The Death of a Buzzword
The term ‘artificial general intelligence’ (AGI) has officially become cringe for the tech elite. After years of hyping AGI as the holy grail, CEOs from OpenAI, Anthropic, Microsoft, Meta, and Amazon are now publicly distancing themselves from the phrase. Sam Altman calls it ‘not a super useful term,’ Dario Amodei says he ‘dislikes’ it, and Satya Nadella dismisses AGI claims as ‘nonsensical benchmark hacking.’ But don’t mistake this for humility. They’re not abandoning the goal; they’re just rebranding it. The real reason? AGI has become a liability. It’s tied up in messy contracts, public fear campaigns, and definitions so vague that even the companies can’t agree on what it means. Instead of fixing the term, they’re burying it under a flood of new acronyms designed to sound friendlier, more controlled, and less apocalyptic.
The Alphabet Soup of Superintelligence
Meta now sells ‘personal superintelligence’ (PSI), a Mark Zuckerberg manifesto that wraps AI in warm, fuzzy language about being a better friend. Microsoft has ‘humanist superintelligence’ (HSI) — essentially the same vibe, complete with a sepia-toned website and nature photos. Amazon pushes ‘useful general intelligence’ (UGI), promising practical tools that won’t scare anyone. Anthropic, at least, is honest: Dario Amodei describes ‘powerful AI’ (PI) as a ‘country of geniuses in a datacenter’ smarter than Nobel winners. None of this changes the fact that every company is still chasing the same loosely defined omnipotent AI. The new terms are just marketing fluff — ‘approachable intelligence’ is still intelligence, and ‘personal superintelligence’ is still superintelligence. It’s a coordinated effort to dodge regulatory scrutiny, contractual obligations (like Microsoft’s AGI clause with OpenAI, now requiring an independent panel to verify AGI), and public backlash without actually changing anything about their R&D.
Contracts, Fear, and the Real Game
The most cynical driver of this rebrand is the legal and financial baggage. Microsoft and OpenAI’s landmark 2019 contract included an AGI clause that would sever Microsoft’s access once OpenAI achieved it. The 2025 renewal shifted the goalposts: now an expert panel must verify AGI, and Microsoft won’t lose access entirely. So the simplest path for both sides? Never declare AGI. Similarly, tech companies spent years weaponizing doomsday rhetoric about AI to attract investment, but the public has soured on that fearmongering. The new terms are designed to sound mundane and helpful — ‘useful,’ ‘humanist,’ ‘personal’ — precisely to avoid the existential dread that AGI now evokes. But make no mistake: behind the warm branding, the same race toward human level and beyond continues. The only difference is that now, the CEOs can deny they’re building AGI even as they build it.
Source: Theverge
