Skip to main content

Posts

Attaching SandCastle to the CesiumJS MCP

 I've been playing wit ionosphere maps and Cesium's MCP controller this week. I wanted to try some quick experiments, but quickly realized I hadn't learned how to hooik connect the Cesium Sandcastle to the their MCP as a client . The method Codex Sol 5.6 and I finally devised and successfully used  is documented below the video. Using that method, I was able to set up the following ionosphere fly-through as seen by a Starlink satellite. More on that soon. Videos Attach the Cesium MCP Camera Bridge to a Sandcastle Map 1. Confirm the camera MCP is running The camera MCP browser bridge listens at: ws://localhost:3002/mcp/ws Check its health from PowerShell: Invoke-RestMethod http://localhost:3002/mcp/health Before attachment, the response should be healthy but may show "connected": false . 2. Open the map in the Sandcastle editor If the map URL uses standalone.html , replace that portion with index.html while preserving the complet...
Post a Comment
Read more
Recent posts

Superconductors as Magnets

 Fun with superconductors on a Monday morning. Did you know that Type II superconductors like the YBCO one shown below can pin magnetic flux? Consequently, if you freeze in the field lines, you've got a superconductor that behaves like a magnet.
Post a Comment
Read more

Lab Book 2026-07-24: Encountering and Hopefully Fixing Subagent Overflow

 I'm still seeing Sonnet fan work out beyond its model capablities using subagents. You can see in the window below that the Gas Town polecat has had to stop and compact its context before proceeding. The good news is that when these subagent work fans don't turn into storms, I'm seeing more thorough reseearch findings returned. The bad news is that with Sonnet's limited context, those findings occasionally overwhelm the polecat that started it all.  I found a preprint from MIT for using what the authors call recurive language models or RLMs. The general idea is basically what I'm seeing with subagents—split the task, split the context—but they have clever ideas about how to manage the output stages to avoid the exact overflows shown above. I'm still harboring my hunch that this all started with the advent of Sonnet 5. Anthropic's release notes include "Sonnet 5 is much more agentic than its predecessors. Testers described how it finishes complex tas...
Post a Comment
Read more

TIL: GasTown Extended git to Offer Rig Setup on first git status

  This is kind of cool! I started a new project today using codex --yolo, (I didn't want to use my Claude tokens until I'd worked out the design and architecture a bit more.) When I got done coding with codex, I asked it to create a repo for the work we did. That seemed to work, but I wanted to verify that the repo was ok, so I typed  gt status and was immediately presented with  I answered y and the extension was off and running adding my new repo to Gas Town. I didn't actually have Gas Town up and running on my Linux client, so I had to get it started and then try again manually with  cd gt && gt rig add ssm_overlay https://github.com/hcarter333/ssm_overlay.git and it just worked!
Post a Comment
Read more

Lab Book 2026-07-20 Claude, Many Agents, and Subagents

 Just noting a feeling, nothing quantitative, but I think I've noticed that if more Claude agents are working at once, they each tend to create fewer subagents. If that's true, and it's a big if, I wonder if there's something at Anthropic that trottles or budgets overall effort to an account. Just to be clear, I've also watched five polecats descend into a complete and utter subagent storm like this one. True or not, I need the storms to either stop or I need to gain the ability to control them, so I'm working more on subagent storm control today. As promised, I did not watch my agents. Intead, I showed the mayor how to watch and mange polecats creating subagents. Here's the start of my conversattion At that stage, I hadn't considered the obvious fact in the figure above that subagents are occasionally beyond recursive. (As an aside, the kid here tells me this is actually a plot point in one of Matrh Well's MrubderBot books.) The next thing I asked...
Post a Comment
Read more

Lab-Book 2026-07-19 How to Get Your Thinking Time back from LLMs

 I've seen a few posts in the last week that talk about engineer's loosing their time to think because they're too busy watching what their LLM agents are doing. Anthropic's models, even under GasTown, have decide to spontaneously fork and find other ways to farm work out to subagents, occasionally causing agent storms on my project in the last week, so I can sympathize a bit more now that I did at the start of the week. There are a few solutions I'd like to propose for this issue. Here goes. How to Get Your Thinking Time Back First, if you're agents are doing marginally what you want, this is pretty simple. Stand up, back away from your desk, and leave . If it makes you feel better, setup a notification so the agent messages you when it's done. Perhaps, something like this If you don't trust your agents for the moment, still don't watch them. Come up with metrics to know if they worked or not and control your agents or modify their activities AUTOM...
Post a Comment
Read more

Lab Book 2026-07-19: Is the Antrhopic Increased Token Deal Related to Fan Experiments?

 Working with Fable last week, I noticed that it can get very excited about fanning out tasks, (to the point of crashing my WSL session.) I experimented with prompting agents to split out tasks on their own, but haven't found a successful to do this yet. Consequnetly, I reverted back to my original prompt which does not explicitly call for task splits. Then, this morning, running on Sonnet 4.6, another task storm sprung up. I wonder if the new 50% higher token limit  till August is Anthropics way of buying themselves some head  room while they experiment with models creating subagents?  Are other  people seeing the same thing? It's particularly worrisome that Sonnet 4.6 has started a subagent storm becuase it really doesn't ahve the context to deal with the results.
Post a Comment
Read more