| Deutsch English Français Italiano |
|
<1033r20$2cg5$2@dont-email.me> View for Bookmarking (what is this?) Look up another Usenet article |
Path: news.eternal-september.org!eternal-september.org!.POSTED!not-for-mail From: Rhino <no_offline_contact@example.com> Newsgroups: rec.arts.tv Subject: Re: Your AI Lawyer Date: Fri, 20 Jun 2025 10:26:08 -0400 Organization: A noiseless patient Spider Lines: 26 Message-ID: <1033r20$2cg5$2@dont-email.me> References: <1032r9v$f3dr$1@dont-email.me> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 8bit Injection-Date: Fri, 20 Jun 2025 16:26:09 +0200 (CEST) Injection-Info: dont-email.me; posting-host="3c18b11359a6d1f690ec270a55b36f20"; logging-data="78341"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1+FyVIeYI5EvIviCX7oPxc2cVnxbZxkGWQ=" User-Agent: Mozilla Thunderbird Cancel-Lock: sha1:ewU33mooCfyfBJt+r6WQ3ndteG8= X-Antivirus-Status: Clean X-Antivirus: Avast (VPS 250620-0, 6/19/2025), Outbound message In-Reply-To: <1032r9v$f3dr$1@dont-email.me> Content-Language: en-CA On 2025-06-20 1:24 a.m., BTR1701 wrote: > I just read that Elon's Grok AI scored in the 92nd percentile on the > California Bar Exam. And it did it in about 10 seconds. <sigh>… > > I'm not surprised an AI could do well in a test like that which is (presumably) a regurgitation of case law and precedents that it has "learned". Then again, when I try to write Javascript code with AI, both chatGPT and claude (emerging as the hottest new AI for developers) do a shit job of it. They confidently deliver code that they assure me will do what I asked - and I deliberately made the request simple - and it inevitably fails. When I point out that the code doesn't work, the AI readily admits that fact and quickly offers me a revised solution that doesn't work either. On and on it goes with each solution utterly failing to work. After the third or fourth attempt to clean up the original solution, you realize that it is going in circles, retrying the same things that have already failed. I don't think I'd want to put my life and liberty in the hands of an AI. Real courts don't give you a lot of do-overs. Even if they did, you can only follow one failing strategy after another so many times before everyone agrees that you've lost the case. -- Rhino