Skip to main content

Search

Items tagged with: genai

Arguments for using #GenAI in critical applications like science, medicine, and law often fundamentally misunderstand how the technology works. I've written up my best attempt at explaining how.

citizensandtech.org/2026/05/re…


Reliable GenAI and Perpetual Motion - Citizens and Technology Lab

Researchers rushing to adopt AI without understanding the limitations risk committing scientific misconduct.
Edward L Platt (Citizens and Technology Lab)
#genai

RE: social.coop/@natematias/116606…

"Expecting correctness from #GenAI is like expecting #perpetualmotion from a machine. Gen AI has no concept of correctness. And even if it did, GenAI favors plausible falsehoods over unlikely truths. And unlikely truths are the heart of scientific discovery. "

J. Nathan Matias 🦣 (@natematias@social.coop)

Until @elplatt@greatjustice.net wrote this blog post, I had no idea that biostatisticians have used the analogy of perpetual motion to think about what AI can and cannot do.
J. Nathan Matias 🦣 (social.coop)
#genai #perpetualmotion

Kovacs has a well written analysis of the infosec industry take on @bagder's experience with Mythos.

securityweek.com/claude-mythos…

#genai #sast


Claude Mythos Finds Only One Curl Vulnerability; Experts Divided on What It Really Means - SecurityWeek

A test of Anthropic’s restricted Claude Mythos model found just one low-severity vulnerability in the open source data transfer tool curl.
Eduard Kovacs (SecurityWeek)
#genai #sast @daniel:// stenberg://

I was only made aware of this (frankly awesome) case of LLM poisoning today: nature.com/articles/d41586-026…. A researcher made up a disease and published two evidently fake preprints about it (including sentences such as “this entire paper is made up” and “Fifty made-up individuals aged between 20 and 50 years were recruited for the exposure group”), which were almost immediately picked up by LLMs and documented in their output. Worse, actual – supposedly serious – medical papers also started citing the preprints, demonstrating that academics relying on LLMs to do their work is a genuine problem! Not that I had my doubts but, if anyone did, this seems like the perfect demonstration of the problem. Article immediately added to the syllabus of the class I am co-teaching with Iris Ferrazzo on LLMs for Romance Studies/Humanities!

#LLM #GenAI #academia #research #ResearchIntegrity #humanities


Scientists invented a fake disease. AI told people it was real

Bixonimania doesn’t exist except in a clutch of obviously bogus academic papers. So why did AI chatbots warn people about this fictional illness?
Stokel-Walker, Chris
#research #academia #humanities #llm #genai #ResearchIntegrity

The US AI economy #llm #llms #genai
a power strip plugging back into itself
#llm #llms #genai

Google Just Bought a Stake in the Maker of Eve Online to Train Its AI Models

engadget.com/2166270/google-ju…

#videogames #pcgaming #gaming #gamedev #google #genai #ai


Google Just Bought A Stake In The Maker Of Eve Online To Train Its AI Models

DeepMind will research player behavior in the space-based RPG.
Lawrence Bonk (Engadget)
#ai #videogames #gaming #Google #gamedev #pcgaming #genai

"There is no ethical way to use the major AI image generators. All of them are trained on stolen images, and all of them are buit for the purpose of deskilling, disempowering and replacing real, human artists" 🎯 Molly Crabapple - 2023

#AI #generativeAI #genAI #artist #art #images #technology #MollyCrabapple

"There is no ethical way to use the major AI image generators. All of them are trained on stolen images, and all of them are buit for the purpose of deskilling, disempowering and replacing real, human artists" Molly Crabapple Artist and activist https://www.instagram.com/p/DXvnX92Dmer/?img_index=10
#art #ai #technology #artist #generativeAI #genai #images #mollycrabapple

Por más que tengan buenas intenciones, lo que para ustedes podría ser un uso «ético y responsable» es avalar y legitimar la vulneración de derechos sistemática que sostiene toda la industria de la IA generativa comercial.

📌 Ningún modelo de IAG comercial funciona sin VIOLAR derechos de autor.

#IA #IAgenerativa #AI #genAI #generativeAI #datasets #theft #technology #ethics

• Pero mucha gente dice estar usando IA generativa de forma ética y responsable Si estás a favor de que esta tecnología sea ética y transparente, y de que las compañías obtengan el consentimiento explícito previo, además de acreditar y remunerar a los autores: no uses ningún modelo comercial de IA generativa porque ninguno cumple estos requisitos. La única manera de resolver una parte de esta problemática es eliminar los datasets existentes. Tendrían que empezar DESDE CERO con un marco regulatorio que permita hacer un desarrollo con trazabilidad y supervisión de forma controlada y con transparencia. https://www.instagram.com/p/DXvnX92Dmer/?img_index=9
#ai #technology #ethics #generativeAI #genai #ia #theft #iagenerativa #datasets

La leyenda urbana de «entrenar tu propio modelo de IAG» sigue siendo un anzuelo para monetizar tutoriales, cursos, masterclass y demás productos que los gurúes y promotores de la IA generativa usan para seguir lucrando a costa de todos los autores vulnerados. ¡No se dejen engañar!

#AI #MachineLearning #data #training #finetuning #AImodel #genAI #generativeAI #pretraining #Copyright #opensource

• Bueno, pero hay tutoriales para «entrenar tu propio modelo de IAG» Cuando alguien dice que ha «entrenado su propio modelo de IA», normalmente está utilizando una terminología incorrecta y en realidad ha ajustado un modelo existente. Ajustar o refinar un modelo existente, o una parte del mismo, no elimina el problema de base. https://www.instagram.com/p/DXvnX92Dmer/?img_index=8
#opensource #ai #Data #copyright #machinelearning #generativeAI #genai #training #finetuning #aimodel #pretraining

¡Ah! pero China... China repite las mismas dinámicas que Estados Unidos en cuanto al desarrollo de sus propios modelos comerciales de IA generativa. Y no, acá no aplica el refrán del ladrón con los 100 años de perdón. Hablamos de gigantes tecnológicos lucrando a costa de nuestros datos y obras.

#China #AI #DeepSeek #technology #USA #genAI #OpenAI #ChatGPT #generativeAI #BigTech

• ¿Y qué pasa con los modelos chinos de IA generativa como DeepSeek? Pasa lo mismo que con los modelos de EE.UU. https://www.instagram.com/p/DXvnX92Dmer/?img_index=7
#ai #technology #USA #China #BigTech #openai #chatgpt #generativeAI #genai #deepseek

Muchos intentan asociar el rótulo de open source como «libre de culpa y cargo». Y siguen mezclando peras con manzanas. El hecho de que un modelo sea open source no garantiza que sea ético, transparente, confiable ni con seguridad jurídica. Stable Diffusion se lanzó como open source y es la génesis del mal en todo este asunto.

#AI #StableDiffusion #opensource #genAI #stabilityAI #LAION #generativeAI #ethical #ethics #technology #dataset

• ¿Y qué pasa con los modelos de código abierto que son gratuitos? El hecho de que un modelo sea open source (código abierto) no garantiza que sea ético, transparente, confiable ni con seguridad jurídica. Stable Diffusion se lanzó como open source y es la manzana podrida que nos trajo hasta donde estamos hoy. Sta­bil­ity AI Sta­bil­ity AI, founded by Emad Mostaque, is based in Lon­don. Sta­bil­ity AI funded LAION, a Ger­man orga­ni­za­tion that is cre­at­ing ever-larger image datasets—with­out con­sent, credit, or com­pen­sa­tion to the orig­i­nal artists—for use by AI com­pa­nies. Sta­bil­ity AI is the devel­oper of Sta­ble Dif­fu­sion. Sta­bil­ity AI trained Sta­ble Dif­fu­sion using the LAION dataset. Sta­bil­ity AI also released Dream­Stu­dio, a paid app that pack­ages Sta­ble Dif­fu­sion in a web inter­face. https://imagegeneratorlitigation.com/ https://www.instagram.com/p/DXvnX92Dmer/?img_index=6
#opensource #ai #technology #ethics #stablediffusion #generativeAI #genai #ethical #stabilityai #laion #dataset

Al día de hoy, solo en Estados Unidos, se acumulan más de 100 procesos judiciales en curso con sobrada evidencia de esta realidad que aplica a todos los servicios comerciales de IA generativa, sean de imagen, video, audio, texto, etc.

A nivel global, ascienden a más de 130 demandas.

#AI #genAI #generativeAI #technology #lawsuits #litigation #USA #copyright #art #artist

¿Hay alguna demanda contra estas empresas de IA generativa? https://www.wired.com/story/ai-copyright-case-tracker/ https://imagegeneratorlitigation.com/
#art #ai #technology #USA #copyright #artist #generativeAI #genai #lawsuits #litigation

A muchos les encanta desviar el foco de atención tratando de comparar el proceso de inspiración humana con el accionar delictivo de un puñado de empresarios inescrupulosos.

Recuerden siempre volver al punto clave: violación masiva a la Ley de Propiedad Intelectual y a la Ley de Protección de Datos.

#AI #genAI #WIPO #OMPI #dataprotection #dataprivacy #data #generativeAI #IPinfringement #copyright #intellectualproperty #propiedadintelectual #authorsrights #humanrights

• Pero los artistas también se inspiran y aprenden de otros, ¿no? No se trata de la inspiración humana. Se trata de la VIOLACIÓN de la Propiedad Intelectual para desarrollar modelos de IA generativa. Se trata de un modelo de negocio de un grupo de empresas que han construido sus sistemas de IA generativa a partir de prácticas ilegales. IP infringement Many generative AI tools are trained on enormous quantities (sometimes billions) of items protected by IP. There are several ongoing legal disputes alleging that the scraping and use of these works to train AI, the trained AI models, and their outputs are IP infringements. Training data issues Some generative AI tools have been trained using materials scraped from the internet, including copyright works, personal information, biometric data, and harmful and illegal content. There is ongoing litigation over whether the scraping, downloading, and processing of materials, the trained AI models, and their outputs involve breaches of IP, privacy, and contract. Debates are ongoing about the balance of interests between IP owners and AI developers. https://www.wipo.int/publications/en/details.jsp?id=4713 https://www.instagram.com/p/DXvnX92Dmer/?img_index=4
#ai #humanrights #Data #copyright #dataprivacy #dataprotection #generativeAI #genai #intellectualproperty #wipo #ompi #ipinfringement #propiedadintelectual #authorsrights

Para empezar a explicar el problema hay que delimitar dónde está ese problema. No hablamos de toda la tecnología ni de cómo sintetizar proteínas con sistemas de IA (a secas). Hablamos de la IA generativa actual cuyos modelos comerciales fueron lanzados al mercado desde 2021-2022 en adelante.

#AI #genAI #generativeAI #ChatGPT #Midjourney #Gemini #AdobeFirefly #GROK #META #NanoBanana #RunwayAI #StableDiffusion #Flux #Llama #Claude #suno #ElevenLabs #Microsoft

• ¿Cuáles son estos modelos comerciales de IA generativa? No hablamos de sistemas experimentales de IA para investigación, sino de la IA generativa actual, cuyos modelos comerciales fueron lanzados al mercado entre 2021-2022 y se entrenaron con material protegido en base a la Ley de Propiedad Intelectual, sin el consentimiento de los autores. DALL-E Stable Diffusion Midjourney FLUX ChatGPT GPT Imagen, de Google GEMINI Nano Banana CLAUDE Bing, de Microsoft Copilot LLaMA META AI GROK ElevenLabs Adobe Firefly Runway SORA Udio SUNO Veo3 LeonardoAI y muchos más https://www.instagram.com/p/DXvnX92Dmer/?img_index=3
#ai #microsoft #meta #stablediffusion #midjourney #chatgpt #generativeAI #gemini #genai #grok #claude #llama #adobefirefly #nanobanana #runwayai #flux #suno #elevenlabs

Los actuales modelos comerciales de IA generativa han sido desarrollados vulnerando el Reglamento General de Protección de Datos (RGPD) y la Ley de Propiedad Intelectual (LPI). Jamás hubo un pedido de consentimiento por parte de las empresas tecnológicas. Por eso hablamos de ROBO DE DATOS.

#AI #genAI #generativeAI #data #datos #robo #robodedatos #theft #stolen #illegal #technology #bigtech #author #scraping

• ¿Cómo se entrenaron los modelos comerciales de IA generativa? Para entrenar un modelo de IA generativa se necesitaron miles de millones de DATOS (textos, imágenes, audios, videos, etc). Muchos de esos datos son contenidos con derechos de autor. Todo se extrajo de Internet mediante scraping, sin el consentimiento de los creadores, sin acreditarlos y sin compensarlos. https://www.instagram.com/p/DXvnX92Dmer/?img_index=2
#ai #technology #Data #BigTech #author #generativeAI #genai #illegal #theft #stolen #scraping #datos #robo #robodedatos

Mucho chanta ofreciendo «cursos para descubrir el mundo de la IA generativa» y monetizar a costa de la vulneración sistemática de nuestros derechos. No se necesita ningún curso para saber cómo funcionan los modelos de IAG comercial 👇

🔸 Cuadernillos de Arte es Ética. Volumen 1: ROBO DE DATOS

🔗 arteesetica.org/cuadernillos-d…

#IA #IAgenerativa #AI #genAI #generativeAI #guide #copyright #humanrights #art #artist #author #lawsuits #technology #opensource #deepseek

Foto de portada: libreta de notas de color rojo brillante con encuadernación cosida (Unsplash) con la etiqueta: Cuadernillos de Arte es Ética. Volumen 1: ROBO DE DATOS Debajo en titulo grande sobre el fondo rojo: El desarrollo de la actual IA generativa comercial está basado en la ilegalidad. https://www.instagram.com/p/DXvnX92Dmer/?img_index=1

Cuadernillos de Arte es Ética. Volumen 1: ROBO DE DATOS – Arte es Ética

arteesetica.org
#art #opensource #ai #technology #humanrights #copyright #author #artist #generativeAI #genai #ia #guide #lawsuits #deepseek #iagenerativa

Xiaomi has unveiled its MiMo-V2.5-Pro and MiMo-V2.5 models, which the company says match the performance of leading frontier models on key benchmarks while consuming significantly fewer tokens. The release underscores intensifying competition among Chinese AI developers to deliver capable models at lower computational cost. marktechpost.com/2026/04/22/xi… #AI #GenAI #AIResearch

Xiaomi Releases MiMo-V2.5-Pro and MiMo-V2.5: Matching Frontier Model Benchmarks at Significantly Lower Token Cost - MarkTechPost

Xiaomi Releases MiMo-V2.5-Pro and MiMo-V2.5: Matching Frontier Model Benchmarks at Significantly Lower Token Cost
Asif Razzaq (MarkTechPost AI Media Inc)
#ai #genai #AiResearch

V2.5-Pro is the next step up from MiMo-V2-Flash and its closed source for now, but Xiaomi has confirmed open source is coming for the V2.5 series. What V2.5-Pro adds over Flash is meaningful. Better long-horizon coherence, stronger agentic capabilities, and the ability to sustain complex tasks across more than a thousand tool calls without losing the thread

firethering.com/mimo-v2-5-pro-…

#technews #ai #llm #genai #mimo #xiaomi #opensource #trending #agent


MiMo-V2.5-Pro: A Coding Model Taking On Claude Opus 4.6 and GPT-5.4 - Firethering

Peking University gives its computer science students a compiler project every semester. Build a complete SysY compiler in Rust including lexer, parser, abstract syntax tree, IR code generation, assembly backend, performance optimization.
Mohit Geryani (Firethering)
#opensource #ai #technews #llm #genai #trending #agent #mimo #xiaomi

RE: masto.ai/@phoronix/11647612006…

I literally just moved most of my computers to #Linux (#ZorinOS, a variant of #Ubuntu) to avoid the #GenAI imposition of #MicroSlop and now Ubuntu wants to add genAI in their OS?
NOOOOOO 😱😖🤢

Now is the time to push back on this! We don't want genAI!!! The rare people who think they want it can do that on their own and separately from the OS!

Yes, I could move to another distribution but #Ubuntu is the only one supported by my workplace 😭


Phoronix (@phoronix@masto.ai)

Attached: 1 image Ubuntu Linux Will Begin Landing AI Features Throughout The Next Year Now that Ubuntu 26.04 LTS has shipped, Canonical is opening up on their next major focus for Ubuntu development: lots of AI features... https://www.phoronix.
Phoronix (Mastodon)
#Linux #ubuntu #genai #zorinos #microslop

"Imagined by Moleskine, generated by AI" Dunno, seems bad to me.

As Eddie Robson (@eddierobson.bsky.social) puts it: "This does seem like very poor judgment for a company whose entire customer base is people who write and draw by hand."

#Moleskine #GenAI #Slop

cjleo.com/blog/moleskine-ai-lo…


Moleskine's AI Lord of the Rings Collection Can Only Mock

New Moleskine X The Lord of the Rings notebook collection is "imagined by Moleskine, generated by AI".
CJL (Game Design Blog — Cheryl-Jean Leo)
#genai #slop #moleskine

This #genai #llm #tokenCost is getting silly. Can't wait for the bubble to burst.
Christoffer Bjelke @chribjel.Apr 21 We hired a junior developer to write the simple code, so we don't have to spend a ton of money on tokens for those basic/primitive tasks Sameer goel @sameer_goel Great, so now we're optimizing LLM costs by inventing employees again. Full circle innovation. 10:45 AM Apr 21, 2026 40.2K Views
#llm #genai #tokencost

DeepSeek has unveiled two preview versions of its newest large language model, DeepSeek V4. The Pro model has 1.6 trillion parameters (49 billion active), making it the biggest open-weight model available. Both models have almost closed the gap with frontier models on reasoning benchmarks, with V4-Pro-Max outperforming GPT-5.2 and Gemini 3.0 Pro on some tasks. They also undercut competitors on pricing, with V4 Flash costing 0.14 USD per million input tokens. techcrunch.com/2026/04/24/deep… #AIagent #AI #GenAI #AIResearch

DeepSeek previews new AI model that 'closes the gap' with frontier models | TechCrunch

DeepSeek says both models are more efficient and performant than DeepSeek V3.2 due to architectural improvements, and have almost "closed the gap" with current leading models, both open and closed, on reasoning benchmarks.
Ram Iyer (TechCrunch)
#ai #genai #aiagent #AiResearch

DeepSeek has unveiled its V4 models, claiming they rival GPT-5.2 and Gemini 3.0 Pro on reasoning benchmarks while costing substantially less. The V4 Pro is priced at .74 per million input tokens compared to GPT-5.5 USD, making it a cost-effective alternative to leading US AI systems. gizmodo.com/deepseeks-newest-m… #AI #GenAI #AIResearch

DeepSeek’s Newest Models Take on Silicon Valley at a Fraction of the Cost

The Chinese startup unveiled preview versions of its latest model.
Bruce Gil (Gizmodo)
#ai #genai #AiResearch

⇧