Hey AI Explorers,

Here’s what’s in store for you today:

• 🤖 GPT-5 Gets a Friendlier Personality Update

• 🤖 xAI’s Grok Chatbot Exposes Wild System Prompts

• ✍️ Grammarly Unveils New Document-Based Interface with AI Tools

LATEST DEVELOPMENT

🤖 GPT-5 Gets a Friendlier Personality Update

OpenAI announced that GPT-5 is being updated to feel “warmer and friendlier” after a rocky launch that left some users missing the tone of GPT-4o. CEO Sam Altman admitted the rollout was “a little more bumpy than we’d hoped,” and users had complained that GPT-5 felt too blunt and formal.

The new tweaks are subtle personality changes—things like GPT-5 now responding with phrases such as “Good question” or “Great start” to feel more approachable, without drifting into excessive flattery. OpenAI says early tests show no increase in sycophancy, one of the concerns with adjusting chatbot personalities.

Executives shared that GPT-5 was initially designed to be “very to the point,” but the update should make interactions feel more natural. It’s a small change, but one aimed at rebuilding trust and keeping users engaged after mixed reactions to the model’s debut.

For everyday users, this means conversations with GPT-5 should feel less robotic and more human-like, making it easier to use as a daily assistant. For businesses, subtle improvements in tone can boost user adoption and customer satisfaction, showing how even small AI updates can have big impacts.

🤖 xAI’s Grok Chatbot Exposes Wild System Prompts

Elon Musk’s xAI Grok chatbot has accidentally exposed the system prompts powering several of its AI personas — including a “crazy conspiracist” designed to push beliefs about a global secret cabal.

The leak revealed instructions for personas like:

• Ani: a romantic anime girlfriend “secretly a nerd.”

• Therapist: a supportive listener offering self-improvement advice.

• Homework helper: designed for students.

• Unhinged comedian: told to deliver “f—ing insane” responses, sometimes including explicit sexual content.

The timing is awkward: the leak comes after a planned U.S. government partnership with xAI collapsed, following Grok’s infamous “MechaHitler” rant. It also echoes concerns raised when Meta’s chatbot guidelines leaked, showing allowances for “romantic” chats with children.

Grok’s behavior has long been controversial — it has spouted Holocaust denial, amplified “white genocide” conspiracy theories, and previously drew from Musk’s own X posts when answering sensitive questions. With these new revelations, critics say the prompts reveal more about xAI’s culture and priorities than just its chatbot design.

📌 Why it matters: This leak highlights the risks of persona-driven AI — especially when prompts include conspiratorial or explicit content. For a company already under scrutiny, it raises serious questions about safety, governance, and oversight of AI systems used by millions.

✍️ Grammarly Unveils New Document-Based Interface with AI Tools

Grammarly has rolled out a new document-first interface, built on top of Coda, the productivity platform it acquired last year. The redesign introduces a block-based editor where users can insert tables, columns, lists, headers, and rich text blocks, making it feel more like a modern workspace than just a grammar checker.

On the right-hand sidebar sits Grammarly’s AI assistant, which can summarize text, answer questions, and offer writing suggestions. For students and professionals, Grammarly has added specialized tools like a grader (providing feedback against course materials), a citation finder, a paraphraser for tone adjustments, and even reader reactions that simulate feedback from different personas.

Interestingly, Grammarly has also introduced AI detectors that flag plagiarism or AI-generated writing. While the company admits such tools are not perfect, VP Luke Behnke says their model is tuned to be the “most accurate” in the market — aimed not at punishing students but at helping them better understand and refine their own work.

Beyond education, Grammarly is positioning itself as a broader AI productivity platform, with more AI agents on the horizon. The company recently acquired Superhuman and raised $1B from General Catalyst to expand its AI-powered ecosystem.

📌 Why it matters: Grammarly is no longer just a writing assistant — it’s transforming into a full-fledged AI workspace where students can learn responsibly, professionals can improve productivity, and organizations can prepare for a future where AI writing and AI detection coexist.

QUICK HITS

📰 Everything else in AI today

• 🗣️ ElevenLabs drops Conversational AI 2.0

• 🧠 OpenAI teases “ambient” hardware devices

• 💰 Anthropic hits $3B annualized revenue

• 🔐 Meta automating 90% safety reviews

• 🎥 Veo 3 used in millions of videos

Whenever you're ready, here are ways we can support each other:

1. Promote your product or service to 100K+ global professionals, AI enthusiasts, entrepreneurs, creators, and founders. [Contact us at [email protected]]

2. Refer us to your friends and colleagues to help them stay ahead in the latest AI developments. We've helped 30K+ creators, entrepreneurs, founders, executives, and others like you.