Gary Marcus argues current AI remains controllable and below singularity
Marcus argued that current systems remain shutdown-capable, complete only a minority of his proposed AGI tasks, and still need human scaffolding, distinguishing a controlled OpenAI cyber drill from spontaneous loss of control.
0 comments · 0 votes
Sign in to join the discussion →
No comments yet. Start the discussion.
Why it moved the index
Specific task limits, retained shutdown capability, and dependence on human-granted permissions provide direct counterevidence to present-day loss-of-control claims, with uncertainty because several examples are interpretive.
Assessment history
-
R1
Away 25 · confidence 70
New dated first-person analysis supplies distinct task-limit and control arguments not present in the durable evidence ledger.
13 Aug 2026
Share this page
-
DoomBench assesses “Gary Marcus argues current AI remains controllable and below singularity” as evidence moving away from doom, with magnitude 25 and confidence 70 out of 100 in the human resilience category.
-
The DoomBench assessment of “Gary Marcus argues current AI remains controllable and below singularity” is based on reporting from Gary Marcus and records the editorial rationale, source quality, attribution, and revision history.
-
DoomBench summarizes “Gary Marcus argues current AI remains controllable and below singularity” as follows: Marcus argued that current systems remain shutdown-capable, complete only a minority of his proposed AGI tasks, and still need...
https://www.doombench.com/news/gary-marcus-argues-current-ai-remains-controllable-and-below-singularity-2026-07-28