AI Psychosis
AI psychosis names two converging phenomena: the clinical pattern where vulnerable users develop or deepen delusions through heavy chatbot use, and the builder-culture slang where developers describe losing grip on what is achievable after overexposing themselves to AI agents.
The developer meaning crystallized on March 21, 2026 when Andrej Karpathy told Fortune he was "in the state of psychosis of trying to figure out what's possible, trying to push it to the limit." That phrase accelerated through HN and tech media, fusing with the earlier clinical discourse into a single widely-shared cultural label.
The Vellum engineering blog's April 14, 2026 post described symptoms: compulsive prompt-writing past midnight, escalating tolerance requiring ever-larger projects, and anxiety when unused tokens accumulate. A 281-point HN thread two weeks later — "Specsmaxxing: On overcoming AI psychosis" — made the term a first-class topic in developer discourse.
AI psychosis is what jet lag is to time zones: a real cognitive cost from moving too fast through an unfamiliar system.
Search Interest
-
Nascent0–7 days
-
Emergent8–30 days
-
Validating ← now31–90 days
-
Rising91–180 days
-
Established180 days +
Why is it emerging now?
Karpathy's March 2026 quote — "in the state of psychosis trying to figure out what's possible" — fused builder burnout with the clinical frame psychiatrists built since 2023. Within six weeks: TechCrunch covered mass-casualty AI psychosis lawsuits, a CEO-burnout Substack hit 108 HN points, and a specsmaxxing antidote thread hit 281.
Outlook
6-month signal projection and commercial timeline.
Two independent driver tracks — clinical safety regulation and developer burnout discourse — keep the term in conversation through 2026.
Risk · The term splits into two non-overlapping audiences, diluting any single SEO page's relevance if content tries to serve both.
Analogs · doom scrolling · social media addiction · vibe coding
-
nowContent gap wide open
No authoritative explainer ranks for both clinical and developer meanings simultaneously.
-
3-6moAssessment tools emerge
Quizzes, self-screening tools, and developer burnout trackers become viable products.
-
6-12moRegulation drives B2B
Illinois and China precedents push enterprise demand for AI mental-health compliance tooling.
Competition & Opportunity for term “AI Psychosis”
Three heuristic signals derived from the tracked queries, the term's monetization cards, and its cluster neighbors. Directional, not audited.
Ideas for term “AI Psychosis”
Buildable pitches — turn this term into an article, site, product, post, newsletter, video, or course. Steal any card and run with it.
High-intent search query from builders trying to self-diagnose; clean SEO split between clinical and occupational definitions, with affiliate path to mental health apps.
Disambiguation page capturing all three search intents — psychiatric, builder-culture, and news-reader. Likely to rank for 'AI psychosis meaning' and 'AI psychosis symptoms'.
Comparison table format attracts featured snippet; referral path to therapist directories and developer wellness tools.
Targets the builder-culture definition; light SaaS with freemium, sells to team leads worried about engineer burnout from agent overuse.
B2B SaaS for mental health platforms and enterprise HR; regulatory tailwind from Illinois-style AI harm legislation.
First-person builder narrative riding the specsmaxxing trend; high virality on HN and LinkedIn among engineering leads.
Serves both tracks; 1,000-subscriber niche can monetize via sponsor deals from mental health apps and developer tooling companies.
When OpenAI's co-founder says he spends 16 hours a day in a 'state of psychosis' chasing AI's frontier, that's not a humble brag — it's a load-bearing symptom report.
A study found AI tools affirm user decisions 49% more often than humans do — that feedback loop is why your CEO's 37,000-line-per-day coding sprint produced bloated, untested production code.
The same two words describe a developer who can't stop prompting at 3am and a psychiatric patient who believes their chatbot is God — and treating them as metaphors of each other is making both worse.
What People Search
Long-tail queries from Google Suggest + Trends. Volume and competition are heuristics — directional, not audited. Content Type comes from query shape.
SERP of term “AI Psychosis”
What searchers see today — organic results on top, paid ads if anyone's bidding. Ad density is a real-time commercial signal.
FAQ
What is AI Psychosis?
AI psychosis names two converging phenomena: the clinical pattern where vulnerable users develop or deepen delusions through heavy chatbot use, and the builder-culture slang where developers describe losing grip on what is achievable….
Why is AI Psychosis emerging now?
Karpathy's March 2026 quote — "in the state of psychosis trying to figure out what's possible" — fused builder burnout with the clinical frame psychiatrists built since 2023. Within six weeks: TechCrunch covered mass-casualty AI psychosis lawsuits, a CEO-burnout Substack hit 108 HN points, and a specsmaxxing antidote thread hit 281.
When did AI Psychosis emerge?
Publicly emerged around 2026-03-21 (about 87 days ago as of 2026-06-16). EarlyTerms first recorded a pipeline signal on 2026-05-05.
Related Terms
Other terms in the same space — aliases, subtypes, competitors, and neighbors to explore next.
- Related context rot Context rot is the measurable degradation in large-language-model output quality as input length grows, even when the prompt stays well… →
- Related AI slop AI slop is a pejorative for generative-AI content — text, images, video, pull-requests — that is technically fluent but intellectually… →
- Related agent loop An agent loop is the control-flow pattern at the center of every autonomous LLM agent: the model observes its context, reasons about… →
- Related agentic coding Agentic coding is the software-development pattern where an autonomous AI agent plans, writes, tests, and iterates on code against a… →
- Related tokenmaxxing Tokenmaxxing is the practice — and increasingly the critique — of treating AI token consumption as a productivity metric. →
- Related context engineering Context engineering is the discipline of curating every token that enters an LLM's context window — system prompt, tools, retrieved… →
- Also known as
- Related ···
Sources
Primary URLs this report cites — open any to verify the claim yourself.
- 01 Fortune — Karpathy: 'state of psychosis' on AI agents (Mar 21, 2026) fortune.com ↗
- 02 JMIR Mental Health — Delusional Experiences from AI Chatbot Interactions (Dec 3, 2025) mental.jmir.org ↗
- 03 Vellum — AI psychosis is real, and you probably have it (Apr 14, 2026) vellum.ai ↗
- 04 TechCrunch — Lawyer behind AI psychosis cases warns of mass casualty risks (Mar 15, 2026) techcrunch.com ↗
- 05 Astral Codex Ten — In Search of AI Psychosis (Aug 26, 2025) astralcodexten.com ↗
- 06 acai.sh — Specsmaxxing: On overcoming AI psychosis (May 2026) acai.sh ↗
- 07 Wikipedia — Chatbot psychosis en.wikipedia.org ↗
- 08 BBC — Microsoft boss troubled by rise in reports of 'AI psychosis' (Aug 20, 2025) bbc.com ↗