X @The Wall Street Journal

The vulnerabilities of large language models such as Grok can be addressed. This should be done right now, as they gain the ability to do things as well as generate language, writes our columnist. https://t.co/LcaT85KeRs ...