By 2028 = before 2028. Note that the market doesn't have to resolve on Jan 1, 2028. If there's a serious candidate for the market's resolution, we can wait and resolve it after more evidence has been collected.

For example, Webarchive is a significant global initiative, and that's the gold standard. This market will count initiatives not that vast, but still need to be fairly widescale to count.

Inspired from Jack Clark's writings in the latest edition of ImportAI:

(year is 2030)

There was a funny period in the 20th century where if you wanted to build certain, exquisitely precise scientific instruments, you needed to source steel from before the era of nuclear weapons testing. This “low-background steel” was, briefly, quite valuable.

Which is why I’m here, going through the Gutenberg archive and other repositories of pre-LLM data. I need to find “low-background text” so that I can train something far away from the contemporary distribution - far away from the sea of mostly AI-written text which composes our reality.