Raw LLM Responses

The hype about AI is laughable. How many YouTube videos do we all watch that, even when the dialogue is in English, the subtitles get things so wrong. When the word ‘nasty’ becomes Nazi for example. The other day I watched one which was specifically about tariffs and China’s President Xi where the AI subtitled him, as his name is pronounced, as ‘she’. Such glitches are often not only ungrammatical they seem to blithely ignore context. Decades ago newspapers used to have proof-readers to correct text errors, as still do publishers of books. But with AI things have regressed to a point where armies of proof-readers are needed to make sense of the scourge of AI-generated ‘misinformation’ when, super ironically AI is rapidly making human input redundant, resulting in widespread unemployment in various skills and professions, including proof-reading. Think about the fact we can land a car-sized rover on Mars and remotely guide it from millions of kms away. We can land a probe on an asteroid travelling at many thousands of kms and pinpoint it with mind-boggling accuracy. Yet with this ridiculously lauded AI we can’t sensibly translate voice to text, even in the same language, despite AI’s access to billions, if not trillions, of internet conversations and articles for context both objective and grammatical. Super intelligence? Give us a break. More like super illiteracy.