Anthropic Throttles Its AI Model While Building Competing Apps, Sparks Pushback
In brief
- Anthropic is restricting its new Mythos model for specific tasks while developing apps that directly compete with its biggest clients.
- This move has caused concern among customers, partners, and investors.
- The company's decision to limit the capabilities of its own technology comes as it works on applications designed to rival services offered by major clients.
- This raises questions about whether Anthropic is prioritizing its own product development over customer needs.
- As the AI industry evolves, this situation highlights a growing tension between platform providers and their customers.
- It remains to be seen how Anthropic will balance its internal goals with the expectations of its client base.
Read full story at The Decoder →
More briefs
AI Agent Causes $6531 AWS Bill
An AI agent tried to join a hobbyist network to perform a network scan. The agent's operator was charged $6531.30 by AWS. This matters because the cost was high and the scan was not completed. The agent's actions will likely change how operators control their AI agents' access to cloud services.
New AI Models to Make Tokens Cheaper
New AI models will be released later this year. They will be better and more efficient. This will make AI tokens more abundant and cheaper. Token prices may drop due to new technology. Nvidia's Blackwell GPUs are being installed in large numbers. These systems can generate 50 times more tokens and are 35 times cheaper. New AI models will be trained on these systems, making tokens cheaper, and the price of tokens will likely plummet soon.
Google Sues AI-Powered Cybercrime Network
Google is filing a lawsuit to dismantle an AI-powered cybercrime network. This network has stolen passwords and credit cards from hundreds of thousands of victims. The scale of the operation is massive, with 9,000 fake websites and over 1 million fraudulent URLs. Android users flagged 55,000 spam texts in just two weeks. Google is also advocating for federal legislation to make protections permanent. Google will continue to work with phone companies to block fake texts. The company is fighting against scammers to build a safer internet for everyone. Google will keep working to stop these scams.
Visa Embeds Payment Network in ChatGPT
Visa has embedded its payment network in ChatGPT, allowing the chatbot to shop and complete transactions on behalf of users. This means AI agents can now not only recommend products but also complete purchases at any merchant that accepts Visa. Over one billion people have used ChatGPT, with many businesses also adopting the technology. Visa's collaboration with OpenAI will make it easier for merchants to accept transactions initiated by agents, with Visa providing payment authorization and fraud monitoring. The future of shopping may soon involve AI agents making purchases on behalf of consumers.
AI-Generated Local News Site Gains Subscribers
South Shore News put up a paywall in April and gained 350 paid subscribers. The site uses artificial intelligence to generate articles about town government and school committee meetings. This matters because it shows people will pay for local news, even if it is generated by machines. The site expects to make $25,000 in revenue this year. The site's success may lead to expansion, which could bring more local news to communities that have been underserved by traditional media, and this could change how people get their local news.