Copasetic Flow

I started my lab book entries when I was a physics graduate student. It's kind of amusing and kind of cool how far I've come. I have the equivalent of a grad student, (aka Claude Opus 4.7), working for me now. I spent some time over the weekend setting up an OCR framework for a book research project of mine. I've been coming up to speed on evals, so I decided to run one to determine which model was the most accurate and cost effective for doing OCR on travel manifest pages. I stepped the eval along rather than automating it and talked the results through with Opus as I went.  First, it turns out that Opus at low effort is the most accurate and the most cost effective choice! That was a surprise. The result has to do with Opus' ability to look at higher res images which means it needs to think less for OCR vs. Sonnet. Second, at the end of the eval, as I was preparing to write up my results it occurred to me that I could ask my grad student to do it instead. Here's...

 I took time to play with a new Dolt enabled app example called Quorum last night. Quorum sets 13 LLM agents with different defined personas  loose on a users question. The agents come up with solutions to the question and then discuss their individual solutions with each other to arrive at a consensus. There's much more detail in this blog post that accompanies the app. Quorum is cool. It is not, however, what I wanted to talk aobut here. Instead, I'm going to focus on the blog post for the app. In short, I'm very excited to see ideas that I've used to manage verification processes for years get codified into tools for LLM agents. Here's one of the important parts " I can shut down the app, lose the server, or disappear entirely — and the deliberation history remains, publicly accessible and cryptographically verified. " Imagine what an engineer can do to work back through their debug hypothesis tree with that sort of infrastructure! As the article'...

 I asked ChatGPT 5.4 Thinking to write some JSON-LD text to summarize a blog post this morning along with alt text for the posts images and got this When I asked it to file a but report on itself, I didn't really expect it to succeed, but I didn't expect more oaicite index listings for the proposed issue description. (Kinda obviously, I also haven't had enough coffee yet :) Anyone else seeing this? Happenings of Interest   (radio and nature) Where was our skip zone at San Bruno? For example. QSO Log Table containing QSOs in text Callsign rx RST tx RST Time (GMT) Frequency KBTEST 539 559 16:42 14058.3 kHz (Add callsigns as post tags?) Unschooling Highlights POTA tx QSL: QSL rx album: References POTA ( Parks on the Air ) Local Ionograms https://lgdc.uml.edu/common/DIDBYearListForStation?ursiCode=PA836 (for example) Videos Demo

  Just a quick note that the AI tutored Project TouCans exams are up and running for the latest US technician class question pool for the license exam.  The exam sprang out of two different projects here at the home QTH. KO6BTY is studying for her extra class license. Also, I'm learning about developing code with AI. And voila: AI tutored ham radio practice exams. Try the AI-tutored Technician Class practice exam now. Project TouCans technician class practice exams . Read more about this project: First  demo  of OpenAI ChatKit enabled exams First release of extra class exam based on OpenAI responses API Removing the vector store  to reduce costs Experiences with Vector Stores Early debug to add contexts by local compute and storage First release of extra class exam  with no AI

 Debug notes from getting the AI help feature of the free ham radio exams to work today. Grabbing a text answer from OpenAI still works: That's from this method async function retrieveTextWithFileSearch ({ system , user }) {   const vsId = localStorage . getItem ( 'vector_store_id' );   if (! vsId ) throw new Error ( 'No vector_store_id found.' );   answText = answText + " " + user ;   const resp = await openai ( '/responses' , {     body : {       model : 'gpt-4.1-mini' ,       input : [         { role : 'system' , content : system },         { role : 'user' ,   content : answText }       ],       tools : [{ type : 'file_search' , vector_store_ids : [ vsId ] }]     }   }); With the agent  flow, it doesn't work async function retrieveTextWithAgent ({ system , user }) {   // 0) Make sure we have ...

 We pushed the teletype prototype for Project TouCans further today! KO6BTY transmitted a CQ call on 2 meters, and I managed to decode it—at least semi-successfully—through audio from my K6 UVK5(8). It’s not perfect yet, but it’s another good sign that Project TouCans’ RTTY experiment is working. For those that don't remember, or weren't follwing along, I started workign on the possibility of teletype using Project TouCans back in May . I didn't have the time to add a frequency change relay and possibly a different internal keyer to TouCan's Rockmite, so I settled on converting the microPython code intended for TouCans into JavaScript that could be run on the blog using GPT. It worked. I knew this right away becasue DroidRTTY decoded the audio output of the blog page app. Meanwhile, CW on 2 meters has become kinda popular lately thanks to  KI7QCF . My mind put the two topics together today, and voila! Here's a video that explains it all. The semi-successful deco...

 This week's AI project is to create an AI Morse code decoder. I've been working with the new ChatGPT 5 model since late last week. I've asked a few different models if they could understand Morse code. ChatGPT 5 couldn't. Gemini couldn't. That's when it occurred to me that this was probably the perfect time to learn how to use TensorFlow to make an AI. So, I changed my question. I asked ChatGPT 5, "If I wanted to setup a model that learned Morse code using Google's Tensor engines, could you describe the entire process and output the code for me?" To which it promptly, (what an awesome pun!), replied, "Heck yes—that’s a super fun project. Here’s a complete, practical path to a TPU-accelerated Morse code recognizer using TensorFlow + CTC (Connectionist Temporal Classification). It generates synthetic Morse audio (with realistic timing/noise/tempo wobble), trains a small CRNN on log-mel spectrograms, and decodes with greedy CTC. You can run it ...

 The TouCans keyer works much better than it ever has in the past, (apologies if you were on the receiving end of a key down lockup), and it's all due to connections I made while unschooling with the 14, 12, and 10 year old gang of kids here. This story starts when they were 9, 7, and 5 years-old, so strap in :) I've probably written about this before, but here's the rough chain of connections The, then, five year old, known as Tawnse here, and I attend a design exhibit at SFMOMA where She and I find a Foldscope , a simple origami-inspired microscope that's pretty fun, and pretty cheap Tawnse develops an interest in microscopy The gang and I join the San Francisco Microscopy Society The Society is working on archiving their old documents. Daize, (aka K06BTY), and I are working on a book about Mike Gladych where we've accumulated a lot of documents. We attend their archivist committee meeting. I read the list of committee members, and there's a name I recognize, ...

 I've read a little bit about 'local data' lately, and I've become fairly excited about the idea. Two of the features of the concept that I like are that your data should be your own, not stuck behind a corporations gated off wall on the internet and that you should have a copy of that data that lives on physical media within arms reach. I won't venture into the second one today except to say that with Google, in particular, somewhat constantly doing away with applications it becomes bored with and Amazon deciding it has and then has not rights to serve movies I 'purchased', having my data in a usable format and in arms reach physically makes more sense than I'd like for it to. Today, I'm going to talk about the first thing, being in control of your own data to use as you see fit. As someone who develops open source apps, there are benefits to everyone beside me holding on to their own data. The biggest one for me being that I don't have to figu...