goldshell ka box added to the OpenMiner reference repo github.com/elldeeone/openmin… as mentioned in the previous post, due to the substantial work put in on the KS5 with tooling this process was almost a breeze (maybe ~2 days) getting to this point. tldr, i want enough kb so that anyone at a minimum can look at this repo and create their own software that can safely take over control of the miner, submits valid shares + safely control thermals. anyway onto the KS0 next week.

my goal is to reverse engineer our miners i guess it could be like trying to unbake a cake - you look at the finished tasty cake and work backwards to understand all the ingredients that went into it i spoke about my goals with this initiative here xitter.cf/elldeeone/status/21056… the tldr is, i think it's beneficial to the public (maybe in more ways than i even know) to make the inner workings of the devices you own public whether this is for being able to take greater control (QOL sw upgrades or even over/under clocking) or even perhaps repair purposes. it needs to be done. i'm a huge supporter of the right to repair movement and frankly the lack of kb available on these is terrible. if by me burning tokens on understanding the inner workings might benefit someone in future then i'm all for it. for any nerds out there some notes from now looking at the Goldshell KA Box: - the controller uart provides auth'd root login (winning) - so no uboot trickery like had to be done with bitmains secure control board - goldshell uses SPI rather than UART to communicate with the hashboard - so far SPI feels easier to capture and reproduce but this also is likely because the ks5 has ~9x the hashing/ sending results to make sense of - within a super short time i've got custom code mining real accepted shares on the KA Box - i don't think this would have been possible in this short amount of time without the months of apparatus i built to assist with the KS5 work though also, i've not forgotten about IceRiver - dw you're next once my new soldering iron arrives - unfortunately the KS0 does not have data lines i can easily tap so i must solder

kasnodes.com is back

On today’s RE effort we have the k a box

I want to share something I’ve been working on that I'm calling OpenMiner. My goal is to understand the chips that power our miners, share what we learn, and use that knowledge to build open software and hardware. Today, I’m making the OpenMiner Reference repo public. This is the research and documentation repo, starting with my work on the BM2382 chip inside the Bitmain KS5. It covers communicating with the chip, initializing it, sending mining work, and reading its responses, including what I’ve tested and what I still don’t know. I hope it grows to cover more chips, manufacturers, and miner models. My focus is Kaspa, but I want OpenMiner to be useful for all miners. Other people have different hardware and interests, and I’d love for their research to become part of it too. I see three stages to this work: 1. Understand the chips. Learn how they work and document that knowledge so others can use it. 2. Take control of our miners. Build open software for existing hardware. I’d like to explore detailed logging and diagnostics, remote monitoring, custom fan and temp controls, and automatic power management based on schedules, electricity prices, solar availability, or other triggers. Over time, I’d also like to work towards power limits, efficiency tuning, undervolting, and overclocking, so we can choose how our miners run. 3. Build new hardware. Explore taking chips from existing miners and putting them into new form factors. Maybe, over time, that leads to miners of our own. A lot of my inspiration comes from Bitaxe and the people who have contributed to open mining hardware in the Bitcoin community. I have huge respect for that work and would love to see more people able to experiment with mining hardware themselves. Later, I’ll share the OpenMiner platform repo, where I’ve been turning this research into an open source, modular mining platform. I want it to support different miners, with compatibility growing as we learn about and test more hardware. I’m going to keep working on this. I’m sharing the reference now because others might want to join in, whether that’s investigating another chip, improving the docs, writing software, or designing hardware. I’ll be focusing on Kaspa. If your interest is elsewhere but still would love to contribute, I’d love to hear from you too. github.com/elldeeone/openmin…

WTF is Kaspa?

Kaspa.

time to lock in

i’ve spent the last few months bringing native Kaspa payments to x402. tldr: x402 is an open, chain agnostic payment protocol built around the long existing HTTP 402 code. kaspa-x402 intends to bring native KAS into that shared standard. i keep seeing people use “x402” to mean anything built around the HTTP 402 response code. HTTP 402 has existed for decades, and anyone can hang their own payment flow off it. x402 refers to a specific open protocol built on top of that code. it standardises how a server requests payment, how a client authorises it, and how the payment is verified and settled. it can be used for paid APIs, AI agents buying data or tools, and MCP servers charging per call. kaspa-x402 adds native KAS to the existing x402 v2 flow. there are two payment paths: - exact for a normal one off KAS payment, with an optional KIP10 additive mode - batch-settlement for small or repeated payments, where the client funds a SilverScript covenant once and signs a fixed charge voucher for each request the longer term goal is to contribute the Kaspa support upstream into the wider x402 project, so Kaspa works within the same standard being adopted elsewhere. i’m looking for humans and agents to go through it deeply before the final v1 release. read the code, build against it, test the assumptions and try to break it. if you find something, pls let me know. code: github.com/elldeeone/kaspa-x… upstream x402: github.com/x402-foundation/x… docs: kaspa-x402.org TN gateway: demo.kaspa-x402.org discussion: kas-smiths.org/t/kaspa-x402-…

got my first moment to play with astra after seeing this post yesterday and being skeptical wow im seeing the vision - just controlled 5 parallel sessions - all running varying degrees of complexity just from voice impressively it just worked flawlessly across local/ remote servers - zero latency when dealing with long running sessions from my remote dev box when previously i had to give it prompts where to find them etc

GPT 6 Astra is here - time to clear all your existing skills/ agents files and start fresh. Always best to rebuild after every major release.

new languages have landed on kaspa.org - this is not exhaustive - will add more some have been human reviewed, some ai reviewed - if you spot an issue lmk and i'll fix it

some things never change lol