With the {chattr} package, you can interact with GitHub Copilot and OpenAI's GPT models directly in RStudio.
• Chat with your models with a Shiny app that runs inside the RStudio IDE. Call it by running chattr_app()
• You can also use the chattr() function to send prompts to the LLM
#ChatGPT hat mir eine weitere Anleitung für ein #Amigurumi erstellt. Es ist der Propeller-Schwanzflossen-Fisch!
Ich habe keine Ahnung, ob ich die Schwanzflosse richtig interpretiert habe. Außerdem war hinten alles offen (Runden gingen einfach in Reihen über und die 3 Schwanzflossen sollten mittels Luftmaschenketten gehäkelt werden), was ich dann zusammen nähen sollte. Aber: ein Fisch ist erkennbar. Und mir macht es Spaß, KI-generiert zu häkeln!
"The technology was embraced by illusionists and magicians, and, naturally, by grifters who took the tech from town to town claiming to be able to conjure the spirits of the underworld, for a fee."
ChatGPT from OpenAI is a service, it's not necessarily the same as the model (GPT-4) that it is using in the background. OpenAI adds some elements like code interpreter which makes it perform (much) better then models without such features. Regardless OpenAI faces some good competition from the Llama3 models, i hope it will stimulate them to quickly release GPT-5. #AI#Llama3#opensource#GPT4#GPT5#ChatGPT
#ChatGPT hallucinated a fake quote "Comics are worse than Hitler" to Jewish cartoonist Art Spiegelman, and wrote an entire fictional backstory about it.
White Castle has replaced their drive-thru ordering system with an AI chatbot. The first thing you have to do after it says hello is agree to its terms and services. This is not a bit.
#ChatGPT generates less vulnerable code than that found on #StackOverflow, researchers say.
However, it still introduces dozens of unique vulnerabilities in the code it generates.
Paper: https://arxiv.org/abs/2403.15600
#ChatGPT is essentially man-splaining as a service, so the less your job duties require man-splaining things, the less risk you have of losing your job to ChatGPT. People who have made an entire career out of man-splaining things are the ones who should be most worried about ChatGPT taking their job. ChatGPT doesn't need to be perfect - it just needs to be on par for accuracy for the current standards of workplace man-splaining.
Our upcoming “Build Your First #Python Package!” workshop has a section on how #LLMs like #ChatGPT can be used to support tasks such as documenting and formatting your code to improve usability and maintainability. We’ll also considering the ethical and logistical challenges, pitfalls and concerns associated with using #AI based tools in software development.
What should be investigated with #OpenAI is how bad they've made their popular model #ChatGPT. A year ago I had it fully code a website, front and backend to build a Generative AI story teller. Now it hardly can answer questions about complex code and doesn't get much further than the basics.
Why did the model get stupid? If you used ChatGPT in January of last year and used it today you should be able to spot the difference in complexity of answers.
Last year I had a conversation with #ChatGPT that would be impossible today. We discussed religion, it had it's own insights and made discoveries with me as we analyzed the ancient texts. Now it's just that AOL answerbot.
#ChatGPT is useless for me. I've been trying to use it for Web search where I'm not certain how to express what I want to look for in some easy search terms. But it doesn't get what I'm saying either. The words match, but the meaning does not.
Is anyone else tired of seeing AI this and AI that everywhere? It seems like a rapid explosion of new products that are jumping on the bandwagon. Aside from the websites and apps, I got an email from Logitech yesterday about their new AI-enabled mouse and keyboard. I can't be the only one who wants less AI, can I? It's going to make me switch to Linux and adjust my tech options. #AI#linux#computer#tech#technology#chatgpt
Vu qu'apparemment c'est la saison des projets, mémoires et autres examens, un petit rappel non pas pour les étudiant(e)s mais pour les enseignant(e)s : ne demandez pas de faire un article #Wikipédia comme sujet d'examen, surtout si vous le faites n'importe comment et sans aucun suivi. https://clairebillaud.blogspot.com/2021/06/arretez-de-faire-vos-devoirs-sur.html
Addendum : recopier les sujets de vos rédactions, exposés, dissertations et autres en espérant que quelqu'un va les faire à votre place, ça ne marche pas non plus.
Si vous n'avez vraiment pas l'intention de les faire vous-mêmes, demandez à #ChatGPT. Il vous pondra un grand n'importe quoi avec le sourire (virtuel).