On June 29, Meta introduced a system that reconstructs typed text from brain activity recordings without surgical intervention, achieving an average word-level accuracy of 61% compared to about 8% with previous non-invasive methods. Brain2Qwerty v2 was trained on approximately 22,000 sentences, and the experiment involved nine healthy volunteers, each spending 10 hours in a magnetoencephalography scanner while typing heard phrases on a keyboard. Meta stated, “Instead of handcrafted schemes to identify neural events, we use end-to-end deep learning to decode data directly from raw brain signals.” The system works with continuous brain activity recordings and reconstructs entire sentences, using a fine-tuned language model that considers semantic context. The best participant achieved 78% word-level accuracy, with more than half of their sentences decoded with one or no errors. The development is not yet ready for clinical or everyday use, as Brain2Qwerty v2 was tested on healthy volunteers who were actually typing text, and for individuals who have lost speech or movement abilities, this approach needs separate testing. The system does not yet function as a full real-time text input, as the model processes entire sentences. The project is linked to the Digital Brain Project, an initiative to create open neuroscience datasets with a budget of $5 million.