jhogendorn

That's shibuya

Yes, you need to design for what jlc stocks as basic parts to retain your sanity. Dont just slap down a 1206 resistor and call it good, jlc prob has no stock in that size in that value, but they probably have it in 0603. Once you get a handle for whats basic or very common extended parts, that sort of thing gets a lot easier. You kind of have to work backwards, ie if i need a diode i see what jlc has in basic and make my part choice accordingly.

Lots of things.

- Got https://beachcomber.sh pretty much stable. Next stage is to propose to various upstreams its worth integrating. - Custom firmware for some ikea symfonisk dials because the oem firmware on them has some pretty bad bugs. Added features like hold and turn. Getting nice smooth dial behaviour over zigbee etc is surprisingly tricky - Built a skill evaluator tool that runs a skill through test suites and then tweaks the skill context and runs again. Its been pretty effective to be honest, almost all skills you do the first version is laughable compared to the one you get after this automated self improvement. - A robust tmux bridge interface for claude to hook into, and then a director layer on top of that for agent orchestration tooling - a stenographer skill that on the fly ripgrep and builds a rag on your on disk conversation history as a form of memory. Pretty effective. - I have just started a tool that brokers woodpecker ci to openbao/vault to give a gitlab like integration for controlled secrets injection for ci. - Been beating my head against a camera tool for a while now, finally making headway. Many ptz cameras dont support fov move, which nvrs need for ml object detection and tracking. They just have a super clunky continuous move and stop. So my tool characterises the camera with cv tools and calibrates movement curves to produce a data file that can be used by my onvif proxy to emulate the more advanced move commands. - Various helper tools for fusion, like csv based parameterised export, and compliant magnet insert generators. - A pipeline that consumes my content backlog, ie instagram saves, reddit saves, hn faves, etc and analyses them with local models and various algorithms steps to categorise and intuit why it was saved and what the key information is and what category it fits into for future reference etc. - A map of my city that shows live river height data with flood map overlays, contour data, predicted overland flow etc. flooding is a regular concern but theres no great resource to know whats going on. I have about 60gb of public datasets it works with. - A package manager for kicad library symbols and footprints, datasheets - skills for kicad so claude can reasonably interpret the schematic and advise on problems, check against datasheets etc. surprisingly effective. - A gcode controlled expansion board for the Carvera Air that gives you 8+8 channels of control for extraction, air assist, vacuum table, timelapse camera, etc. you only have 1 pwm pin so the protocol encodes over that. - A novel exploration interface for vitamins that renders them in a network graph, showing relationships. When you select one it rearranges around it into a kind of valance orbit style so you can explore chains of effect. Turns out, lots and lots of things relate to magnesium. - A comprehensive usb c pd board with 4s battery management. 3a or 8a depending on version. Trying to do proper pd in is nontrivial so this is a drop in solve. - A new brain pcb for Kinesis Advantage Pro keyboards to give modern firmware, bluetooth etc. - Repacked my rack UPS battery with LiFePo cells, and built an induction/resistive series battery balancer pcb for it. - Playing around with a new debug header/connector concept thats tiny footprint and zero cost to add.

The hard part is getting things over the line, publishing and seeing if theres interest. A thing can be largely done but theres a lot of detail work polishing it up so its public ready.

So a lot of this goes away if you try to stick to jlcpcb basic parts from lcsc, and you sort of develop preferred components for things. Another trick is when in a category, for example angular hall sensors, once you have the key attributes like vin, smt, etc, sort by quantity desc to see what parts get heavily stocked, that tends to indicate what is popular in industry and likely to be a good generic fit. Out of the top 10-20 pick according to stock level against price. If youre ordering from jlc, look through lcsc and you can do pcba with them, or get the parts shipped with the pcbs. Otherwise digikey, mouser (good for connectors), farnell, rs etc, roughly in that order. You will quickly learn for a lot of components there are very standard parts everyone uses. Diodes, inductors for example. To some extent if your footprint is already in the jlcpcb library, its popular and use that. Ie theres 1 smd push button that is basic on jlc and has a footprint in kicad default library. Youd be a little crazy to use anything else. Make all your resistors and caps 0603 or 0402, jlc stocks those the most, vs spotty coverage in 0805 or 1206.

But you find each time you find something in a category, a hall sensors, an encoder, a capacitance sensor, gpio expander, adc, you just file that away as kind of solved and next time you hit that part first before finding alts. Over time you have to dig about less and less. But dont bother with things like resistors, caps, etc that are basic. Even mosfets, unless you have special needs the 3 main types jlc has as basic are fine. Dont get some odd package thats technically two mosfets in one, just put down two basic mosfets. And if you do need a more 'advanced' part, look what is on jlc's promo list atm. Theres very popular stm's for example that are prettymuch constantly on promo (promo is they have the part loaded in mid term for some big customer so you might as well piggyback. Means no loading fee. Adds up)

Llm's can be useful for sifting the sand. Narrowed it to like 30 options and still dont know whats the best pick? Export search as csv and ask your favourite stochastic parrot which part would work best for your application and why. Dont ask them for part numbers though, completely pointless.

I built https://beachcomber.sh after one day getting huge lag and asking claude to investigate found that one of the factors was thousands of resources purely for giving my prompt, tmux statusline, nvim statusline, claude statusline the same identical information.

I probably would not have bothered to allocate time to this pre ai, the juice wasnt really worth the squeeze. But I approached it with an initial amusing naivete about it being 'super simple'. As is almost always the case with software theres a reasonable amount of hidden complexity. But I have been using it as a sort of learning proving ground for how to work with agentic development. For example I got to a point where claude wouldnt implement properly and would argue with me about changes because it would read the current/old docs in the repo and get confused about reality. So right now I'm experimenting with 'canonical specs' that can only be changed modally with gates and a defined cascade from canon, to code, to docs in that order. Otherwise you end up in a weird thrash about the docs and the code disagree and which one will the agent decide to change for consistency?

Anyway, its been interesting and its v0.6 and at a point where Ive not hit a sharp edge dogfooding for a while and some beta testers would be valuable. Right now you have to manually wire it into your stack, once some others have kicked the tyres hard enough I will make some pr's to the popular tools to consider integrating it.

If you like that, you should check out my project https://beachcomber.sh . Its about time I take it from dogfooding to beta users if you want to give it a go.

As far as I understand it, the movies were actually very successful, but there was some kind of internal politicking or studio issue and they just didnt continue.

Compaction wont save you, in fact calling compaction will eat about 3-5x the cold cache cost in usage ive found.

I dont believe the fixes offered by boris et al are genuine. What i can get done on my max plan has gone down by about 80%. I have often seen my usage meter sit at 3-7% before a single message has been sent that session, and ive seen simple fresh prompts jump usage 15-20%. Theres clearly some fundamental underlying issue with the usage monitoring and billing platform.

If you hold command while you restore the window from the dock it will restore into the current space. I wish i knew how to make this the default. Getting whisked away to a random space is one of the most irritating issues. Like when you want a finder window for downloads and instead of a new window in the current space you get taken cross country to on thats already open.

I wonder if you can buy a replacement headband yet, since they have a lifespan of about a year before the mesh collapses. Which probably happens because they are so god-damn heavy.

I really wish the audio technica headband style was more common, even they use it less now. The two winglets as the head contact are just so much more comfortable than literally any other configuration.

Has anyone done a teardown and figured out where all the gravity in these things comes from?

I saw an artist say recently on insta reels that if their fanbase switched to Apple Music it would go from beer money to more than their day job. And apparently even more from Tidal. They acknowledged that spotify is the elephant in the room with 80% of their audience on it.

Inventree or homebox are what you want.

Keep in mind that POTS is specifically heart rate and not blood pressure. Blood pressure drop is orthostatic hypotension. They can both exists or one or the other, but are not the same thing and POTS is a specific term not an umbrella one.

Try subsai on github. I recently used it to bulk subscribe and subtitle many hours of video of an accented speaker talking in a speciality domain with the appropriate jargon and i would say using the openai api model for whisper it was about 95-97% accurate. Super impressive.

The web interface has an 'update all' button thats just as convenient. I find if theres ones i want to not update i just temp break their yaml file with an unexpected keyword and it fails to compile and then update.

I wrote an integration to make this available as part of kroki, if anyone wants to use this as a service.

As a person with EDS, i would counter that it is way more common than realised. As its a systemic genetic driven inability to produce collagen correctly, the only cure will be when we can rewrite and organisms genetics wholesale, which we are not exactly close to.

Treatments are poor, the condition itself and its systemic effects are almost unstudied. Most doctors either dont know about it or their knowledge stops at 'flexible'. Some doctors (surgeons esp) may even decline to work with you if you disclose EDS.

CSF leak, either from a lumbar puncture, or from things like chiari malformation and other cervical abnormalities are a well known phenomenon within the community. A person complaining of persistent postural headaches usually gets csf leak as the first response from said community. It is however, incredibly difficult to get a doctor to be willing to diagnose it, to have it competently found, and then patched.

Yeah thats not what happened, if we are referring to the same incident.

He had had a close friend pass recently and he had, within the last few weeks, used mushrooms to help treat the depression/grief. Ie he did not crash the plane mid psychotic break slash psychadelic trip.

Huh, didnt realise they had one to use. Every time we have tried to pay with card in japan it has been either impossible or a huge song and dance, even involving the old click clack slide impression carbon paper tools.

How do you make clones of a yubikey? Is this normal functionality I've missed or a hack?

Would fusionauth meet all these needs?

Yes, agree totally because thats me and a few other core people. As a very much self starter it can be hard to build systems that are enough support but not too much for those who aren't as comfortable or 'entitled' as you put it.

We absolutely run into this issue of new people floating ideas and the old hats going 'for the 100th time that will not work because x y z' and the new person going 'jesus, ok then'. Theres a lot of history and learnings thats hard to solidify and pass on and get people into the same book, not even the same page. I still haven't cracked this nut, it feels like i should write a bunch of documentation but i suspect it will not be read.

I've been the polite but chronic reminder/nagger. It does work but its extremely wearing on all participants. I've found the best way to do it is to pair or more people up, in a physical location at the same time to do things. Body doubling is like crack to my audhd brain but it works great for most people as well. Bonus points if you feed them. However, thats a high level of organisation and it can be really tricky to pull it off with the regularity and scale the organisation might require.

I've said in other comments i'd do unseemly things for a volunteer coordinator, but i'd make the devil blush if i could pin down a project manager. Most people just need structure, thats just the long and short of it I've found. You've gotta provide it some way or another, and doing it effectively is a full role and I have enough hats as it is.

How do you combat the old knowledge vs new (naive?) ideas issue? Do you try to reevaluate that 'old knowledge'? Just let the new idea down gently and explain the history? Talk about what you havent tried that might be relevant?

Huh, I had not considered that failing their own expectations of what they can volunteer might be a factor in leaving. I'm going to have to think on that for a bit.

We definitely caution against taking on too much at first, or we give them clear deescalation paths should they struggle.

Hey, hey don't call me out like that. I'm the problem, its me. I absolutely have a tendency to do most of the work but want to run/control things tightly. Always pushing back on myself to try and chill out.

What do you think the main reasons volunteers quit are?

I recently had a new person, strong neurodivergent who struggled greatly with the fact we were/are an organisation with many flaws we are working on. They came to me after their first week with a 9 point plan to fix volunteering. Point 1: Get a volunteer coordinator. I felt like hank scorpio palm to forehead, a volunteer coordinator, why didn't I think of that? We were of course super aware of it, nearly impossible role to fill. I'd give a limb for one almost.

Burnout is _the_ monster that I'm trying to fight, more people, smaller workloads. Predicated on getting more people however.

Setting tasks is... well you learn that you have to train people how to do it. And how SMART Goals are one thing, and actionable tasks are another (subset thing). If you write the task as a Goal, when it comes time to assess if its done, you play the interpretation game and it gets a bit rough.

I think skill building is a big factor that, while we talk about it, we should highlight it more front and center.

Recognition is definitely powerful, as I said in another comment, the issue I've found with it is that if you miss just one person, they tend to feel much worse than if you thank no-one. Hard to catch everyone in that net.

We have been talking about gamification ideas lately, letting volunteers get some form of karma point for doing certain tasks, letting volunteers 'fist bump' each other to tip karma as thanks for a job. Then its a self reporting issue and harder to be left out. Still chewing the idea over, feels like capitalism with extra steps.

Volunteers are definitely hard. If i ever get a coordinator i may nail them to the floorboards.

How did you organise your charity? heirarchical, groups by area or topic? how did they work together?

Do you have any tips on how to get people to do the boring jobs. "Please update the asset register for your area with the msds for anything you keep in stock" 2 hour job. Like pulling teeth, but i need it for compliance.

We have learnt these lessons the hard way, repeatedly, until maybe some of it has sunk in.

I've lost count of how many times i've pointed out that you cannot put good and bad actions on a scale and try to balance them. Good actions are good in one pile, and bad actions are bad in their own pile, and that pile has to stay real, real small. Conflict avoidant people who are new (to you, ie me the 10yr veteran vs the 3yr 'newbie') really, really struggle with 'ostracising'.

Our code of conduct is pretty average tbqh. Do you have any examples of ones that you would consider exemplary or have particularly interesting points or approaches?

Great points, my thoughts:

  * Direct positive results, absolutely yes. The problem I've found here is that if you are 95% good at this, you are still failing and someone is very upset they got missed/jilted, and you're none the wiser until they take some action based on that.
  * Our environment we pitch as 'volunteering is a great way to pick up new skills'. Do you think theres value in 'gamification' ie, issuing badges or collectable somethings to indicate skills gained? We cant officially certify you as a book-keeper but we can teach you how to operate xero for example. Useful as kudos, maybe marginally as a resume padder.
  * We do a lot of this, it works okay-ish. It's hard to get 'full compliance', ie you give a volunteer ownership of an area, theres fun parts and not fun parts and weirdly enough the not fun parts get missed every month...
  * Historically we've been terrible at letting go but we are getting better and better. The geek social fallacies are very very strong.
  * This last point feels like going to the dentist. I know i need to floss more <insert excuse>. You're 100% right though, hyperfocusing on the onboarding and training should probably be one of our highest priorities because its making every other fire downstream bigger.

Are you able to provide examples of how you've done these points in your organisation from an implementation perspective?

These are all great, although they're more concepts than implementable things. I see a lot of mirrors in my own experiences for sure.

Our volunteers seem to fall into:

  * Obsessive Crazy People (myself): Very committed, for no discernable reason. Organisation survives because a few people are like this at any time.
  * I Want Control: Someone wants to exercise their will or way of doing things over an area or similar. Put a bridle on them and lets go
  * I Enjoy Tinkering: I like to help out and scratch my itches, but I wont commit to anything specifically and its mostly about fun and following my personal interests.
  * I Will Volunteer for X: A discount, kibble, free soft drink. Extrinsicaly motivated by some external factor or gamification.
  * Saints: I just want to help and give back.

So I see some overlap in your descriptions there. Attrition and Job handover is super tricky. Or really, formalising the training of things is a whole big job in itself, so often people get thrown into things and told to swim. Not great, but as a firefighter you probably know sometimes its hard to know which fire to fight...

Do you have any suggestions re sophisticated collaborative structures?

Would it be viable to just state that you are hard of hearing and require text based communications? While its not the strict truth, its a much simpler and more understood 'hard' disability, whereas when you state a disability that is perceived as 'soft' (ie any given mental issue neurotypical people assume you can just think through/get over) theres more pushback on you to normalise yourself rather than for them to accommodate.