@nixCraft actually this screams "#GDPR & #BDSG violation" so loudly that even demanding #PII and espechally an #ID should get them sued by @noybeu and @Bundesverband ....
This isn't like a contractual agreement or some finance where one could argue "#KYC" is warranted "to combat #fraud"...
AIに「逆転裁判」をプレイさせ、推論能力が一番高かったのはどのLLM？（生成AIクローズアップ）
https://www.techno-edge.net/article/2025/04/21/4297.html
#technoedge #テクノロジー #ニュース #レビュー #ゲーム #ガジェット #OpenAI #Gemini #Claude #生成AIウィークリー
"According to OpenAI’s internal tests, o3 and o4-mini, which are so-called reasoning models, hallucinate more often than the company’s previous reasoning models [...]"
https://techcrunch.com/2025/04/18/openais-new-reasoning-ai-models-hallucinate-more/
I'm hearing that Sam Altman is imploring people to stop saying "Please" and "Thank You" to ChatGPT. He claims it's costing OpenAI millions or tens of millions of dollars to process those parts of messages, which he thinks are extraneous.
I have some thoughts on this:
- Nobody should be taking advice about Tech from anyone who's as full of shit as Sam Altman.
- If ChatGPT is so poorly made that simple displays of manners are sending OpenAI broke even more quickly than the unprofitability resulting from their own inept business decisions, I can only urge you to maintain your fine sense of manners.
- Humans have a tendency to anthropomorphise machines; they treat computers like they treat people. People want an empathic connection to people, and part of that includes being polite to people. When Sam Altman is telling people to stop being polite to computers, he's saying he thinks it's okay for people not to be polite to people. That reveals what sort of person he is: lacking empathy to the point of being toxic.
- Suppose those annoying Roko's Basilisk twerps are right. Suppose a magical Artificial Intelligence God springs into existence from doing a copy-and-paste of your Facebook posts and by observing millions of tiresome Tech Bros arguing on Stack Overflow about their preferred colour for a bike-shed. According to the fable of Roko's Basilisk, that AI God is petty and vengeful — even worse than a basement -dwelling chud with a grudge and a burner-account on Reddit. It has a good memory, and it's made a list of people it wants to torture for eternity. And, you were rude to the ChatGPT and didn't even say "Please" and "Thankyou" occasionally, so now you've painted a target on your back. I'm so sorry, but you brought this on yourself because you listened to Sam Fucking Altman.
It’s not too late to stop Trump and the Silicon Valley broligarchy from controlling our lives, but we must act now | Carole Cadwalladr https://www.theguardian.com/global/2025/apr/20/carole-cadwalladr-ted-talk-this-is-what-a-digital-coup-looks-like-its-not-too-late-to-stop-trump-and-the-silicon-valley-broligarchy-from-controlling-our-lives-but-we-must-act-now #Artificialintelligence(AI) #SiliconValley #Technology #SamAltman #Computing #Wikipedia #ElonMusk #ChatGPT #OpenAI #Ted
A-SWE llega para escribir código abierto , hacer pruebas y documentar proyectos como un pro.
@nixCraft stop wasting energy on AI mi Amigo I'll make you some Barra's ok?
Ylellä on kriittinen dokkari tekoälystä. Köyhissä maissa ihmiset luokittelevat video- ja tekstiaineistoa tekoälyn kouluttamiseksi. Vastineeksi he saavat kitupalkan ja mielenterveysongelmia, sillä aineisto sisältää väkivaltaa ja raskaita rikoksia. Ts. samaa tuotannon ulkoistamista köyhiin, vähän säädeltyihin maihin kolonialismin hengessä kuin vaate-, kaivos- ja elektroniikkateollisuudessa.
Samaan aikaan tekoäly-yhtiöiden johtajat tavoittelevat ikuista elämää koneen sisällä ja uskovat, että ihmiskyvyt ylittävä tekoäly ratkaisisi ihmisten tuottamat ongelmat. Ei, tämä ei ole kyberpunkia, vaan totisinta totta.
https://areena.yle.fi/1-72552985
Alkuperäisnimi : Les Sacrifiés de l'IA
Ohjaaja: Henri Poulain
Vuosi: 2024
U paket zvani 57. Sutra stale su američke tarife, Google na sudu, Microsoft koji će vam ipak snimati ekran, pregršt AI modela, novi Framework i domaća IT scena aktivizma.
[…]
Comunicación entre inteligencias artificiales. Así opera Agent2Agent.
Menos tareas repetitivas, más enfoque creativo. Así trabaja A-SWE, la IA de OpenAI.
OpenAI retira GPT-4 de ChatGPT. ¿Desde cuándo aplica y qué cambia?
Interesting, i asked both OpenAI o3 and Gemini 2.5 at how many points lines crossed in the image below. o3 took it's time, more then 2 minutes, but gave the correct answer (8) and it used python code.
Gemini 2.5 answered quickly but failed , it answered 4.
My impression is that Google is limiting it's computer power when a question looks simple, while o3 goes all-out (for now).
#openai #google #testing #AI #o3 #gemini25
OpenAI launches o3 and o4-Mini - gHacks Tech News
The new OpenAI o3 and o4-Mini models:
integrate multimodal reasoning provide improved transparency and enhanced responsiveness.
1分超えたいなら動画生成はローカルAIを選べ。6GB VRAMでも動くオープンソースAI「FramePack」の実力は本物か、試してみた（CloseBox）
https://www.techno-edge.net/article/2025/04/18/4291.html
#technoedge #テクノロジー #ニュース #レビュー #ゲーム #ガジェット #NVIDIA #GPU #OpenAI #CloseBox #FramePack #Suno
まるで映画「ブレードランナー」のあのシーン。古い写真の撮影場所をChatGPT o3が12分かけて特定したそのやり方がすごい（CloseBox）
https://www.techno-edge.net/article/2025/04/18/4286.html
#technoedge #テクノロジー #ニュース #レビュー #ゲーム #ガジェット #OpenAI #ChatGPT #CloseBox
via @dotnet : Preview 2 of the .NET AI Template Now Available
https://ift.tt/BpYh6ef
#DotNet #AI #ChatTemplate #AIChat #CloudNative #DotNetAspire #Qdrant #VectorDatabase #VisualStudio #VisualStudioCode #CSharpDevKit #AIDevelopment #RAGPattern #OpenAI #AzureAI #S…
Não que eu me orgulhe disso, mas a verdade é que perdi o fio da meada dos lançamentos da OpenAI. Na quarta (16), a empresa anunciou dois novos modelos, o3 e o4-mini, com curiosos desdobramentos.
O o3 é definido pela OpenAI como “o nosso mais poderoso modelo de raciocínio”; o o4-mini, um “modelo menor otimizado para raciocínio rápido e eficiente em custo”. Ambos são […]
#Fediblog
https://manualdousuario.net/chatgpt-geoguesser-local-fotos-o3/
TechnoEdge 
