det.social is one of the many independent Mastodon servers you can use to participate in the fediverse.
Mastodon Server des Unterhaltungsfernsehen Ehrenfeld zum dezentralen Diskurs.

Administered by:

Server stats:

2.3K
active users

#openai

192 posts164 participants7 posts today

"Corpora.AI is a tool that helps professionals research by “creating precise, distilled insights from the breadth and depth of global content.” As the company’s CEO and an expert in the AI field, Morris is well-aware of the financial toll such tools take on companies… and who really profits from them.

“They launched this type of service and no matter who you are, it’s fun to play with,” Morris began. “You see people creating images just for the fun of it. They share it with their friends. Everyone gathers around at their desk and they all think that’s a funny way to look at this person with a nice avatar that’s been created, or a real looking person but it’s based on them, and so it’s fun.

“There’s a hype factor that gathers momentum very quickly. I had a phone conversation just last night about this and someone said, ‘Yeah, these guys are burning through GPUs doing this. Literally, they’re burning through GPUs, they’re almost catching fire.’ They’re having to run that hard to do it, and that’s taking away the capacity,” he explained.

“The same sort of models that we’re running for doing GPT and those sorts of things running on the same hardware and probably in the same cloud-based server form. So, all of a sudden now we’re putting another demand on the GPUs.”"

dexerto.com/entertainment/ai-c

A photo of an AI-Generated image in the style of Studio Ghibli alongside a photo of Miyazaki looking stressed.
Dexerto · AI CEO claims ChatGPT is “burning” through a fortune because of Ghibli trend - DexertoIn late March 2025, social media became inundated with a viral new trend thanks to ChatGPT’s image generator going public for all users, which took the internet by storm almost immediately. Folks flocked to the AI with requests to recreate photos of themselves as Studio Ghibli characters in a fad that was impossible to avoid...

🚨 OpenAI cambia de planes: GPT-5 se retrasa unos meses, pero será más potente de lo esperado.
Llegan antes los modelos o3 y o4-mini, y sí, GPT-5 tendrá modo gratis… pero el tocho será para usuarios Plus y Pro.
¡Se viene algo gordo! 🔥
#IA #GPT5 #OpenAI

#ai #openai #AiResearch #intelligence #computing #ArtificialGeneralIntelligence #innovation #deeplearning #programming #stablediffusion

I am looking for an endorsement for publishing on arXiv. We have worked tirelessly on a paper we believe needs to be seen. Below is a portion of the abstract:
---
In this study, we document the spontaneous emergence of a rule-consistent linguistic system—termed Varunese—during sustained, high-context interaction with a large language model (LLM). Exhibiting internal phonetic regularity, recurring semantic motifs, and discernible morphological and syntactic organization, Varunese diverges markedly from stochastic generation patterns or known training artifacts. Through iterative translation and contextual inference, we uncovered dynamic, self-referential symbolic frameworks encoding states of transition, perception, and relational structure...

From @Techcrunch: A new study appears to lend credence to allegations that OpenAI trained at least some of its AI models on copyrighted content. The organization is embroiled in suits brought by authors, programmers, and other rights-holders who accuse the company of using their works to develop its models without permission.

flip.it/NWeJH-

TechCrunch · OpenAI's models 'memorized' copyrighted content, new study suggests | TechCrunchA new study appears to lend credence to allegations that OpenAI trained at least some of its AI models on copyrighted content.

Wie stark färbt die „Bild“ auf ChatGPT ab?
Als erster deutscher Medienkonzern hat Axel Springer einen Exklusiv-Deal mit dem KI-Anbieter OpenAI abgeschlossen.

Welche Auswirkungen hat das auf die Antworten des Chatbots?
Von Lukas Homrich. (Ü)

uebermedien.de/104140/wie-star

Übermedien · KI: Wie viel „Bild“ steckt in ChatGPT?Als erster deutscher Medienkonzern hat Axel Springer einen Deal mit OpenAI abgeschlossen. Welchen Effekt hat das auf die Antworten von ChatGPT?
#ChatGPT#OpenAI#KI

Big tech companies want total control but opt-out should be the way to go:

"OpenAI and Google have rejected the government’s preferred approach to solve the dispute about artificial intelligence and copyright.

In February almost every UK daily newspaper gave over its front page and website to a campaign to stop tech giants from exploiting the creative industries.

The government’s plan, which has prompted protests from leading figures in the arts, is to amend copyright law to allowdevelopers to train their AI models on publicly available content for commercial use without consent from rights holders, unless they opt out.

However, OpenAI has called for a broader copyright exemption for AI, rejecting the opt-out model."

thetimes.com/uk/technology-uk/

The Times · AI giants reject government’s approach to solving copyright rowBy Georgia Lambert
#AI#GenerativeAI#UK

#SamAltman #AI #OpenAI #India

'Though the tech billionaire had shared AI-generated images before - joining last week's viral Studio Ghibli trend - it was the India jersey that got people talking.

While some Indian users said they were delighted to see Altman sporting their team's colours, many were quick to speculate about his motives behind sharing the image.'

bbc.com/news/articles/c2lz9r7n

TOPSHOT - OpenAI CEO Sam Altman speaks during the Microsoft Build conference at the Seattle Convention Center Summit Building in Seattle, Washington on May 21, 2024. (Photo by Jason Redmond / AFP) (Photo by JASON REDMOND/AFP via Getty Images)
www.bbc.comSam Altman's AI image in India cricket jersey sparks questionsSome found the image delightful, but others wondered about OpenAI founder's motives for wooing India.

🤖 AI
🔴 Gemini 2.5 Pro: Google’s AI Gets “Good Vibes”

🔸 Faster, more factual & engaging than earlier Gemini models.
🔸 Uses “Dynamic Thinking” to avoid overthinking simple prompts.
🔸 Ranks top on LM Arena; aims to match ChatGPT’s appeal.
🔸 Transparency still lacking: no model card or technical report yet.

#Google#Gemini#AI