OpenAI has rolled out a stripped-down version of the ChatGPT-4o large language model. Dubbed GPT-4o Mini, it is claimed to have better accuracy than GPT-4 on tasks, and costs substantially less than GPT-3.5 Turbo. Zdnet.com reports that OpenAI is claiming that the new AI model is “the most cost-efficient small model in the market.” It is worth noting that there aren’t any parameters defining large or small models, so this may just be puffing. At any rate, GPT-4o Mini is priced at 15 cents per million input tokens and 60 cents per million output tokens…vastly more affordable than the previous models and 60% cheaper than GPT-3.5 Turbo. The model only offers text and image support right now, with audio and video expected to be added at a later date. Its training data is current through October 2023. 

Meta has looked at investing billions in eyewear giant EssilorLuxottica. According to theverge.com, the blast of cash would be in furtherance of Meta’s partnership with the owner of Ray-Ban and numerous other eyewear brands. The Ray-Ban Meta smart glasses released last year now support multimodal AI to identify what wearers are seeing. They also sold more in a few months than the previous pair did in two years, according to EssilorLuxottica’s CEO.

Amazon is teaming up with the Better Business Bureau to fight fake review brokers…starting off with a lawsuit against a firm called ReviewServiceUSA.com. Geekwire.com says the suit claims that Review Services  allegedly facilitates the selling of fake positive reviews for products on Amazon listings or Better Business Bureau profile pages. Amazon said it blocked more than 250 million suspected fake reviews in 2023.

We had reported yesterday that Apple was among some firms that used its OpenELM model to train AI on YouTube Subtitles. Now, 9to5mac.com reports that Apple says that’s not true. Cupertino claims that the subtitles were only used for research purposes, and that the YouTube subtitles were not used to power Apple Intelligence. Apple has said that Apple Intelligence models were trained “on licensed data, including data selected to enhance specific features, as well as publicly available data collected by our web-crawler.” Apple also stated that it has no plans to build new versions of the OpenELM model. 

I’m Clark Reid and you’re ‘Technified’ for now.