AI Mania Is Eviscerating Global Decision‑Making – A Consultant’s Front‑Row View

TL;DR – AI hype is destroying rational decision‑making

The author, a consultant who has run sales and technical work for dozens of companies, reports that every AI project he has observed has failed, and that the resulting mania forces executives to publicly profess belief in AI or risk being fired. This creates a coordination problem that stalls sensible governance and drives wasteful spending.

1. AI Investments Deliver Little to No Value

"All of the AI projects we have observed as a team are failing. Every single one – we have seen 0 % success in a year and a half…"

Zero‑success claim – Across more than a year and a half of engagements, the author’s team saw no AI project deliver measurable productivity gains. The failure rate is not just high; it is total for the sample they examined.
Mis‑reporting is systemic – Executives who overstate AI benefits are quickly removed; honest employees are often laid off. Boards, vendors, and consultants all have incentives to obfuscate true outcomes.
Typical failures – Internal chatbots see almost no adoption because documentation is poor and LLMs are not omniscient. Customer‑facing bots often look polished but never resolve issues (e.g., a Mitsubishi support bot that promised a callback that never arrived).
Root causes – Most failures stem from poor software‑project execution, not the LLMs themselves. Companies lack the discipline to ship reliable software, and the novelty of AI adds extra risk.

"Even if some companies are seeing clear gains, this is the exception, not the norm."

2. Dissent Is Punished – “Heretics Will Be Shot”

"People who raise the possibility that AI might not be the solution are outright dangerous to their careers."

Mandatory AI evangelism – In firms with 500+ employees, executives are forced to make religious‑style proclamations about AI, often without having used the technology themselves.
Career‑risk for skeptics – The author witnessed high‑performers being fired for achieving results without LLMs. Expressing doubt can lead to termination or being labeled a liar.
AI‑wash – Engineers lie about using LLMs to protect their jobs (e.g., claiming a Go rewrite was done by Claude). Token‑usage leaderboards become gamed metrics.
Commentary – A software engineer on the thread confirmed this behavior:

"Checking out a parallel copy of our Go repository and telling the AI to rewrite the whole thing in Zig while I work on something else just so I can keep my job."

3. Flashy Demos Create Irresistible Buying Frenzy

"When we demonstrated Snowflake’s Cortex chatbot, every lukewarm client turned into a hot buyer."

Demo‑driven hype – Even a half‑working natural‑language query demo can trigger instant purchasing pressure, regardless of actual suitability.
Ethical restraint – The author’s team declined to exploit this frenzy, removing the demo from future pitches.
Market effect – Vendors with weak AI expertise can still sell because customers have no better reference points; a mediocre demo appears superior to nothing.

4. Executives Face a Prisoner‑Dilemma

"If an executive admits AI gains are implausible, they risk being seen as a heretic and losing their contract."

Game‑theoretic stalemate – Executives must either co‑operate (admit the truth) and risk being fired, or defect (double‑down on hype) and preserve their position.
Board anxiety – Board members admit skepticism but feel their jobs depend on AI investment, leading to a self‑reinforcing cycle of false claims.
Coordination failure – No mechanism exists for all leaders to simultaneously reveal the truth, so the status‑quo persists.

5. AI‑Native Is Now a Hiring Requirement

"Every large organisation now demands AI alignment even when the value proposition is ambiguous."

AI‑as‑a‑checkbox – Projects are retro‑fitted with a token‑usage phase to satisfy internal politics, even when the core work is a standard database migration.
Token‑leaderboards – Employees are judged on AI spend; those who cannot justify AI usage are labeled “bad at AI” and risk termination.
Hiring policies – Some firms require staff to prove they tried AI before requesting headcount; failure to do so can lead to dismissal.

6. Surviving the Crisis – Practical Advice

When You Have Another Objective

Raise issues privately – Avoid group discussions about AI projects; use one‑on‑ones to protect sources.
Anonymous polls – Ask team members to rate project success anonymously; a bimodal split often reveals hidden failure.
Involve end‑users – Ground truth comes from the people who actually use the tools; many are unaware they even have AI licenses.
Don’t challenge the mantra – If someone says “AI is changing everything,” let it slide unless you’re trying to fix a specific problem.
Build trust before confronting – Private meals and personal rapport are more effective than public confrontation.

When You’re Just Trying to Survive

Accept limited influence – Recognize that pushing back is often futile; the problem is organizational, not technical.
Consider contracting – Contractors earn more and stay out of internal politics, with a clear end‑date.
Limit AI news consumption – Reduce exposure to hype to preserve mental health.
Feign agreement when safe – Nod to AI usage if it won’t cause harm; focus on personal well‑being.
Prepare an exit strategy – Treat every large AI‑driven workload as a signal to start a job search.

7. Community Reactions – Highlights from Hacker News

Agreement on burnout – A commenter echoed the author’s warning about reviewing massive AI codebases and burning out.
Critique of hyperbole – Several users (e.g., @A1kmm, @jdw64) argued that the 0 % success claim is an exaggeration and that many modest AI uses (semantic search, regression) do improve productivity.
Selection‑bias concern – @kbar13 and @Aurornis noted that the author’s consulting firm likely attracts struggling clients, inflating the failure rate.
Real‑world counterexamples – @LOCNE55 shared personal success using Claude for advanced SQL and Python, suggesting pockets of genuine value.
Cultural analogy – @zkmon compared AI mania to historic “manias” (tulip, agile, timesheets), emphasizing that such hype cycles are inevitable.

8. Final Thought

The AI frenzy has turned many organizations into cult‑like ecosystems where dissent is punished, metrics are gamed, and genuine productivity gains are rare. While the bubble will eventually deflate, the damage to decision‑making structures may linger, echoing past tech manias. For now, the safest path is to protect oneself, document reality discreetly, and prepare to move when the hype inevitably wanes.