.000] Yeah.