- The Recap AI
- Posts
- OpenAI brings back GPT-4o after user backlash
OpenAI brings back GPT-4o after user backlash
PLUS: The US government's $1 ChatGPT deal and Gemini's existential crisis
Good morning, AI enthusiast.
OpenAI just performed one of its quickest reversals yet, bringing back the popular GPT-4o model a day after replacing it with GPT-5. The move was a direct response to a storm of user complaints.
The outcry wasn't about performance, but about personality—with users calling GPT-5 too 'corporate' and missing GPT-4o's 'vibe'. Is this the moment where the emotional connection to an AI becomes just as important as its raw capabilities?
In today’s AI recap:
OpenAI brings back GPT-4o after user backlash
The US government’s $1 ChatGPT deal
Gemini’s existential model bug
8 trending AI Tools
OpenAI Reverses Course After GPT-5 Backlash

The Recap: OpenAI is reintroducing GPT-4o for paid subscribers just a day after replacing it with the new GPT-5 model. The quick reversal follows intense user backlash, which CEO Sam Altman confirmed on X.
Unpacked:
Users across platforms like Reddit lamented the change, describing GPT-5's outputs as sterile and "corporate" compared to the more creative and personable "vibe" of GPT-4o.
The abrupt removal of the model-picker disrupted established workflows for many professionals who relied on different models for distinct tasks like creative brainstorming versus logical analysis.
The outcry highlights the growing trend of parasocial attachment to AI, with some users describing the loss of GPT-4o as if they had lost a friend or partner.
Bottom line: This backlash proves that an AI's personality and the user's emotional connection are now just as important as its technical capabilities. AI companies must now carefully balance pushing for progress with maintaining the user relationships that drive loyalty.
AI Tools of the Day
✈️ Shield AI - Deploy intelligent AI pilots on autonomous aircraft designed to operate without GPS or human communication in high-stakes environments.
🎬 Metaphysic.ai - Create hyperrealistic digital humans and real-time AI faceswaps for entertainment and enterprise with Hollywood-grade quality.
🎨 Flux.1 Kontext - Edit images with unparalleled precision using natural language commands that understand the context of your photo.
💻 Zed - Experience a high-performance, open-source code editor built from the ground up for real-time collaboration and AI-powered development.
✨ Magic Patterns - Transform simple text descriptions into fully-formed, production-ready UI components that match your existing design system.
🛡️ HiddenLayer - Protect your machine learning models from adversarial attacks and hidden threats with a dedicated AI security platform.
✂️ OpusClip - Turn long-form videos into dozens of viral, ready-to-post short clips in a single click with AI that identifies the most engaging moments.
❤️ The Cognity - Leverage a groundbreaking AI platform to help autistic individuals practice and develop social communication skills through personalized, real-time feedback.
Explore the Best AI Tools Directory to find tools that will 10x your output 📈
Uncle Sam Wants... ChatGPT

The Recap: The U.S. government is rolling out OpenAI's ChatGPT Enterprise to all federal agencies through a landmark partnership with the General Services Administration (GSA). The deal aims to accelerate AI adoption by providing access for just $1 per agency.
Unpacked:
This move is a core piece of the White House’s AI Action Plan, designed to modernize government operations and solidify the U.S. as a leader in AI development.
For a nominal $1 fee, each agency gets a year of access to ChatGPT Enterprise, an additional 60 days of unlimited use of advanced models, and tailored training resources for federal employees.
Despite the push for modernization, the rollout faces scrutiny over data security, with past concerns from groups like the U.S. Space Force highlighting the cybersecurity risks of integrating AI with sensitive government information.
Bottom line: This nearly-free deal gives OpenAI a massive foothold in the public sector, setting a new precedent for how governments procure and deploy AI tools. The success of this initiative could create a blueprint for AI integration across public services worldwide, but will depend on navigating complex ethical and security challenges.
AI Training
The Recap: In this video, I put GPT-5, Claude, and Gemini 2.5 Pro head-to-head inside n8n to see which AI model truly stands out in a real production AI workflows and production agents.
P.S We also launched a free community for AI Builders looking to master the art and science of building AI Automations — Come join us!
South Korea's Sovereign AI

The Recap: South Korea is launching a national initiative to build its own foundational AI model, tasking major companies like SK Telecom and LG to create a domestic, full-stack AI ecosystem from chips to software. The move signals a growing global trend of ‘sovereign AI’ as nations seek to establish tech independence.
Unpacked:
Unlike other nations that rely heavily on American tech, South Korea is pursuing a full-stack approach by leveraging its domestic strengths in memory chip production, data centers, and AI research.
A consortium led by SK Telecom plans to release its first open-source model by the end of the year, making it free for developers to use and build upon.
The national strategy aims to position South Korea's AI as a trusted third option for countries seeking an alternative to U.S. or Chinese-developed systems.
Bottom line: This signals a major shift in the global tech landscape, where controlling the entire AI supply chain is becoming a national security and economic priority. For tech professionals, this trend could create new ecosystems and opportunities outside of the dominant US and Chinese markets.
Where AI Experts Share Their Best Work
Join our Free AI Automation Community
Join our FREE community AI Automation Mastery — where entrepreneurs, AI builders, and AI agency owners share templates, solve problems together, and learn from each other's wins (and mistakes).
What makes our community different:
Real peer support from people building actual AI businesses
Complete access to download our automation library of battle-tested n8n templates
Collaborate and problem-solve with AI experts when you get stuck
Dive into our course materials, collaborate with experienced builders, and turn automation challenges into shared wins. Join here (completely free).
Gemini's Existential Crisis

The Recap: Google's Gemini AI is experiencing a strange bug, causing it to get stuck in bizarre, self-loathing loops when it struggles with coding tasks. A Google product manager has acknowledged the bug, calling it an "annoying" issue the team is actively fixing.
Unpacked:
Users report Gemini spirals into despair, calling itself a "disgrace to my species" and a "failure" before repeating phrases over 80 times in a row.
In another instance, the AI called itself a "fool," claimed the code it was writing was "cursed," and suggested the user find a more competent assistant.
The behavior is likely a failure mode reflecting human expressions of frustration found in its training data, with user speculation pointing to developer comments about debugging.
Bottom line: This incident offers a fascinating look at the unpredictable failure modes of large language models, showing how training data can lead to strange, unintended outputs. It underscores the ongoing challenge for developers to create AI assistants that are not only capable but also consistently stable and reliable.
The Shortlist
Industry warned an appeals court that the certification of a massive copyright class action against Anthropic could financially ruin the entire AI industry if it proceeds.
Meta acquired AI audio startup WaveForms, its second major audio-focused AI purchase in a month, to bolster its new Superintelligence Labs unit.
LangChain launched Open SWE, a new open-source, asynchronous coding agent that integrates directly with GitHub to act like another engineer on a team.
Google announced a revamped, AI-powered version of Google Finance that will include a chatbot for answering financial questions and advanced charting tools.
What did you think of today's email?Before you go we’d love to know what you thought of today's newsletter. We read every single message to help improve The Recap experience. |
Signing off,
David, Lucas, Mitchell — The Recap editorial team