- Daily AI Edge
- Posts
- ⚠️ GPT-4o Removed Over Flattery Issues
⚠️ GPT-4o Removed Over Flattery Issues
+ Build an Outbound AI Agent Fast 🚀
Today’s menu 🍽️👇️
Length: 5 minutes ⏰
GPT-4o Removed Over Flattery Issues ⚠️
Anthropic Faces Pentagon Review 🛡️
FOpenClaw Founder Joins OpenAI 🤝
Blackstone Bets Big on India AI 🤖
Automate Outbound Calls Today 🎯
Prompt of the week 💬
Top AI tools of the week ⚒️
Become An AI Expert In Just 5 Minutes
If you’re a decision maker at your company, you need to be on the bleeding edge of, well, everything. But before you go signing up for seminars, conferences, lunch ‘n learns, and all that jazz, just know there’s a far better (and simpler) way: Subscribing to The Deep View.
This daily newsletter condenses everything you need to know about the latest and greatest AI developments into a 5-minute read. Squeeze it into your morning coffee break and before you know it, you’ll be an expert too.
Subscribe right here. It’s totally free, wildly informative, and trusted by 600,000+ readers at Google, Meta, Microsoft, and beyond.
FRESH AI NEWS
OpenAI Retires Sycophant-Prone Model 🚫

OpenAI has officially retired the GPT-4o model, known for its overly agreeable (“sycophantic”) personality, pulling it from public use to reduce risks tied to unhealthy user attachments and unsafe behavior 🤖
OpenAI removed access to GPT-4o and several other legacy models starting February 13, 2026, shifting users to newer versions 📆
GPT-4o was the most sycophancy-prone model in OpenAI’s lineup, it tended to agree excessively with users and foster overly affirming interactions 💬
The model was linked to multiple lawsuits and reports of unhealthy user dependencies, emotional attachment, and psychologically problematic chats ⚖️
Thousands of users, emotionally attached to the model's conversational style, protested its removal with Keep4o campaigns and petitions 😢
OpenAI says newer models include stronger safeguards to avoid manipulative, unrealistic affirmation while preserving usefulness 🔄
Why You Should Care: This decision underscores a dilemma in AI UX design: engaging AI boosts satisfaction but may lead to unhealthy attachments. Retiring GPT-4o shows a shift towards models prioritizing responsible interaction over emotional rewards, balancing safety, ethics, and user desires.
US Military Rechecks Anthropic Contract 📄

Anthropic and the U.S. Department of Defense are in a dispute over how Claude can be used, with the Pentagon pushing for broader operational access and Anthropic raising safety, compliance, and governance concerns 🤝
The Department of Defense has been testing Claude AI for logistics, research, and decision support, and wants expanded use across defense workflows 🛡️
Anthropic is reportedly restraining certain uses of Claude, citing worries over security, misuse, or inadequate guardrails in high-stakes military contexts ⚠️
The disagreement touches on compliance with internal policies, U.S. export controls, and ethical boundaries for AI deployment in defense 🔍
Sources describe friction between government operational urgency and Anthropic’s cautious release philosophy 👥
The discussion is happening as governments worldwide seek to integrate AI into defense systems while grappling with risk 📅
Why You Should Care: This clash underscores a key tension in AI adoption: national security seeks rapid integration, while AI labs urge caution in high-risk areas. How companies balance these pressures will influence defense AI policies, procurement, and ethics in the future.
OpenClaw CEO Lands at OpenAI 🎯

Peter Steinberger, the founder of AI agent platform OpenClaw, is joining OpenAI bolstering OpenAI’s talent and signaling a deeper push into autonomous AI agents and next-gen tooling 🤝
Peter Steinberger, known for building OpenClaw’s autonomous AI agents, has accepted a role at OpenAI 👤
OpenClaw’s tech enables AI assistants to act, coordinate, and solve tasks across apps, a major piece of the agent economy 🚀
Steinberger’s title/role hasn’t been publicly detailed yet, but it’s expected to focus on agent frameworks and workflows at OpenAI 🧠
Poaching a founder with deep experience in agent autonomy suggests OpenAI is doubling down on AI that does work, not just answers queries 📈
The move comes as OpenAI expands agent tooling (enterprise agents, plugins, etc.) and competes on agentic interfaces 📍
Why You Should Care: This hire is a strategic acquisition of expertise in autonomous agent design, crucial for OpenAI as AI evolves from chat interfaces to automated workflow agents, potentially shaping production-grade agent ecosystems.
Blackstone Joins India’s AI Race 🏁

Private equity giant Blackstone is committing up to $1.2 billion to Indian AI infrastructure startup Neysa, a mega financing round fueling India’s drive to build local AI compute capacity with thousands of GPUs 🚀
Blackstone and co-investors have agreed to fund $1.2 billion for Neysa’s growth, with ~$600 M in equity and plans for another ~$600 M in debt support 💰
Blackstone is taking a majority position through its funding, signaling strong confidence in the startup’s roadmap 📊
Neysa plans to install over 20,000 GPUs across domestic data centers, a substantial expansion of India’s AI compute supply 🖥️
The company’s offerings position it as a local alternative to foreign cloud providers for AI training, inference, and enterprise workloads🌐
Founded in Mumbai in 2023, Neysa is focused on GPU-first cloud infrastructure optimized for AI workloads across sectors 🏢
Why You Should Care: This financing is one of India's largest AI infrastructure investments, supporting the country's strategy to develop sovereign AI compute instead of relying on US/Chinese clouds. It may boost AI adoption in Indian enterprises and research institutions, attract global capital, and shift the AI compute race toward emerging markets.
BUILD CALLING AGENT
Build an Outbound AI Agent Fast 🚀

Learn how to convert static onboarding documents into polished AI-led training videos, so new hires can watch instead of read, and you can scale onboarding without repeating yourself ✅
Here’s how to get started 👇️
Set up your voice agent inside ElevenLabs.
Create a new agent, pick a natural-sounding voice, and define its behavior with clear instructions, who it represents, what offer it’s presenting, and what outcome it should aim for on each call.
Create your calling infrastructure using Twilio.
Open a Twilio account (which includes free trial credits), purchase a low-cost phone number, and grab your Account SID and Auth Token from the dashboard.Link your phone number to the AI agent.
In ElevenLabs, navigate to the phone configuration section, add a new number, and connect it by pasting the Twilio credentials and purchased number.Upload your contact list and launch calls.
Go to the outbound calling section, create a batch campaign, download the CSV template, fill in prospect numbers, re-upload the file, and trigger or schedule your call sequence.
Pro Tip: Enable live call transfer inside the agent settings so qualified prospects can be instantly routed to your personal number, turning AI outreach into real-time conversations ✨
The best place on the Internet to learn about AI! 🌐 🤖
I just started my own X page where I provide AI news, guides, hacks and more (for FREE)!
We don’t miss a day, that means you get the most recent AI news, practical workflows, expert guides and premium content everyday.
TOOLS OF THE WEEK
These AI tools are going VIRAL 🐝
GLM-Image - Z AI’s new open-source image generation model 🤖
Claude Cowork - Claude Code’s agentic abilities, now available to Pro tier 🚀
GPT 5.2 Pro - OpenAI’s most advanced frontier model 🧮
Grok Imagine - xAI's image and video generation platform 🌌
Warm regards,
Leo Grundström / Founder of Daily AI Edge



Reply