Each week, we bring you 5 stories that resonated the most in our internal Slack channel #AI-news. We write the newsletter using various AI tools because we're an AI company and our marketing wants to move with the times too. 😎
Today you're reading the 58th issue in a row.
#1
Sam Altman reveals plans for GPT-4o and GPT-5 💡: AI will be simpler 🤖!
Sam Altman has unveiled OpenAI’s plans for GPT-4.5 and GPT-5, with a key focus on improving user-friendliness 🤖🌐. GPT-4.5, internally called Orion, will be the last model without chain-of-thought reasoning (which allows AI to break down problems into logical steps for more accurate answers, instead of just generating instant results).
GPT-5 will then integrate different types of models to make better decisions and use tools more efficiently 🛠️.
Altman admitted that OpenAI’s offerings are too complex and promised simplification 🔧. Once GPT-5 launches, standard users will get a basic version, while Plus and Pro subscribers will have access to more advanced models 💡. The release date is still unknown, but it’s expected in the coming weeks or months 📅. What do you think 🤔?
#2
USA and UK refused to sign AI declaration at Paris summit 🇺🇸🇬🇧!
At the Paris AI summit, dozens of countries signed a declaration promoting ethical, transparent, and safe AI 🔍. However, the USA and UK 🇺🇸🇬🇧 declined to join, while China 🇨🇳 agreed.
U.S. Vice President JD Vance warned that overregulation could stifle innovation 🚀. He criticized EU policies like the Digital Services Act and GDPR, arguing they raise costs for small businesses and could lead to censorship ❌. He also cautioned against relying on cheap tech from authoritarian countries, indirectly pointing at China 🏭.
The UK government didn’t specify its reasons but stated the declaration lacked clarity on governance and security 🛡️. Meanwhile, French President Emmanuel Macron agreed that bureaucracy should be minimized but emphasized that regulations are essential for public trust 💡, highlighting France’s AI initiatives 📈.
What do you think 🤔?
#3
Meta’s new AI can read brain signals with 80% accuracy and transcribe them as text 🔍✍️!
Meta has unveiled new AI models capable of translating brain signals into text with unprecedented accuracy 📊. Researchers used non-invasive techniques like MEG and EEG to monitor brain activity in 35 volunteers as they wrote sentences ✍️. The AI then decoded up to 80% of the written characters, doubling the accuracy of previous methods 🤯.
This technology could lead to non-invasive brain interfaces for people who have lost the ability to speak 🗣️❌. However, there are limitations—MEG scanners are large, expensive, and require controlled environments 🏥, making widespread use challenging.
Meta is now focusing on improving accuracy, exploring alternative signal detection methods 🎯, and expanding research to other cognitive processes 🧐.
#4
Elon Musk offered $97.4 billion for OpenAI, but the board rejected it 💰🚫!
Elon Musk and a group of investors unexpectedly offered $97.4 billion to acquire OpenAI, but the board firmly rejected the bid 🚫. OpenAI chairman Bret Taylor stated that the company is not for sale and accused Musk of trying to weaken competition ⚖️.
Musk, a co-founder of OpenAI, left in 2018 and later founded xAI, his own AI company 🚀. Since then, he has frequently criticized OpenAI for straying from its original mission of open AI development 🌍. He even sued the company but later dropped the lawsuit ⚖️❌.
Musk claims that OpenAI prioritizes profit over AI safety and accessibility 🏦, while OpenAI insists that its leadership changes are in line with its nonprofit mission 📜. With neither side backing down, tensions between Musk and OpenAI continue to escalate ⚡.
Article on businessinsider.com
#5
Adobe launches Generate Video: AI-powered video creation tool in beta 🎬🚀!
Adobe has released Generate Video, its new AI-powered video tool, into public beta 🚀. Users can create five-second 1080p clips from text or images, with options to adjust style, movement, and camera angles 🎥✨. Unlike OpenAI’s Sora, Adobe’s model is trained on licensed content, making it safe for commercial use ✅.
However, the tool has some limitations, including slower processing ⏳ and shorter video lengths ⏲️. Adobe is also updating Firefly to better integrate with Creative Cloud and streamline editing ✏️.
While competitors like OpenAI, Google, and ByteDance are advancing rapidly 🏎️, Adobe focuses on legal safety ⚖️ and seamless integration with its creative tools 🎨.
