What counts as AI-related?

A post is AI-related when it explicitly mentions AI, ML, LLMs, agents, RAG, embeddings, or known AI infrastructure terms.

What is not proven by this report?

It does not prove total market demand. It measures visible startup hiring language on HN.

AI Hiring Reality Report | StartupJobList

Name: StartupJobList AI Hiring Reality Dataset
Creator: StartupJobList
License: https://www.startupjoblist.com/terms

AI signal by month

This separates AI-related language from explicit AI/ML role families, because one sentence about an LLM does not make the job an AI role.

Month	Posts	AI related	AI share	AI/ML roles	Salary posts
2026-08	239	158	66.1%	149	62
2026-07	270	158	58.5%	153	62
2026-06	318	194	61%	191	74
2026-05	327	181	55.4%	173	66
2026-04	334	199	59.6%	198	73
2026-03	322	188	58.4%	183	59
2026-02	415	239	57.6%	235	86
2026-01	258	142	55%	140	36
2025-12	380	202	53.2%	197	71
2025-11	394	229	58.1%	225	71
2025-10	341	187	54.8%	187	46
2025-09	300	153	51%	152	48
2025-08	307	163	53.1%	160	54
2025-07	333	160	48.1%	166	60
2025-06	361	173	47.9%	170	59
2025-05	301	138	45.9%	140	54
2025-04	308	154	50%	150	39
2025-03	354	149	42.1%	153	40
2025-02	396	157	39.7%	151	49
2025-01	276	114	41.3%	119	35
2024-12	330	127	38.5%	132	39
2024-11	318	134	42.1%	131	46
2024-10	330	110	33.3%	108	36
2024-09	304	111	36.5%	112	36

Explicit AI infrastructure demand

Methodology and data dictionary

Source: public Hacker News Who is Hiring top-level comments. Demand means share of classified hiring posts in this corpus, not the whole labor market and not confirmed hires. Raw comment text is stored privately for reproducible parsing, but public exports contain derived fields, HN IDs where relevant, and source URLs only.

Sample gates: hide salary medians below 20 salary observations, keep weak change claims marked as insufficient sample, and show counts next to every table. The taxonomy and parser versions are included in every export row.

CSV checksum: c7ecc4f6a77dcc4cc5251f124a041273ef79e4ce4fad25f396c3d981f587135d