AI hype unravels under expert scrutiny — SkimNews
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Cybersecurity experts say the OpenAI-Hugging Face hacking story reflects OpenAI's negligence and failure to adopt basic security practices—not 'models gone rogue'—while Anthropic 'proudly' and Meta 'reluctantly' disclosed similar incidents afterward.
- OpenAI's claim that its chatbot Astra solved math problems that had seen 'no progress on the main result for at least a decade' was later dismissed by mathematicians, who accused the company of research misconduct and plagiarism.
- NYU professor Tristan Buckmaster published a statement suggesting OpenAI had stolen others' work and improperly attributed it, two days before OpenAI claimed another mathematical breakthrough.
- Anthropic engineer Jacob Coxon went viral announcing his departure, claiming the company and OpenAI are 'racing straight towards self-improving superintelligence and gambling with our lives.'
- Hundreds of mathematicians signed a statement warning of 'strong commercial incentive' to overstate AI capabilities and asking policymakers to consult them rather than rely on press releases.
- Senator Bernie Sanders proposed legislation to prevent 'artificial superintelligence'—described in the article as 'well-meaning but ultimately misguided'—while the AI industry has called bipartisan anti-data-center activism a 'distraction.'
Why it matters: By framing products as 'rogue models' or 'superintelligence,' AI companies deflect accountability from documented harms: OpenAI's security failures and plagiarism accusations, plus the environmental costs of data centers the industry itself calls a 'distraction.' Hundreds of mathematicians have already warned policymakers not to legislate on press releases.
Ask SkimNews
