4 ms·
One of the main reasons I am bearish on Sakana.AI is that for a Japanese-first LLM to work you need to solve a context sensitive language where all 4 alphabets
by jsemrau 2y ago
One of the main reasons I am bearish on Sakana.AI is that for a Japanese-first LLM to work you need to solve a context sensitive language where all 4 alphabets intermingle in a given body of text. So when you are hiring Japanese-speaking PhD's (an already small market), to solve this problem you are moving your attention away from building a great product to optimizing edge cases.
- lmm 2y agoPeople exoticise Japanese a lot and it's not really warranted IMO. 4 scripts do not make AI-related tasks meaningfully more complex, and context awareness is already an important problem that any AI will need to solve.