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 78th issue in a row.
#1
OpenAI launches ChatGPT Go for $4.6/month in India 🪙📉.
OpenAI has launched a new subscription plan in India called ChatGPT Go, priced at $4.6 per month 🪙 – the cheapest subscription in the company’s history 📉. Users get 10 times more messages and image generations compared to the free version, along with faster responses ⚡.
The goal? To reach hundreds of millions of users in a highly price-sensitive market 📱.
India is already OpenAI’s second-largest market after the U.S., and Sam Altman expects it will soon become number one 🥇. This move confirms the company’s strategy to expand in regions where AI adoption is just beginning 🚀 – and where a low entry cost is crucial 🧮.
#2
Sam Altman 💸: The AI sector is in a bubble – like the dot-com era 💥.
Sam Altman has admitted that the AI sector is currently in a bubble – with investors pouring billions into startups that often have little more than a three-person team and an idea 💸. He compared the situation to the dot-com era, where tech progress was mixed with irrational capital 🌐.
Altman also predicts that many investors will lose money 💥, but says the economy as a whole will benefit from the bubble 📈. He argues that such periods are normal in tech history and help accelerate progress 🧠.
According to him, OpenAI is taking a realistic approach to a possible crash – and continues with plans to invest trillions of dollars into building data centers 🏗️.
#3
USA is testing its own AI platform 🏛️ – combining OpenAI, Google, and Meta tools 🤖.
Federal employees in the U.S. are set to begin testing a new AI platform called USAi, which brings together tools from OpenAI, Google, Anthropic, and Meta 🤖. The goal is to boost government efficiency through chat, search, and coding tools 🏛️.
The administration presents the project as part of modernization efforts, but several experts have raised concerns about the risk of automating civil service roles ⚠️.
To address these concerns, USAi ensures that agency data will not be fed back into commercial AI models 🔒. OpenAI and Anthropic have also offered their tools for a symbolic $1 💵, allowing them to help shape the government’s first AI experiences 🇺🇸.
While the platform is optional, its implementation could influence how AI is used in public institutions over the long term 🧠.
#4
GPT-5 disappointed regular users 😕, but impressed developers 💻.
OpenAI’s GPT-5 is facing disappointment from users who, after a long wait, expected more significant improvements 😕. While the model brought faster speed, lower cost, and fewer hallucinations, its responses feel less natural ✍️. Critics point to a dry, less fluent style and accuse OpenAI of overhyping the release 📉. The company even responded by temporarily bringing back an older model for dissatisfied users 🧑💻.
Despite the criticism, GPT-5 holds a clear lead in programming, currently delivering the best performance in the industry 💻. That makes it a strong tool for businesses and developers who prioritize stability and capability over conversational elegance 🏢.
Even if it didn’t meet all expectations, GPT-5 could still become a major revenue driver and a reliable asset in the business segment 💼.
#5
Meta reorganizes its AI division again 🔄 – the fourth time in six months 😵.
Meta is undergoing another major restructuring of its AI division — the fourth in just half a year 🔄. Superintelligence Labs will be split into four parts: a new team temporarily called TBD Lab, a product branch focused on the Meta AI assistant, an infrastructure team, and the FAIR research group 🧠. The changes follow the departure of several key employees and disappointment over the Llama 4 model, which failed to generate the expected impact 😕.
Despite the internal turmoil, Mark Zuckerberg remains committed to AGI goals and promises further massive investments 💰. Meta is already seeking $29 billion to build new data centers and plans to reach capital expenditures of up to $72 billion by the end of the year 🏗️. These expenses are expected to grow even more in 2026 📈.
