OpenAI acaba de lanzar para todo el mundo su motor de búsqueda potenciado por IA integrado en ChatGPT, tanto en web como en aplicaciones: https://www.dekazeta.net/openai-lanza-su-motor-de-busqueda-integrado-en-chatgpt-para-todo-el-mundo/

Je n'avais pas trop entendu de ça #AI #OpenAI #chatpgt https://www.zdnet.fr/actualites/viree-dopenai-elle-balance-altman-a-ignore-le-conseil-dadministration-pendant-des-annees-392265.htm?utm_campaign=NL_Zdnet_30052024&utm_content=30052024&utm_medium=email&utm_source=EMAIL&rwid=B9689340B924A9213960F4B5ADB3BAEB042B7B7A9B04D8E7D0201065F76576C7

C’mon, we’ve all been using LLMs on a daily basis for more than year now. You’d think people would slowly start to get the idea. But no.

Why is everyone still using an LLM and looking for accurate answers? They are obviously not built for that.

You use an LLM to get yourself off your writer’s block, to get a few ideas how to move on, not for it to solve the end equation for you. Duh!

—
#AI #ChatPGT #LLM

Uncomfortable thought of the day (via [1]):

Have there already been sightings of LLMs *asking* questions on forums like StackOverflow?
Seems like something that some hot AI startup should be doing right now.

"ChatGPT not giving you good answers? Let us ask your question on fifty programming forums in parallel and generate a response based on expert consensus! To reduce latency, we'll identify the hottest upcoming questions and pre-post them!!"

[1] https://forum.dokuwiki.org/d/21215-please-do-not-post-ai-generated-answers/4

Wie man einem großen Sprachmodell Mathe beibringt

Eine neue Methode zeigt, wie große Sprachmodelle wie GPT-3 zuverlässiger Matheaufgaben lösen können.

#fedilz #fedicampus #openai #chatpgt #mathe

https://bit.ly/3vRJJSY