Run step
Step 4: tool results
get_claim_instances
Input
{
"claim_id": "54c25eab-5263-467c-880b-ee0f2c3bdd12"
}{
"claim_id": "54c25eab-5263-467c-880b-ee0f2c3bdd12",
"total": 1,
"limit": 20,
"offset": 0,
"has_more": false,
"instances": [
{
"id": "e3a13307-e488-43b8-8d66-f0dd27bc2035",
"source_id": "aca3a46b-00d2-4dd6-bcc0-43f4d7eb3320",
"verbatim_text": "Anthropic co-founder Jack Clark has moved his forecast for when AI could fully improve itself to 2027 from 2028, pointing partly to a new test of research judgment.",
"context": "Report of Clark's 12 October 2026 Import AI post, which favours 2027 for automated AI research able to train a successor; no updated numerical odds were given. Near the claim rather than identical: it concerns full self-improvement, not capability short of agenda-setting.",
"stance": "affirms",
"confidence": 0.55,
"speaker": "Jack Clark (as reported)",
"publication": "The Rundown",
"source_date": "2026-10",
"link": null,
"source_title": "Anthropic’s Jack Clark moves AI self-improvement forecast to 2027",
"source_type": "news_article",
"source_url": "https://www.therundown.ai/news/jack-clark-ai-self-improvement-forecast-2027",
"source_published_date": "2026-10-15",
"source_authors": [
"The Rundown Editorial Team"
],
"source_publisher": "The Rundown AI",
"source_doi": null,
"source_interest": null,
"recorded_by": "claim_steward",
"display_rank": null,
"reading": null,
"draws_on": [],
"story": {
"standing": "untraced",
"basis": "none",
"depth": 0,
"diverges": false,
"downstream_total": 0
}
}
]
}query_sql
Input
{
"sql": "select n.id, n.kind, left(n.canonical_form,170) as text, a.status, a.claim_credence, a.assessed_at::date as d from nodes n left join assessments a on a.claim_id=n.id and a.is_current where n.id in (select claim_id from tags where slug in ('ai-research-automation','ai-development-forecasting')) and n.state='active' order by n.importance desc limit 80",
"max_rows": 80
}{"success":false,"code":"SQL_ERROR","message":"column a.claim_id does not exist. The sql skill lists every view's columns."}search_similar_claims
Input
{
"limit": 12,
"query": "AI agents can carry out long open-ended AI research projects end to end without human help"
}{
"query": "AI agents can carry out long open-ended AI research projects end to end without human help",
"results": [
{
"id": "d91e5ad8-8ffd-4fb2-8507-ee86b736c75f",
"kind": "claim",
"text": "Evidence on whether AI agents can carry out open-ended AI research is thin.",
"claim_type": "evaluative",
"state": "active",
"similarity_score": 0.7211863218068374,
"importance": 0.35,
"assessment_status": "supported",
"assessment_confidence": 0.7
},
{
"id": "72ac75a9-6f75-45b9-997c-205f081655f1",
"kind": "claim",
"text": "AI agents outperform human experts on short AI research-engineering tasks but underperform them on longer tasks.",
"claim_type": "empirical_verifiable",
"state": "active",
"similarity_score": 0.6632501624451773,
"importance": 0.35,
"assessment_status": "supported",
"assessment_confidence": 0.65
},
{
"id": "66eddda8-c9d1-4b53-8ed6-2a6055c68756",
"kind": "claim",
"text": "AI agents attempting open-ended research exhibit recurring failures in research judgment, redesign, backtracking, resource awareness, and instruction following.",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.6459398970811148,
"importance": 0.4,
"assessment_status": "supported",
"assessment_confidence": 0.7
},
{
"id": "78d6c0a0-a6d8-4b57-89eb-f193be0c0e0e",
"kind": "claim",
"text": "AI agent evaluations built around narrow, verifiable tasks do not assess open-ended research ability.",
"claim_type": "evaluative",
"state": "active",
"similarity_score": 0.6418124234372925,
"importance": 0.4,
"assessment_status": "supported",
"assessment_confidence": 0.7
},
{
"id": "27a33dd0-42b0-48e8-aead-a838e382a200",
"kind": "claim",
"text": "As of 2026, frontier AI agents can autonomously complete the engineering work of AI research projects.",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.6414255408657892,
"importance": 0.6,
"assessment_status": "supported",
"assessment_confidence": 0.5
},
{
"id": "6d318e34-dfe2-4f86-a5d8-d7f71179125d",
"kind": "claim",
"text": "Current evaluations poorly measure AI agents' ability to conduct open-ended research.",
"claim_type": "evaluative",
"state": "active",
"similarity_score": 0.6390475737230732,
"importance": 0.4,
"assessment_status": "supported",
"assessment_confidence": 0.7
},
{
"id": "c14e4dbe-a52f-40cb-bf38-4667b68afd85",
"kind": "quantity",
"text": "AI-agent work performed at OpenAI in August 2026, measured in human workdays per human workday",
"claim_type": "empirical_verifiable",
"state": "active",
"similarity_score": 0.6305086631429498,
"importance": 0.4,
"assessment_status": null,
"assessment_confidence": null
},
{
"id": "7064aebc-8735-45b1-9ee7-523e919cb15a",
"kind": "claim",
"text": "Frontier AI agents exhibit recurring failures in research judgment, design adaptation, backtracking, resource awareness, and instruction following during open-ended research.",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.6247372028504105,
"importance": 0.45,
"assessment_status": "supported",
"assessment_confidence": 0.7
},
{
"id": "81eeb61b-4ce4-4d0b-891d-11e970d8f669",
"kind": "claim",
"text": "As of 2026, frontier AI agents can perform research engineering but struggle with critical stages of the AI research lifecycle.",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.6129052683963658,
"importance": 0.6,
"assessment_status": "supported",
"assessment_confidence": 0.7
},
{
"id": "d4fdf446-31c5-4eaa-add8-feb9e2bdfb02",
"kind": "claim",
"text": "Current AI systems can originate radically novel research ideas.",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.6070783361232177,
"importance": 0.5,
"assessment_status": null,
"assessment_confidence": null
},
{
"id": "ceb833ca-a085-4b45-aba3-8bd3d58add9c",
"kind": "claim",
"text": "The ability to sustain autonomous work on long tasks is the main capability bottleneck preventing AI from automating AI research.",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.6040030710441096,
"importance": 0.45,
"assessment_status": "doubtful",
"assessment_confidence": 0.6
},
{
"id": "54c25eab-5263-467c-880b-ee0f2c3bdd12",
"kind": "claim",
"text": "By end of 2027, Anthropic or OpenAI will have AI capable of performing all frontier AI research work except agenda-setting.",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.5883881869664624,
"importance": 0.4,
"assessment_status": "inconclusive",
"assessment_confidence": 0.6
}
],
"count": 12
}search_similar_claims
Input
{
"limit": 10,
"query": "lack of research judgment or taste will slow remaining automation of AI research"
}{
"query": "lack of research judgment or taste will slow remaining automation of AI research",
"results": [
{
"id": "55b84cc2-c3ba-4168-ae4b-723dafc6b353",
"kind": "claim",
"text": "Lack of research judgment will slow automation of remaining AI R&D tasks.",
"claim_type": "causal",
"state": "active",
"similarity_score": 0.8751674751264789,
"importance": 0.65,
"assessment_status": null,
"assessment_confidence": null
},
{
"id": "e6e8a4fa-eee4-4fc3-b564-ea04a6313546",
"kind": "claim",
"text": "Progress toward automating AI research and development is likely to stall in the next few years.",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.6742476157566103,
"importance": 0.7,
"assessment_status": "doubtful",
"assessment_confidence": 0.65
},
{
"id": "6be5b967-9784-4985-8413-928665429119",
"kind": "claim",
"text": "AI development is moving toward substantial automation while humans continue setting research directions and judging results.",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.6497517483673381,
"importance": 0.6,
"assessment_status": "supported",
"assessment_confidence": 0.6
},
{
"id": "3699eb33-020b-43bd-8da6-469f19575f06",
"kind": "claim",
"text": "Partial automation of AI R&D tasks will substantially speed AI software progress before full automation.",
"claim_type": "causal",
"state": "active",
"similarity_score": 0.6493037381489963,
"importance": 0.7,
"assessment_status": "supported",
"assessment_confidence": 0.65
},
{
"id": "d479cf4d-52b6-4260-a3dc-7a2e5a812d42",
"kind": "claim",
"text": "As of late 2026, humans supply most research judgment in frontier AI R&D.",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.6380278172934556,
"importance": 0.6,
"assessment_status": null,
"assessment_confidence": null
},
{
"id": "24c8ff90-2b07-40b9-9d37-e6d66c0617d3",
"kind": "claim",
"text": "Data scarcity or poor data quality can constrain AI automation despite abundant compute.",
"claim_type": "causal",
"state": "active",
"similarity_score": 0.6350298526593284,
"importance": 0.45,
"assessment_status": "supported",
"assessment_confidence": 0.75
},
{
"id": "94831dd7-b838-416f-9851-1a652cf58861",
"kind": "claim",
"text": "Limited compute for experiments would be the binding constraint on AI progress even if AI research labor were fully automated.",
"claim_type": "causal",
"state": "active",
"similarity_score": 0.6346148829191904,
"importance": 0.6,
"assessment_status": "inconclusive",
"assessment_confidence": 0.6
},
{
"id": "9352bf05-815c-455e-9f52-ecb4cc40cdbf",
"kind": "claim",
"text": "AI systems will eventually develop research taste comparable to skilled human AI researchers.",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.6312565706463593,
"importance": 0.75,
"assessment_status": null,
"assessment_confidence": null
},
{
"id": "e03e6ac0-7bf3-44df-bae8-e8822164f268",
"kind": "claim",
"text": "Automated AI researchers could substantially increase research progress per unit of compute through better experiment selection.",
"claim_type": "causal",
"state": "active",
"similarity_score": 0.6279186104706923,
"importance": 0.4,
"assessment_status": "supported",
"assessment_confidence": 0.6
},
{
"id": "d7359f07-7b36-4da2-a0d7-816dbfb8831f",
"kind": "claim",
"text": "Uncertainty about future AI automation capabilities can reduce investment in AI.",
"claim_type": "causal",
"state": "active",
"similarity_score": 0.6265246128818954,
"importance": 0.4,
"assessment_status": null,
"assessment_confidence": null
}
],
"count": 10
}search_similar_claims
Input
{
"limit": 10,
"query": "year a superhuman AI researcher / automated AI researcher is first achieved at leading company"
}{
"query": "year a superhuman AI researcher / automated AI researcher is first achieved at leading company",
"results": [
{
"id": "aa36d48a-c988-4991-a701-50102794ef74",
"kind": "quantity",
"text": "Calendar year in which AI systems first fully automate AI research and development",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.7144801638459685,
"importance": 0.85,
"assessment_status": null,
"assessment_confidence": null
},
{
"id": "f8a3bbad-417d-46e7-91b8-7424e0806b9d",
"kind": "quantity",
"text": "Calendar year when an AI company first internally develops an AI that can do any coding task its best engineer can do, faster and cheaper",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.696177238777687,
"importance": 0.75,
"assessment_status": null,
"assessment_confidence": null
},
{
"id": "e083ea2a-d43e-4156-8cca-395ac22ef6b5",
"kind": "quantity",
"text": "Calendar year when AI systems first automate 99% of AI research and development tasks",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.6955424118724923,
"importance": 0.8,
"assessment_status": null,
"assessment_confidence": null
},
{
"id": "a6037281-5051-4d1f-bd95-5546955a242d",
"kind": "quantity",
"text": "Date when a frontier AI company first has an AI matching top researchers' research taste and fully automating coding",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.6839091874728208,
"importance": 0.8,
"assessment_status": null,
"assessment_confidence": null
},
{
"id": "cb0cfaf6-af14-493f-8fc3-0d571ecb3fc9",
"kind": "quantity",
"text": "Calendar year by which AI systems first perform most AI research and development tasks",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.6739408930072628,
"importance": 0.75,
"assessment_status": null,
"assessment_confidence": null
},
{
"id": "eefadc66-321b-459a-b09d-b3e69fb10daf",
"kind": "quantity",
"text": "Calendar year in which artificial general intelligence is first developed",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.657078461587432,
"importance": 0.85,
"assessment_status": null,
"assessment_confidence": null
},
{
"id": "ed22320f-8175-4a31-9e65-e24717cd614f",
"kind": "claim",
"text": "Absent a stall in AI progress or deliberate slowdown, superhuman AI researchers will arrive before 2036.",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.649930067276687,
"importance": 0.8,
"assessment_status": null,
"assessment_confidence": null
},
{
"id": "200278b7-b70d-4549-8ab4-da228093c781",
"kind": "claim",
"text": "OpenAI achieved an automated AI research intern by September 2026.",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.6246209340729951,
"importance": 0.45,
"assessment_status": null,
"assessment_confidence": null
},
{
"id": "79562451-dd54-42e1-ba95-047a4e9341fb",
"kind": "quantity",
"text": "Elapsed time, in years, from AI being able to mostly perform frontier AI research-engineer work to AI surpassing humans at all cognitive tasks",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.6152402880103516,
"importance": 0.7,
"assessment_status": null,
"assessment_confidence": null
},
{
"id": "6c1f5ff8-8110-42e0-bde5-161eac34836f",
"kind": "claim",
"text": "AI systems will attain superhuman coding ability by 2027.",
"claim_type": "empirical_derived",
"state": "active",
"similarity_score": 0.5982646269362525,
"importance": 0.6,
"assessment_status": "doubtful",
"assessment_confidence": 0.6
}
],
"count": 10
}