Development

The near-term framing shifts from superintelligence to replication

development2026-04-02

Claim

Dwarkesh Patel framed the near-term implication not as superintelligence but as replication, imagining a thousand copies of a leading mathematician each given millions in inference and pointed at separate Millennium Prize Problems for a hundred subjective years, while Arcee released the strongest open model outside China at 76.3% on GPQA-D.

Reported metric: 76.3%

Reporting date: 2026-04-02 (not necessarily the event date)

Sources

Attribution and reporting linked by this record; links do not imply independent verification.

Welcome to April 2, 2026 (theinnermostloop.substack.com)

Knowledge graphopen in browser →

Reported in

Welcome to April 2, 2026

Actor

Arcee AI

Evidence for

Supersedes

A single-bit 8B model matches full precision in 1.15 GB

Referenced by

Cite or correct this record

Citation formatCitation text

The near-term framing shifts from superintelligence to replication
Innermost Loop KB — secondary record
Reporting date: 2026-04-02
KB permalink: https://nicholsn.github.io/innermost-loop-kb/developments/2026-04-02-clone-terry-tao-a-thousand-times
Claim: Dwarkesh Patel framed the near-term implication not as superintelligence but as replication, imagining a thousand copies of a leading mathematician each given millions in inference and pointed at separate Millennium Prize Problems for a hundred subjective years, while Arcee released the strongest open model outside China at 76.3% on GPQA-D.
Attributed sources:
- Welcome to April 2, 2026 — https://theinnermostloop.substack.com/p/welcome-to-april-2-2026
This citation does not imply independent verification of linked-source claims.

Suggest a correction — opens a GitHub draft for you to review and submit.