Written by John Marshall
In the fast-evolving world of artificial intelligence, where hype often outpaces reality, a recent incident at OpenAI has underscored the perils of premature proclamations. Kevin Weil, a vice president at the company, took to social media last week to boast about what he described as a groundbreaking achievement by OpenAI’s latest large language model, dubbed GPT-5. In a now-deleted tweet, Weil claimed the AI had “found solutions to 10 (!) previously unsolved Erdős problems and made progress on 11 others,” referring to a set of challenging mathematical conjectures posed by the renowned Hungarian mathematician Paul Erdős.
This assertion quickly drew scrutiny from experts, revealing a gap between the claim and the actual capabilities demonstrated. Thomas Bloom, a mathematician at the University of Manchester who maintains the website erdosproblems.com, publicly rebuked the statement as a “dramatic misrepresentation.” According to Bloom, the AI hadn’t discovered new solutions but had merely regurgitated existing work that already addressed those problems, a far cry from the innovative breakthroughs Weil suggested.
The Hype Machine in Overdrive
The episode, detailed in a report by Futurism, highlights ongoing tensions within the AI industry about transparency and the accurate portrayal of technological advancements. Weil’s tweet was part of a broader promotional push for GPT-5, which OpenAI has positioned as possessing “PhD-level” intelligence. Yet,...
Former Northamptonshire Police Chief Constable Nick Adderley has been charged with fraud and misconduct in public office over allegations he lied about his military service and academic background...