Software and published archives are experimental. Code might contain bugs. Archives published should not be assumed to be authenticated, atleast for now.
Why?
Ethereum blobdata is the most censorship storage solution on Earth, while also being affordable enough for daily use. Anything published to blobdata gets immediately mirrored to thousands of ethereum nodes across multiple nuclear-armed states.
Ethereum blobdata is stored for approx 18 days by default. For longer-term storage, blobscan providers are maintaining IPFS archives, but availability of these depend on how many pin the archives (this is similar to torrent seeding).
I wanted to experiment with this in a use case with non-trivial stakes. I care about my writings being censorship-resistant, even if future Samuel faces political or other pressure to take the writings down.
Also, other people with even higher stakes use cases than me can now use my public repo, or build based on it.
How?
I used gpt-6-astra to write an automated script to publish an archive of my website to ethereum blobdata. It splits a tarball containing only markdown files (no images, pdf, videos, etc) into chunks and publishes them.
ERC7588 metadata format consumes too much calldata gas, so I didn't use it. I invented a custom metadata format instead, that is human-readable enough to decode without requiring a proper spec document, and yet efficient in terms of calldata gas used. This format is documented in the calldata of the first tx of each published archive.
Subscribe
To subscribe, enter "Subscribe " followed by Signal username, email address or whatsapp phone number. You will receive atmost one update per month.
Alternatively, you can subscribe via RSS, or a third-party service like blogtrottr