Language is hard. Particularly if you have a foreign speaker. Among the challenges, so many exceptions to the rules. However, a few languages are more consistent. Programming language, for example. That’s why LLMs are so darn good. Another comparatively consistent language: Law.
The impact of LLMs on the legal profession is a hot source of discussion and disruption. Consider the following:
“What if we told you that the LLM (large language model) Claude can consistently decide Supreme Court cases correctly when you take just 10 seconds to upload merit briefs, without any complex prompt engineering or hand-holding, and produces detailed and well-reasoned decisions in less than a minute. “
“Not only can Claude reproduce simple case law, it can produce completely novel new legal standards. And remember, this is the least powerful AI that will ever exist and AI will be 5-10x more powerful in a matter of years. “
The attached article is really detailed, but endeavors to prove the point.
Michael J. Silva, Edward Caja considering your backgrounds, I’m betting you have great thoughts and insights. Others?
https://adamunikowsky.substack.com/p/in-ai-we-trust-part-ii
