Anthropic will have to defend a proposed class action alleging that its marketing overstated the value of Claude Max, its $100- and $200-per-month subscription plans. Two subscribers say the promise of “5x or 20x more usage” led customers to expect five or 20 times as much use over a month compared with lower-priced plans. The suit says the multipliers apply only within each five-hour session, while weekly usage caps are not increased by the same amounts. One plaintiff, computer-science student Solomon Smith, says he used Claude for coding and other work.
A federal judge in California ruled that the subscribers’ interpretation of the marketing was plausible and allowed claims of deceptive advertising, negligent misrepresentation and breach of contract to proceed. The judge rejected Anthropic’s argument that reasonable customers would understand the figures to refer only to per-session use, finding that relevant website disclosures were too far removed from the main sales pages to resolve the dispute at this stage. Other claims, including one under California’s unfair competition law, were dismissed, but the plaintiffs may amend their complaint. The case can now move into evidence-gathering, and the proposed class action seeks to represent thousands of Claude subscribers.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
Researchers tested 16 LLMs on all 55 TIS-100 puzzles, giving each one attempt and time to refine its code. The models found 133 new performance tradeoffs across 18 puzzles and set three outright cycle records, though human players improved on most solutions within days. (tis-100.wincon.dev)
Australia is considering a new criminal offence for AI companies that fail to take reasonable steps to prevent harm caused by their models. The government is also weighing an AI safety watchdog with power to impose significant fines, following a hack of a Medicare statistics portal involving OpenAI. (abc.net.au)
AI is increasingly becoming a gateway through which people find information. Audiences still resist having AI produce their news. (niemanlab.org)
Nvidia holds a central position in the AI industry, making disruption to the company potentially catastrophic for the wider ecosystem. The race for advanced AI is also intensifying economic and geopolitical conflict, while concentrated corporate power and infrastructure face growing political resistance. (france24.com)
LLM inference can require reading about 26 GB of model weights to generate each token from a 13-billion-parameter model. The explanation shows how model size drives memory bandwidth demands during text generation. (youtube.com)
Deletion Tuning matches full fine-tuning across three LLM post-training methods by zeroing fewer than 0.05% of model weights. Its compact masks also preserve prior-task accuracy better than full fine-tuning, LoRA, and random sparse tuning. (arxiv.org)
Today's best Hacker News stories, summarized and screenshotted, one email a day.