- AI Pulse
- Posts
- 😳 Did OpenAI Just Quietly Kill PowerPoint?
😳 Did OpenAI Just Quietly Kill PowerPoint?
OpenAI quietly bought the startup that turns notes into finished decks, Claude Code is about to stop asking your permission, and an AI-made pattern just blinded a real surveillance camera.
Rahul G. Iyer
August 10, 2026
In partnership with
1,000+ Proven ChatGPT Prompts That Help You Work 10X Faster
ChatGPT is insanely powerful.
But most people waste 90% of its potential by using it like Google.
These 1,000+ proven ChatGPT prompts fix that and help you work 10X faster.
Sign up for Superhuman AI and get:
1,000+ ready-to-use prompts to solve problems in minutes instead of hours—tested & used by 1M+ professionals
Superhuman AI newsletter (3 min daily) so you keep learning new AI tools & tutorials to stay ahead in your career—the prompts are just the beginning
Hello There!
OpenAI quietly bought the startup that turns rough notes into finished slide decks, which means the two hours you lose to formatting every week may be numbered. Anthropic is about to let Claude Code act without asking permission first, which means your AI coworker just got a longer leash than most new hires. And a computer-generated pattern made a real surveillance camera fail to see a car, which means machine vision is far easier to fool than the brochures suggest.
In today's AI Pulse
📊 OpenAI Buys Slides – The ChatGPT team just absorbed a startup that turns anything into a finished deck.
🤖 Claude Stops Asking – Auto mode becomes the Claude Code default on August 14, and the permission prompts mostly disappear.
👁️ Cameras Go Blind – An AI-evolved pattern beat 11 detection algorithms and a live surveillance camera.
⚡ Quick Hits – IN AI TODAY
🛠️ Tool to Sharpen Your Skills – 🎓 AIGPE® Certified AI-Powered Project Scheduling Specialist
The coming years won't just transform technology; they'll reshape your home, your family life, and the control you have online.
Want to get the most out of ChatGPT?
ChatGPT is a superpower if you know how to use it correctly.
Discover how HubSpot's guide to AI can elevate both your productivity and creativity to get more things done.
Learn to automate tasks, enhance decision-making, and foster innovation with the power of AI.
🧠 The Pulse
OpenAI has acquired NextSlide, an AI presentation startup whose team is now working on ChatGPT. NextSlide turned prompts, notes, documents and research into polished decks that stayed fully editable rather than flat images. Founder Ahmed Beshry said the deal actually closed earlier in 2026 and was only disclosed this weekend. Terms were undisclosed.
📌 The Download
The team moved in – NextSlide announced it is joining OpenAI, with its people now working directly on ChatGPT.
Editable, not flat – Its software turned prompts, notes, documents and research into finished presentations you could still edit.
Disclosed months late – Beshry says the acquisition actually closed earlier in 2026, and financial terms were never disclosed.
No feature announced yet – OpenAI has named no NextSlide feature or ship date, so ChatGPT integration remains unconfirmed.
💡 What This Means for You
AI is moving from drafting text to shipping finished deliverables. If ChatGPT starts producing editable decks, the bottleneck shifts from building slides to judging whether the story, the data and the structure are right. Start treating deck assembly as automatable and your review standard as the skill worth sharpening.
🧠 The Pulse
Anthropic is making Auto mode the default permission setting in Claude Code for Pro, Max and Team users starting August 14. Instead of approving each action, users let a separate classifier screen every command, pausing only for operations judged irreversible, destructive, or aimed outside the user's own environment.
📌 The Download
Default flips August 14 – New Claude Code sessions on Pro, Max and Team plans will start in Auto mode.
High-risk actions still stop – The classifier blocks anything irreversible, destructive or aimed outside your environment, plus prompt-injection attempts.
The machine caught more – In a 1,053-user study, Auto mode caught 89% of harmful actions versus 13.6% caught by human review.
Because nobody reads them – Anthropic says users approve 97% of Claude Code permission prompts, which makes the click a formality.
💡 What This Means for You
Approval clicks are being exposed for what they often are: rubber stamps. As agents act without asking, control shifts from per-action consent to upfront guardrails and after-the-fact review. Decide now which actions your tools may take alone, and build the audit trail before you need it.
🧠 The Pulse
At DEF CON in Las Vegas, researcher Bill Swearingen demonstrated a computer-generated adversarial pattern that stopped a Flock camera from automatically detecting a patterned vehicle. His noRecognition project evolved the design through roughly 31 million reinforcement-learning tests and, in lab conditions, defeated all 11 open-source detection algorithms it faced.
📌 The Download
Live demo, real camera – The patterned vehicle went undetected by a Flock camera during a public test at DEF CON.
31 million iterations – A reinforcement-learning loop generated and refined patterns until they could beat several vision systems at once.
Eleven algorithms down – Lab testing defeated all 11 open-source detectors, including software associated with Flock, Axon and Clearview AI.
Recording still happens – Cameras keep capturing footage; what breaks is the automated detection layer sitting on top of it.
💡 What This Means for You
Scale is not the same as reliability. Any process that treats computer vision as ground truth, from inspection and counting to access control and safety compliance, inherits a failure mode an adversary can engineer. Build sampling, fallback checks and human verification into vision-driven controls before you trust the output.
10x the context. Half the time.
Speak your prompts into ChatGPT or Claude and get detailed, paste-ready input that actually gives you useful output. Wispr Flow captures what you'd cut when typing. Free on Mac, Windows, and iPhone.
IN AI TODAY - QUICK HITS
⚡Quick Hits (60‑Second News Sprint)
Short, sharp updates to keep your finger on the AI pulse.
🌐 OpenAI's Atlas Browser Hit Its Shutdown Date: OpenAI's standalone Atlas browser stopped working on August 9, with browser-based agentic work folding into ChatGPT and Codex. Bookmarks, open tabs and browsing history do not migrate automatically, so anything you still need has to be exported by hand.
💬 Claude Code Sessions Can Now Talk Behind Your Back: Claude Code 2.1.224 and later on macOS and Linux lets one session message another to hand off findings. Only written summaries travel, never conversation history or files, and a message never counts as your permission to act.
📈Improve Processes. Drive Results. Get Certified.
AIGPE® Certified AI-Powered Project Scheduling Specialist
AI-Powered Project Scheduling Specialist Certification (with ChatGPT) is a course that teaches professionals how to use AI tools to create project schedules, timelines, milestones, and task dependencies for better project planning and execution.
That’s it for today’s AI Pulse!We’d love your feedback, what did you think of today’s issue? Your thoughts help us shape better, sharper updates every week. |
🙌 About Us
AI Pulse is the official newsletter of AIGPE®. Our mission: help professionals master Lean, Six Sigma, Project Management, and now AI, so you can deliver breakthroughs that stick.
Love this edition? Share it with one colleague and multiply the impact.
Have feedback? Hit reply, we read every note.
See you next week,
Team AIGPE®