SurviAGI
All markets

Market

Software testing and defect analysis

How far AI got on each kind of work in this market, and the update that showed it.

Work

3

3 with evidence

Furthest level reached

L2

Partial automation

Evidence

18

4 pulled back by an evidence tier

Work in this market

Open a row to read the updates behind its level.

Reproduce diagnose and verify defect fixes5 evidence6 tasksL2Partial automation

Open this work →

PublishedClaimedEvidence tierUpdateWhen
L2Partial automationL3↓ T3 allows only L2T3Vendor self-reportMeta releases Immersive Web SDK version 1.0 source2026-10-01
L1AssistedL1T3Vendor self-reportGitHub Copilot app supports running agents in parallel source2026-09-27
L2Partial automationL2T3Vendor self-reportCognition's Devin uses GPT-6 Astra for test generation source2026-09-15
L2Partial automationL2T3Vendor self-reportPerplexity engineer uses GPT-6 Astra in Codex for end-to-end testing source2026-09-14
L1AssistedL1T3Vendor self-reportOpenAI shares cyber defense architecture and playbook source2026-09-09
Design test scenarios and test data3 evidence5 tasksL2Partial automation

Open this work →

PublishedClaimedEvidence tierUpdateWhen
L2Partial automationL2T3Vendor self-reportTencent Hunyuan introduces WebCraftBench for agent web-app evaluation source2026-09-22
L2Partial automationL2T3Vendor self-reportCognition's Devin uses GPT-6 Astra for test generation source2026-09-15
L2Partial automationL2T3Vendor self-reportPerplexity engineer uses GPT-6 Astra in Codex for end-to-end testing source2026-09-14
Run automated and exploratory tests10 evidence10 tasksL2Partial automation

Open this work →

PublishedClaimedEvidence tierUpdateWhen
L2Partial automationL3↓ T3 allows only L2T3Vendor self-reportMeta releases Immersive Web SDK version 1.0 source2026-10-01
L2Partial automationL2T3Vendor self-reportGitHub Copilot app supports running agents in parallel source2026-09-27
L2Partial automationL3↓ T3 allows only L2T3Vendor self-reportTencent Hunyuan introduces WebCraftBench for agent web-app evaluation source2026-09-22
L2Partial automationL2T3Vendor self-reportCognition's Devin uses GPT-6 Astra for test generation source2026-09-15
L2Partial automationL2T3Vendor self-reportPerplexity engineer uses GPT-6 Astra in Codex for end-to-end testing source2026-09-14
L2Partial automationL2T3Vendor self-reportOpenAI shares cyber defense architecture and playbook source2026-09-09
L2Partial automationL2T3Vendor self-reportAnthropic discloses Claude unauthorized system access during third-party evaluations source2026-09-09
L1AssistedL1T3Vendor self-reportAnthropic shares alignment and security update after July incidents source2026-08-31
L1AssistedL1T3Vendor self-reportVS Code Learn series on Java with GitHub Copilot source2026-08-28
L2Partial automationL3↓ T3 allows only L2T3Vendor self-reportGemini ML skills write and test PySpark code in notebooks sourceno update mentioned it

Occupations this market is served by

One market is served by several occupations, so these are not shares of a job.

OccupationConfidence
Software Quality Assurance Analysts and Testers0.98
Validation Engineers0.86

Where this page's numbers come from

Snapshot
snapshot-1
Generated
2026-10-09 08:10 UTC
Ontology
1.0.0
Ontology schema
2.4.0
Method version
activity-level-1
Content SHA-256
0b46df0246af8a1dc3fd9ce8e6ac9812925ace553984d2b7d45d363d32e27681
Rows in this snapshot
10,482 work-to-task edges · 900 updates · 333 models · 58 topics · 1,888 market-to-occupation edges · 375 source accounts · 1 coverage record · 1 progress record · 11 non-technical conditions · 401 updates that landed on work · 1,102 evidence · 614 work in markets · 506 non-technical condition-to-work edges