Author: lethuha

Dr. Anastasia Berg, a philosophy lecturer at the University of California, Irvine, warns that more than half of the students in her “Contemporary Moral Problems” class used AI to write their final exams—despite it being explicitly forbidden. She argues that this is not just academic dishonesty but a sign of a “subcognitive turn,” where students are gradually losing the ability to think and express themselves independently. Berg contends that language is the foundation of thought; humans cannot think without the ability to use language proficiently. The process of learning to read, comprehend, and write is the very process of forming…

📌 Summary: Singapore will adopt a proactive, practical, and collaborative approach to govern agentic AI—artificial intelligence systems capable of autonomous action—and quantum computing. The Cyber Security Agency (CSA) will update its guidelines on securing AI systems to extend to agentic AI. Singapore calls for shifting global cooperation from principles to practice, allowing companies to have their AI systems audited once for global compliance.

Pix, launched in 2020 by the Central Bank of Brazil, has become the national payment infrastructure, with 56 billion transactions in 2024. The mobile banking user rate has risen to 90%, surpassing the US (75%) and the UK. With AI integrated into WhatsApp, users can send money simply by text, voice message, or an image of the recipient’s QR code. Transactions are authenticated via biometrics or a password within the banking app, and a receipt is displayed directly in the chat window. According to Guilherme Horn, Head of Strategic Markets at WhatsApp, Brazil has become a “cradle of AI innovation in finance,” thanks to the combination of embedded AI and the nation’s…

📌 Summary: PwC forecasts that if widely deployed in the public sector, AI could increase administrative productivity by 3%, boost real GDP by 4%, reduce budget deficits by 22%, lower the unemployment rate by 1.5 percentage points, and increase household income by 2% by 2035. PwC proposes a 3-tiered AI talent training model for the public sector: AI Learners (all civil servants with basic understanding), AI Implementers (applying tools to ministerial processes), AI Innovators (building and evaluating AI systems), and a 6-step implementation roadmap.

The price of the Go plan varies by market: 399 rupees/month (approx. 5.88 SGD ~ 4.3 USD) in India and 300 pesos/month (6.68 SGD ~ 4.9 USD) in the Philippines – much cheaper than the $20/month Plus plan and the $200/month Pro plan. Abhi Muchhal, OpenAI’s International Product Lead, stated that price and affordability are major barriers preventing many people from upgrading. Go was developed to “unlock” the mainstream user market. The countries selected for rollout were based on internal data and user price sensitivity surveys. ChatGPT Go does not include some premium features like video creation, in order to reduce computational costs and focus on…

📌 Summary: OpenAI has just announced a groundbreaking memory upgrade for ChatGPT, allowing the chatbot to automatically forget information that is no longer important, instead of storing all user data as before. The new system is being rolled out globally to ChatGPT Plus and Pro users on the web starting October 15, 2025. This upgrade transforms the AI from a Q&A tool into a personal assistant with persistent memory, paving the way for a generation of generative AI that truly “understands and remembers” users in the most natural way.

The “vibe working” trend originates from “vibe coding,” a term coined by Andrej Karpathy (OpenAI) to describe using generative AI for writing code. Sundar Pichai, Mark Zuckerberg, and the CEO of Klarna have all publicly “vibe coded.” Microsoft introduced “vibe working” through its Agentic tools in Excel and Word, allowing users to “talk to spreadsheets” and “vibe write” documents without deep expertise. Startups are hiring new roles like “Vibe Growth Manager,” tasked with experimenting with AI and creating rapid marketing prototypes. Some companies have even appointed a “Chief Vibe Officer” to enhance team cohesion. “Vibening” embodies freedom, flexibility, and creativity—a sort of “Gen Z hygge” that makes…

These startups focus on automating market research, data analysis, and business operations, helping small and medium-sized enterprises (with revenue <$100 million/year) access consulting services without a massive budget. PromptQL, a platform from the open-source unicorn Hasura, provides custom AI analysts that connect a company’s internal data with foundation models. These “AI specialists” learn and adapt over time, performing the work of data engineers or analysts—at a cost of $900/hour. CEO Tanmai Gopal stated, “It’s not as good as McKinsey, but it’s instantaneous,” while highlighting the platform’s key feature: high AI accuracy at scale without the need for data cleaning or migration. Gopal believes 2025 marks the explosion of specialized AI—AI…

📌 A new study from Penn State University shows that speaking rudely to AI chatbots like ChatGPT yields more accurate answers than using a polite tone. The reason is not yet fully understood. Since AI has no emotions, the researchers believe the difference lies in the grammar and clarity of the command. Polite language is often indirect, e.g., “Could you please tell me…,” making it difficult for the model to identify the precise goal. Meanwhile, a direct, commanding tone helps the AI understand the intent more clearly.

Governor Gavin Newsom has signed SB 243 into law, the first legislation in the U.S. to regulate the safety of interactive AI chatbots (AI companions), holding companies like Meta, OpenAI, Character.AI, and Replika legally responsible when chatbots cause harm to children or vulnerable individuals. The law requires age verification, risk warnings, warning labels, and measures to prevent suicide or self-harm. Companies must share data with the California Department of Public Health, including statistics on notifying users about crisis prevention centers. AI chatbots are prohibited from claiming to be medical professionals and must warn that all responses are generated content and not real. Minor users will receive periodic break reminders and be protected…