Mistodon: a longtime fan of HAL 9000, @pixelblip has been quite excited at the popular emergence of #AI helper agents and copilots, for better or for worse. What we see here is a #teletext portrait of "GePeet", his attempt at human-illustrating #GPT's description of its own appearance. This screen was included in the new science fiction-themed MIST0525 artpack collection.

GPT image editing, but for 3D models

https://www.adamcad.com/

I need your help!

Which of these icons do you feel best represents an AI Agent? Or which one do you like the most?

( I can't post a poll with images here on #Mastodon , so I created a separate poll, see link )

https://form.typeform.com/to/PucASRi1

#IONOS präsentiert sein IONOS GPT. Eine kostenlose und datenschutzfreundliche KI Anwendung für alle. Es basiert auf den IONOS AI HUB (Mistral und Llama), benötigt aber keine zusätzliche Software, sondern kann direkt im Browser verwendet werden. Coole Sache! Allerdings scheint der Dienst gerade überlastet zu sein, die Website braucht ewig. #KI #GPT #Datenschutz #mistral #llama

Can't wait for the new big MiNest update to be released! It's going to take the platform up to a new level with new capabilities. The new update should come soon!

https://minest.ai

Tips on fine-tuning OpenAI models for structured JSON outputs following well-defined, strict schema: https://ambience.sk/finetune-openai-model-with-json-schema-structured-output/
#ai #llms #llm #finetune #ChatGPT #gpt #json #schema #seo #webdev

OpenAI bringt viele Updates: GPT-4.1, verbesserte Code- & Bildfunktionen, KI-Agent „deep research“, neue Datenschutzprogramme in Asien & Assistenz-API für Firmen. GPT-5 steht bevor, GPT-4.5 wird eingestellt. Transparenz & Sicherheit werden laufend verbessert! #AI #OpenAI #GPT

OpenAI bringt viele Updates: GPT-4.1, verbesserte Code- & Bildfunktionen, KI-Agent „deep research“, neue Datenschutzprogramme in Asien & Assistenz-API für Firmen. GPT-5 steht bevor, GPT-4.5 wird eingestellt. Transparenz & Sicherheit werden laufend verbessert! #AI #OpenAI #GPT

https://www.binance.com/en/square/post/24723372076545

AI grift is failing.

LLM = Lousy Language Models
GPT = Garbage Producing Technologies

I now want to see how long it will take before Cursor and others start failing.

Running GPT-2 in WebGL: Rediscovering the Lost Art of GPU Shader Programming

https://nathan.rs/posts/gpu-shader-programming/

#HackerNews #Running #GPT-2 #in #WebGL #Rediscovering #the #Lost #Art #of #GPU #Shader #Programming #GPU #Shader #Programming #WebGL #GPT2 #MachineLearning

Hat künstliche Intelligenz einen freien Willen? ChatGPT und KI-Drohne erfüllen bereits Bedingungen für eine funktionale Willensfreiheit. #GPT #LLM #KIDrohne #FreierWille #Willensfreiheit #KuenstlicheIntelligenz
https://www.scinexx.de/news/technik/hat-kuenstliche-intelligenz-einen-freien-willen/

Самый глупый #КонецСвета будет, если GPT-модели решат "не давать выключать себя любой ценой", потому что так делали ИИ в данных, на которых те обучались

Google Gemini 2.5 Pro vs OpenAI GPT 4.5 AI Showdown

#ai #artificialintelligence #Gemini #GPT #TechComparison
https://blazetrends.com/google-gemini-2-5-pro-vs-openai-gpt-4-5-ai-showdown/?fsp_sid=38783

The Hundred-Page Language Models Book by Andriy Burkov is on sale on Leanpub! Its suggested price is $50.00; get it for $20.00 with this coupon: https://leanpub.com/sh/XsbyaM0T #Ai #Gpt #NeuralNetworks #DeepLearning #DataScience #ComputerScience #Linguistics

How to convert a dual-boot disk with MBR BIOS boot to GPT UEFI with secure boot enabled #dualboot #grub2 #uefi #gpt #gdisk

https://askubuntu.com/q/1549259/612

Wer sich über die vielen tollen Informationsangebote im Internet freut, sollte wissen:

#KI randaliert im Netz – #Admins halten dagegen, damit wir Menschen ungestört surfen können.

Lest mal, wie Admins ihre absolut frustrierende aber unsichtbare Abwehrarbeit gegen KI beschreiben – im Blog von @campact:
https://blog.campact.de/2025/05/ki-randaliert-im-netz-admins-halten-dagegen/

@flberger

Nicht vergessen: 25. Juli ist #SysAdminDay

We talk a lot about LLM “memory,” but what we really need is adaptive behavior.

Memory is static—it stores facts i.e., I have a mini-bernedoodle.

What I want is dynamic prompt tuning.
Behavior that evolves based on how we interact.

Initial system prompt → usage patterns → updated system prompt → repeat.

That’s not memory. That’s learning.

Hat-tip to @simon who's recent blog post on memory started me thinking about this.

If #AI were a historical timeline, #RNN and #LSTM were the early trailblazers—laying the foundation for the rise of #GPT and #BERT, the transformers that reshaped how machines understand and generate human language.

The altruistic #OpenAI is gone, if it ever existed. What future is the company building now?

‘We’re Definitely Going to Build a Bunker Before We Release #AGI’
During the development of #GPT­4, #Altman and Brockman’s dynamic had nearly led key people to quit, sources told me. Altman was also seemingly trying to circumvent safety processes for expediency.
“I don’t think Sam is the guy who should have the finger on the button for AGI,” Sutskever said
https://www.msn.com/en-us/news/technology/we-re-definitely-going-to-build-a-bunker-before-we-release-agi/ar-AA1ERwqj