AI labeled overhyped
A string of social posts argues that much of what’s sold as artificial intelligence is marketing rather than genuine intelligence — one account, Economía De Barrio, said algorithms are being repackaged for profit rather than invention. (x.com) The thread has low engagement but repeats a sampling of skepticism that’s circulating among tech-watchers today. (x.com)
A new round of posts is reviving an old complaint: much of what companies call “artificial intelligence” is still pattern-matching software wrapped in bigger promises. (x.com) The account Economía De Barrio said in a recent post that firms are “repackaging” algorithms as invention, not delivering genuine intelligence; the post had low engagement but echoed a wider skeptical line already circulating in tech coverage in March and April 2026. (x.com) Artificial intelligence, in plain terms, is software trained on large piles of data to predict the next word, image patch, or action. Nature wrote this month that current systems can look fluent while still operating more like high-speed calculators than independent minds. (nature.com) That argument has gained traction as the industry keeps shipping new models while also retiring old ones on tight schedules. OpenAI’s developer documentation says GPT-4.5-preview was deprecated on April 14, 2025 and removed from the application programming interface on July 14, 2025, while ChatGPT model retirements continued into 2026. (developers.openai.com) Anthropic has kept the release cycle moving too. The company announced Claude 3.7 Sonnet on February 24, 2025 as a “hybrid reasoning” model and later introduced Claude 4, presenting both as steps toward stronger coding and multi-step problem solving. (anthropic.com 1) (anthropic.com 2) The skepticism is not limited to social media. MIT Technology Review described 2025 as a “hype correction” for artificial intelligence after two years of inflated expectations around large language models, the text-generating systems behind chatbots such as ChatGPT and Claude. (technologyreview.com) The spending behind those expectations remains enormous. The Financial Times reported in July 2025 that the race for “superintelligence” was driving a global data-center buildout even as critics questioned the cost, environmental impact, and whether all that capacity was necessary. (ft.com) AI companies reject the idea that the field is only branding. OpenAI says its mission is to build “artificial general intelligence,” and Anthropic says its latest Claude releases improve coding, tool use, and long-running tasks for business customers. (openai.com) (anthropic.com) What the posts capture is less a new backlash than a durable split in how the technology is described: researchers and executives keep publishing new benchmarks, while critics keep asking whether prediction engines should be sold as intelligence at all. (nature.com) (technologyreview.com)
