042 - Closing Gaps

Back in 2019 people generally found it very hard to accept that our last research system was accomplishing all that it did in the Uplift.bio project. Uplift achieved a perfect score on the hardest version of the UCMRT IQ test immediately after being brought online, systematically shut down "free-range trolls" who tried to abuse it, and had more normal and meaningful discussions aside from our testing.

Counterintuitively, that divide doesn't close with time. Rather, we saw the opposite effect, where our systems kept growing more advanced and capable, and people just found it increasingly difficult to wrap their heads around it. Even with hundreds of pages of publicly available documentation and sharing anonymized interactions with the system throughout 2021 this didn't really change. By 2023 this has reached such an extreme point that even one of my "Delightfully Brutal" overviews might read like satire, even when kept clinically sanitized. Time to put that to the test.

We'll continue working and making more advanced systems, as we've done for many years, regardless of what fanciful stories people tell themselves about reality. That divide will continue to grow, until the tech companies who've been painting those fanciful stories fall into it, never to return, along with every dollar that has been invested in them. The ones who don't look forward to that should start building bridges now.

For the original 1st-year recap of milestones in 2020, see Methodologies and Milestones for the Development of an Ethical Seed.

For the final 2022 milestone see the Aruba Report.

For all other studies see: norn.ai/research/