November 25, 2025
Titanic vibes, Phoenix pay fries
Trillions Spent and Big Software Projects Are Still Failing
Trillions spent, projects still sink — the crowd says AI won’t save the ship
TLDR: Big software keeps failing despite trillions spent, and the article argues AI won’t fix the management chaos behind it. Comments clash: some blame architecture and cuts, others call out Titanic-level hubris, with many warning AI could worsen the mess—proof that leadership, not code, sinks projects.
The article drops a cold splash of reality: despite sky-high IT (information technology) spending—now over $5.6 trillion—big software projects keep faceplanting. And no, AI (artificial intelligence) isn’t swooping in to fix bad management, politics, and wishful thinking. Cue the comment section going full popcorn. One camp, like franktankbank, resists the “doomed from day one” vibe and suggests the infamous Canadian Phoenix pay system might’ve been crippled by architecture and cuts, not destiny. Another crew, led by mdavid626, fires back that AI will actually make things worse, adding “hallucinations” to human delusions. Meanwhile, JohnMakin brings the drama, drawing Titanic parallels: full speed ahead, warnings ignored, and a very expensive iceberg.
There’s spicy debate over whether we need stricter, engineer-style accountability for software, sparked by ZeroConcerns’s “licensed engineers?” jab. History nerd BirAdam chimes in that hardware folks learn from the past while software folks keep reinventing the same mess—cue “failure déjà vu” memes and Chernobyl jokes after the article’s meltdown comparison. The vibe: money isn’t the problem; leadership is. The punchline: Phoenix didn’t rise—it burned paychecks. Commenters turned “mindblowing-units-of-money” into a catchphrase while dunking on magical AI expectations, with the consensus landing on this: it’s not code, it’s choices.
Key Points
- •Global IT spending rose from US $1.7T (2005) to $5.6T (2025, constant dollars), yet software success rates have not significantly improved.
- •The article argues AI tools and coding copilots won’t quickly fix management, governance, and organizational issues driving large IT project failures.
- •Recurring causes include unclear or unrealistic goals, unmanaged risks, complexity, and long-documented, avoidable failure modes.
- •Stephen Andriole’s 2021 framework (Forbes) is cited to show failure factors have been known for decades and repeatedly recur.
- •Canada’s CA $310M Phoenix payroll system (launched April 2016) is presented as a case: extensive customization, vast rule complexity, many integrations, reduced testing and staffing, and aggressive budgeting contributed to failure.