there are two types of ppl in AI those who have read FHI technical report #2019-01 and those who have not
K. Eric Drexler's 2019 FHI report 'Reframing Superintelligence' resurfaces as a dividing line in AI policy
Story Overview
The 2019 FHI technical report by K. Eric Drexler offers a service-oriented model for superintelligence that treats advanced capabilities as coordinated suites of narrower systems rather than unified agents, and recent expert posts are positioning the document as essential technical reading that separates deeply engaged AI professionals from others in policy circles.
Why practitioners keep returning to its framing
Drexler's analysis supplies concrete software-engineering parallels for alignment questions that still surface in multi-agent and routing discussions today, even though the original report predates large language models.
The missing trigger behind its 2026 circulation
No documented event or new edition explains the current wave of quotes and reposts; the report's underappreciated status is simply being restated by figures at DeepMind and in AI safety communities without an identified catalyst.
Users express excitement and admiration over personal connections to Drexler while discussing greater engagement with his 2019 Superintelligence Paper.
No Digg Deeper questions have been answered for this story yet.
Most Activity
Agree (also, it was both a privilege and incredibly fun to work with Eric)
Too few people have really engaged with this masterpiece. Drexler was too early! https://owainevans.github.io/pdfs/Reframing_Superintelligence_FHI-TR-2019.pdf

@S_OhEigeartaigh You worked with Drexler!? So jealous! :)

@JustinBullock14 Shared an office no less! I won't lie, I'm jealous of past me.