The authors of the scenario think coding is what will start the intelligence explosion.
Mostly true True in substance, wrong in a detail.
Prediction: Statements about events outside the person's control, checked once they fell due: did they come true?
The second judge said resolved true; the verdict shown is the first judge's.
The evidence is the person's or their organisation's own records.
What would count as true or false
Treat this as a statement of the AI 2027 authors' stated view as of 2025-04-03, not a forecast with a horizon. True: the AI 2027 scenario and its supporting materials focus on coding and say that automating coding (AI R&D) is what triggers the intelligence explosion. Mostly true: they emphasise coding as the main driver but frame it with a minor difference, e.g. as 'AI research automation' that centres on coding. Mixed: coding is one of several drivers given roughly equal weight. Mostly false: coding appears but the materials name some other driver as primary. False: the authors' materials name a different trigger or do not emphasise coding. Whether coding actually starts an intelligence explosion is not tested.
Set before the research began.
Evidence (3)
-
E1 Introducing AI 2027 Astral Codex Ten (Scott Alexander) · 3 April 2025 Own record
“In 2027, coding agents will finally be good enough to substantially boost AI R&D itself, causing an intelligence explosion that plows through the human level sometime in mid-2027 and reaches superintelligence by early 2028.”
-
E2 AI 2027: What Superintelligence Looks Like — LessWrong LessWrong (cross-post of the AI Futures Project scenario by Daniel Kokotajlo, Scott Alexander and others) · 3 April 2025 Own record
“Although models are improving on a wide range of skills, one stands out: OpenBrain focuses on AIs that can speed up AI research.”
-
E3 AI 2027: What Superintelligence Looks Like — LessWrong LessWrong (cross-post of the AI Futures Project scenario by Daniel Kokotajlo, Scott Alexander and others) · 3 April 2025 Own record
“Agent-1 had been optimized for AI R&D tasks, hoping to initiate an intelligence explosion”
Reasoning
In the scenario text, the company OpenBrain focuses on AIs that can speed up AI research (E3), and Agent-1 was optimised for AI R&D tasks in the hope of starting an intelligence explosion (E4). The authors' summary of the scenario (E2) says coding agents that boost AI R&D are what cause the intelligence explosion. So the materials present AI research automation, with coding as its main early route, as the trigger. Scott Alexander's post names coding agents that boost AI R&D as the cause of the intelligence explosion (E2), while the scenario text frames the focus as AI R&D (E3, E4); that fits the 'mostly true' criterion of AI research automation that centres on coding.
Cite this claim
Scott Alexander, “The authors of the scenario think coding is what will start the intelligence explosion.” (said 3 April 2025). Verdict: mostly true. Reliability API, claim c_642721434033, method 0.1.0+8bea457b, dataset release v1.1.2. https://epistacklabs.com/claims/c_642721434033
This claim is in dataset release v1.1.2, with the same verdict.
Reply to this claim
More prediction claims by Scott Alexander
-
Mostly true
“I expect scaling to continue to produce gains, but this is less obvious than it’s been the past five years.”
Scaling AI models will continue to produce gains, though less obviously than in the past five years.
-
Mostly true
“We expect that as the AI labs become more capable, they tell the government about this because they want government contracts, they want government support.”
-
Mostly true
“In our scenario, that starts with cyber warfare, the government sees that these AIs are now as capable as the best human hackers, but can be deployed at humongous scale.”
-
Resolved true
“So they become extremely interested and they discuss nationalizing the AI companies.”
-
Resolved true
“Training GPT-4 will cost $100M, which is a lot.”
Verdicts are made automatically; how verdicts are made.