Making AI Headlines · Page 3
OpenAI Enters the Market for Short Answers
· New York time
OpenAI has made its Decisions API widely available, entering a market for models that quickly choose among options and return a confidence rating. The competition includes TypeSafe's Jev, which helped establish the category of specialized models sometimes called super classifiers or System One models. Developers have found that these focused models can replace large language models in some workflows while operating faster and at much lower cost, according to TechCrunch. TypeSafe, founded by a former frontier-lab researcher, has reached a $7.6 billion valuation. There is apparently considerable money in teaching a computer not to turn every choice into an essay. The appeal is narrower work performed efficiently. Applications that need repeated decisions may not require the full conversational range of a general-purpose model. A confidence rating adds useful information, although confidence remains something to assess rather than a decorative certificate of correctness. OpenAI's entry makes this a contest over a less visible layer of software intelligence. Instead of greeting users with paragraphs, the model may sit behind an application making focused selections. The chatbot has discovered a back-office career: fewer speeches, more decisions, and a performance review increasingly concerned with how much each answer costs.
Read this story in the full newspaper →Satire. Not factual reporting. Read the original reporting from the linked publishers.