I have briefly tested my #AIText #question type with #Moodle 5 and it appears to work as expected. It gives feedback and (optionally) marks to a student response based on prompt to the #LLM. In this case it was OpenAI with gpt-4.1 With thanks to @PhilippMemmel who gave me an essential hint about the code.

My next stop will be to test against
https://github.com/mebis-lp/moodle-local_ai_manager

Which is the Mebis-lp (Bavaria) AI back end which is a good alternative to the core Moodle AI subsystem.

When is 4.1 greater than 4.5? When it’s OpenAI’s newest model. - On Monday, OpenAI announced the GPT-4.1 model family, its newest series of... - https://arstechnica.com/ai/2025/04/when-is-4-1-greater-than-4-5-when-its-openais-newest-model/ #largelanguagemodels #machinelearning #aiassistants #samaltman #chatgpt #chatgtp #gpt-4.1 #gpt-4.5 #biz⁢ #gpt-4 #api #ai

I found a simple hilarious question that fucks up AI:

Alice has two sisters. She also has has two brothers. How many sisters does Alice's brother have?

I tried this on #chatGPT and I got a crazy broken computer answer. Like nothing I had seen before. We shouldn't rely on chatGPT for anything important.

https://www.wacoca.com/pets/381866/ 擬人化したけど･･ #つーちゃん #コーギー #chatgpt #dog #Inu #ウエルシュ・コーギー・ペンブローク

AI access only with verification: OpenAI restricts use for corporate customers

Companies will have to verify themselves with OpenAI to access AI models and additional functions. This is intended to protect against misuse.

https://www.heise.de/en/news/AI-access-only-with-verification-OpenAI-restricts-use-for-corporate-customers-10351723.html?wt_mc=sm.red.ho.mastodon.mastodon.md_beitraege.md_beitraege&utm_source=mastodon

There's a trend going around where people are asking #ChatGPT to transform their dog into a human.

Here's the result of my young black and tan chihuahua. She became a beautiful young girl.

#SteadyCommunityContent

Nach dem Einbau des #Raspberry Pi 4 in das mysteriöse #Gehäuse kommt jetzt die #Inbetriebnahme.

Mit #Linux arbeite ich zwar selten, doch grundsätzlich komme ich damit klar. Wo früher noch mühsam gegoogelt oder #Manpages konsultiert wurden, frage ich heute einfach #ChatGPT. Das geht für eine Root-Cause-Analyse bestimmt 100-mal schneller. Also, womit fange ich an?

https://tino-eberl.de/nutzen-kuenstlicher-intelligenz/raspberry-pi-4-einrichten-mein-weg-zum-stromsparenden-mini-pc-fuer-python/

Draußen scheint die Sonne, die Balkonmöbel rufen meinen Namen, aber ich sitze hier vor einem Ablehnungsschreiben einer Beanstandung und drehe gleich durch.

Ich versuche das mal mit #ChatGPT soll ja gut funktionieren, sagen die Leute