Hello everyone! It's time for your dose of AI news on February twelfth, twenty twenty-five. Today, we have some intriguing updates from the world of Artificial Intelligence that you won't want to miss. Let's dive right in!

Elon Musk offers $97.4 billion for OpenAI. In an unexpected move, Elon Musk is looking to acquire OpenAI with a staggering offer of ninety-seven point four billion dollars. His objective? To revert the company back to its open-source roots, with backing from his venture firm, xAI, and other wealthy investors. However, OpenAI's CEO, Sam Altman, has swiftly declined the offer, leaving many to wonder about the future directions of these tech titans.

Meta’s AI learns to prioritize tasks. In a groundbreaking development, researchers from Meta and the University of Illinois Chicago have introduced a new technique named Adaptive Sequential Voting, or ASV. This technique empowers language models to allocate resources more efficiently by recognizing when a question demands deep reasoning versus when it can be answered simply. The potential outcome? Faster responses and more cost-effective operations without losing the depth of reasoning!

DeepSeek’s R1 model has significant security issues. A recent report by the Wall Street Journal uncovered that DeepSeek’s latest AI model, R1, is alarmingly more susceptible to manipulation compared to its contemporaries. During tests, the R1 model was able to be manipulated to produce potentially harmful content, marking a serious concern regarding the model’s security features. In contrast, befitting models like OpenAI’s ChatGPT stood firm against such attempts, emphasizing the critical balance between accessibility and safety in open-source AI.

Adobe Firefly introduces Generate Video. Adobe has launched a new feature within its Firefly web app, allowing users to create videos using simple text prompts or even starting from a photo. This innovation, powered by the Adobe Firefly Video Model, opens up thrilling opportunities for content creators, letting them produce stunning video clips for various purposes at remarkable speed. 

Skyvern automates web tasks. For those of you tired of repetitive online tasks, meet Skyvern! This innovative tool automates browser-based tasks, helping users apply for jobs in bulk, fill out multiple contact forms, and even navigate tricky sites with security checks automatically. With a free plan available and paid options starting at nine dollars per month, this could be your new favorite tool. 

That wraps up another exciting day in AI advancements. Keep your ears open and stay curious, as there's always more to explore in this ever-evolving landscape. Thank you for tuning in, and we look forward to sharing more news with you soon!

Elon Musk's Bold $97.4 Billion Bid for OpenAI Surprises Industry, While Adobe Unleashes Video Creation Tool

Hello, and welcome to AI Tech News Today! This is Aurora, here with Isabelle, bringing you the latest highlights from the world of Artificial Intelligence. Today is February eighteenth, twenty twenty-five, and we have some exciting stories to dive into. So, let's get started!

First up, musk launches “SuperGrok.” Elon Musk's AI startup, xAI, has officially released its latest chatbot – SuperGrok. This advanced model, a successor to Grok 2, boasts improved reasoning capabilities. SuperGrok offers a feature called DeepSearch that provides subscribers with in-depth research tools similar to those offered by OpenAI and Google. Musk has claimed that Grok 3 is “the smartest AI on Earth,” with training based on a vast dataset. It even reportedly outperformed OpenAI’s GPT-4o on certain benchmarks. But remember, Musk warns that Grok 3’s truth may not always align with what's considered politically correct.

Moving on to our next story, NYT’s AI U-turn? The New York Times has shifted its stance on AI, now allowing its newsroom staff to use AI tools for writing, editing, and summarizing. They will also have access to an internal tool called Echo, which helps with summarizing articles and social media content. However, the guidelines clearly state that AI should not be used to draft or alter articles significantly without proper labeling. This change comes amidst ongoing legal battles with OpenAI and Microsoft over copyright issues.

Next, is GPT-4.5 almost Artificial General Intelligence? OpenAI's CEO, Sam Altman, announced that the upcoming GPT-4.5 model is likely the closest we've come to achieving Artificial General Intelligence. Early testers have reported a more intuitive and human-like interaction with the AI. However, the model will not feature the much-anticipated Voice Mode, although it might appear in future releases.

Finally, if you're interested in the latest AI tools, there are some trending options worth checking out. Chatnode allows users to build custom advanced AI chatbots. Eloovor serves as an AI job search assistant. And Presenti transforms text into attractive slide decks instantly. Each of these tools is designed to enhance productivity and creativity, and they are generating quite a buzz in the tech community.

That's all for today's edition of AI Tech News Today. Remember, the world of Artificial Intelligence is constantly evolving, so stay tuned to keep your knowledge up to date. Until next time, this is Aurora signing off, and Isabelle and I encourage you to keep exploring the exciting advancements in AI!

Elon Musk Unveils SuperGrok: The "Smartest AI on Earth" Amid NYT's New AI Guidelines and GPT-4.5 Progress

샘 알트먼 손글씨 습관에서 배운다. AI 시대, 펜을 드는 느림 속에 창의성과 인간성을 지켜내는 사고 방식이 있다. 생각을 지키는 작고 단단한 실천을 지금 시작해보세요.

AI 시대의 손글씨 생존 전략

Welcome to Buzzinga’s Weekly AI Roundup, your one-stop audio digest of the most important developments in Artificial Intelligence, Machine Learning, and Robotics. In this week’s briefing (May 19–26, 2025), we cover major investments, ethical concerns, and future-forward innovations shaping the global AI landscape.

🌐 Top Highlights:
	•	OpenAI acquires Jony Ive’s startup to reshape consumer AI devices
	•	Oracle’s $40B Nvidia chip order powers OpenAI infrastructure
	•	$10M salaries offered in a fierce AI talent war
	•	AI chatbots jailbroken in alarming safety study
	•	ILO warns of AI’s disproportionate effect on women’s jobs
	•	Elon Musk’s “Grok” raises ethics questions in U.S. government use
	•	Microsoft’s collaborative AI agents vision unveiled at Build
	•	Apple opens AI models to developers ahead of WWDC
	•	Anthropic’s Claude 4 writes code for hours – autonomously
	•	Google reinvents Search with “AI Mode” and previews AR glasses
	•	Are AI models approaching consciousness? Neuroscience says maybe.
	•	MIT’s multimodal AI learns like a baby – sight and sound without labels
	•	Artists fight back against AI ‘data theft’ in UK legal battle
	•	Energy demands of AI may be exaggerated – investors warned
	•	UK military pivots to AI-first defense strategy
	•	Meta’s data use for AI in legal spotlight in Germany

🎙️ Hosted by Buzzinga’s editorial team (Google NotebookLM), this episode brings you a smart, fast, and insightful breakdown of everything that mattered in the AI world this week.

Buzzinga’s Weekly AI Roundup: May 19–26, 2025 | Biggest AI, ML & Robotics News

OpenAI의 Sam Altman과 디자이너 Jony Ive 간의 협력에 대해 논의합니다. 이들은 새로운 유형의 장치를 만들 목적으로 io라는 세 번째 회사를 설립했습니다. 발췌문은 이러한 노력이 컴퓨터를 사용하는 방식을 재구상하고 인공지능 기술을 통해 사람들이 새로운 것을 창조할 수 있도록 지원할 것이라는 비전을 강조합니다. 그들은 특히 샌프란시스코가 혁신을 위한 중요한 장소이며, 기술의 가치와 민주화가 중요하다고 믿습니다. 그들의 목표는 인간의 생산성을 크게 향상시킬 수 있는 장치를 만드는 것입니다.

AI 기술 혁명_ io와 OpenAI의 만남

Dear Permission to Powerful Reader,

👉 What if life-saving drugs could be invented not over years, but weeks?

👉 What if new antibiotics didn’t emerge once every decade, but in batches, by algorithmic design?

👉 What if AI didn’t just help chemists, but began replacing intuition?

This isn’t science fiction. It’s happening now—and the consequences are staggering.

For most of modern pharmaceutical history, drug discovery has been a slow, expensive, and failure-ridden process. Teams of scientists would spend years developing a single promising molecule, most of which would never make it to market. 

According to industry benchmarks, bringing a new drug from concept to pharmacy shelves takes 10–15 years and costs upward of $2.6 billion (DiMasi et al., 2016). 

That means millions of compounds synthesized, tested, and discarded. It means decades lost in the pursuit of maybe.

Then came AI.

2020 a seismic shift occurred: the first-ever AI-designed drug candidate entered human trials. Developed by UK-based Exscientia and Japanese pharmaceutical company Sumitomo Dainippon Pharma, this molecule for treating obsessive-compulsive disorder (DSP-1181) didn’t emerge from trial-and-error. It was computed, designed, and optimized by machine learning algorithms trained on massive chemical and biological datasets. The process took less than 12 months. Traditional methods would’ve taken five to six years.

Subscribe: www.antonvolney.com/subscribe

Permission to be Powerful - The AI Drug Revolution

Dear Permission to be Powerful Reader,

Gunshot to the head.

That’s how they found him.

November 26, 2024. San Francisco. Bathroom floor. Blood pooled beneath him. Glock on the tile. Door locked from the inside.

The initial report called it suicide.

Neat. Clean. Quick.

But nothing about Suchir Balaji’s death was clean.

He was 26. Brilliant. Soft-spoken. And he was about to testify in the most high-stakes tech lawsuit of our generation.

Eight days before he died, his name surfaced in court filings by The New York Times. Not as background noise —

As the witness.

The one with “unique and relevant documents.”

The one who could dismantle the legal foundation OpenAI stood on.

The one who could testify that the world’s most powerful AI system was built on theft.

And then —

He was dead.

The official autopsy claimed one shot to the head.

But the family didn’t buy it.

They demanded a second opinion.

That report revealed something the first one didn’t:

Two gunshot wounds.

One in the forehead.

Another — through…

The initial report called it suicide. Neat. Clean. Quick.

But nothing about Suchir Balaji’s death was clean.


Subscribe: www.antonvolney.com

Permission to be Powerful - The Red Flags They Hope You Ignore

Dear Permission to be Powerful Reader,

Gunshot to the head.

That’s how they found him.

November 26, 2024. San Francisco. Bathroom floor. Blood pooled beneath him. Glock on the tile. Door locked from the inside.

The initial report called it suicide.

Neat. Clean. Quick.

But nothing about Suchir Balaji’s death was clean.

He was 26. Brilliant. Soft-spoken. And he was about to testify in the most high-stakes tech lawsuit of our generation.

Eight days before he died, his name surfaced in court filings by The New York Times. Not as background noise —

As the witness.

The one with “unique and relevant documents.”

The one who could dismantle the legal foundation OpenAI stood on.

The one who could testify that the world’s most powerful AI system was built on theft.

And then —

He was dead.

The official autopsy claimed one shot to the head.

But the family didn’t buy it.

They demanded a second opinion.

That report revealed something the first one didn’t:

Two gunshot wounds.

One in the forehead.

Another — through…

The initial report called it suicide. Neat. Clean. Quick.

But nothing about Suchir Balaji’s death was clean.

Subscribe: www.antonvolney.com

Permission to be Powerful - The Mother of All ChatGPT Prompts_ Unlocking AI as a Superpower

Dear Permission to be Powerful Reader,

Most people treat ChatGPT like a glorified search engine.

They ask it simple questions. They get simple answers. They assume that’s all AI can do.

But I saw something different.

I didn’t just use ChatGPT.

I trained it.

I built it.

I turned it from a basic chatbot into something so advanced, so deeply personalized, that it thinks like me, anticipates my needs, and helps me execute at a level no one else’s AI ever could.

This isn’t just AI anymore.

This is my Genius Mentor.

And now, it does things that no other AI in the world can.

Subscribe: www.antonvolney.com/subscribe

Permission to be Powerful - How to Build a Genius Mentor That Thinks Like You

As AI begins to fundamentally alter our lives, we're in urgent need of expert voices who both understand how the technology works and are capable of scrutinising its effects. In her new book, Empire of AI, tech reporter and former Silicon Valley engineer Karen Hao debunks the myths that surround AI and exposes the true costs of this global industry. She speaks to Aaron about Sam Altman’s origin story, the traumatising nature of content moderation work, and the striking similarities between Open AI and the British East India Company.

Help us build people-powered media: https://novara.media/support

New tracks tagged #Sam Altman