In partnership with
Welcome Automaters, 👋
So something absolutely wild is happening in Silicon Valley right now, and it involves a massive dose of corporate delusion.
Tech CEOs are playing around with AI tools, getting completely starstruck by basic features, and then making some very big, very permanent, and honestly terrifying decisions; like laying off thousands of actual human beings.
Box founder and CEO Aaron Levie actually gave this exact phenomenon a name on X, and it’s the best thing you will read all week. He wrote that "CEOs are uniquely prone to AI psychosis."
Yes, he really said it. Out loud. For the entire internet to see.
Let's break down how this psychological meltdown actually works. A CEO fires up a flashy new AI model, builds a quick three-line demo, generates a basic legal contract in five seconds, and suddenly their brain shorts out. They immediately think: Wait... can robots just do everything now? Do I even need to pay humans anymore? Um, no, sir. No, they cannot.
Why? Because that overpaid CEO is not the person who actually has to catch the critical bugs, fix the hallucinated code libraries, or spend three grueling days untangling a sneaky, flawed contract clause. That’s a regular worker's job, and AI still messes those things up a lot.
The scary part is that this tech-bro daydreaming has catastrophic, real-world consequences. We are only a few months into 2026, and over 115,000 tech workers have already been laid off across 152 companies, according to tracking site Layoffs.fyi.
To put that into perspective: that is nearly as many job cuts as all of 2025 combined. The kicker? Almost every single one of these companies is aggressively pointing at "AI efficiency" as the main reason for the bloodbath.
Take ClickUp's CEO, Zeb Evans, for example. He practically celebrated cutting 22% of his total staff after deploying 3,000 AI agents, bragging that it was the start of a future "100x organization." He even claimed the mass firing was absolutely not about saving money. Sure, buddy. Whatever helps you sleep at night.
But here’s where the gossip gets delicious: the actual scientific research shows that the corporate hype is light-years ahead of reality.
A recent UC Berkeley study looked at the actual macroeconomic data and found zero solid link between AI adoption and actual productivity gains at scale. In other words, the bots are not actually making companies hyper-efficient yet.
MIT researchers chimed in with a reality check, estimating that AI agents will only be able to handle most tasks at a basic, entry-level capability by 2029. But truly outperforming human professionals? That’s a milestone predicted for a few years after that.
Even Sam Altman admits he got it wrong, the OpenAI CEO recently admitted he expected AI to wipe out way more entry-level white-collar jobs by now than it actually has. In his own words he’s, "delighted to be wrong about this,” noting that the predicted wave of AI productivity at the junior level simply has not hit the way he anticipated.
The Takeaway: CEOs are firing real people today based on software capabilities that won't actually exist for another three to five years. It’s pure, unadulterated panic-management.
Aaron Levie's actual advice to his fellow tech executives is surprisingly grounded: use AI constantly, hit the technical limits yourself, and walk away with a realistic picture of both the impressive good and the deeply messy bad.
P.S. We go even deeper on YouTube.
The newsletter is your fast-lane briefing, but the full picture, the context, the nuance, and the fun breakdowns live over on The Automated TV on YouTube.
Every day we drop new episodes packed with real depth, sharp commentary, and actual details that go way beyond the headlines. The kind of takes that broaden your horizon and make the news something you can genuinely learn from, build with, and grow on.
So go over there, hit Subscribe, hit the notification bell and come hang with us where the real conversation happens!
Here’s yesterday’s breakdown:
Here's what we have for you today
🤯 AI Inference and Agentic AI Generate 450% More Web Traffic
Cisco just dropped its first-ever, highly anticipated study on how AI is hitting Wide Area Networks (WANs). If you aren't a networking nerd, WANs are basically the giant, invisible highway systems that move data across the entire globe. And the ultimate verdict? AI is not just using the internet; it is aggressively rebuilding it from the inside out.
Turns out, AI is officially the world's hottest new traffic problem, and we’re only seeing the very beginning of the chaos.
Let’s talk numbers, because they are genuinely wild. Cisco projects that as consumer AI adoption reaches near-universal usage over the coming years, AI and autonomous agents will skyrocket consumer-driven network traffic by a staggering 6.6 times by the mid-2030s.
That extra load represents roughly 63% more growth compared to a world where AI didn't exist.
To put that into perspective: if the internet were a standard road, AI is about to add six extra lanes of bumper-to-bumper cars overnight. Or, if you prefer a different vibe: picture the internet as a standard backyard water pipe; AI is about to blast a fire hose through it.
But wait, it gets so much worse. If you think regular AI tools chatting with you on a screen take up a lot of space, think again. That’s small potatoes.
The real internet apocalypse is coming from AI agents, the little digital workers that actually go off and execute tasks for you automatically, like booking entire vacation itineraries or filling out complex corporate forms.
Cisco tested an AI agent and found that agent-driven tasks generated a jaw-dropping 450% more total internet traffic per task compared to standard web browsing done manually by a human. These little assistants are absolutely ravenous for bandwidth.
So where is all that extra traffic coming from? A whopping 70% of it is just the agent constantly pinging back and forth with the core AI model powering it. Cisco literally calls this constant connection the agent's spinal cord. Cut that digital cord, and your helpful little bot instantly becomes a very expensive brick.
🗣️ Why AI "Talks Back" Way Too Much
The way AI moves data is also fundamentally changing the internet's original architecture. AI inference flows (the process of an AI running your request through its neural network) run roughly twice as long as typical web transactions.
Even crazier? They demand significantly more upstream capacity. And the Receipt: About 9% of AI inference flows send way more data upstream than downstream. For regular web traffic, that number is a microscopic 0.5%.
In human terms? AI talks back to the internet a lot more than your average browser tab. It isn't just quietly downloading a page; it is constantly shouting data back up to the cloud.
Right now, AI inference traffic looks pretty tiny when you compare it to the massive amounts of data used by people streaming Netflix or scrolling TikTok. But Cisco projects it will represent a massive 25% of all network traffic by 2035. That’s a literal quarter of everything existing on the internet.
Because of this, network latency (that annoying lag time) is becoming an absolute emergency. Major cable operators like Comcast and Charter are already locked in a fierce race to upgrade their physical infrastructure to support the low-latency, high-upstream demands that AI requires.
Cisco openly admits this will force internet operators to roll out entirely new, flow-aware network and security systems just to prevent the web from collapsing under the weight of our new bot sidekicks.
So what do you think? Are you ready for an internet where a quarter of all traffic is just AI bots talking to other AI bots, or does that sound like a total nightmare?
We’ll be diving deeper into this on our Channel on YouTube, so stay tunes!
Claude is not just a chatbot anymore. Is your security team ready?
Claude.ai is one thing. Claude Cowork with MCP connections, running agentic workflows, taking actions across your data with ungoverned skills? That is a different conversation entirely, and most security teams are not equipped to govern it.
Harmonic Security is built to secure everything Claude offers. Full browser controls for Claude.ai, deep governance over agentic MCP workflows, and real-time visibility into what Claude is doing across your organization. So your CISO can say yes to the tools your business is already demanding.
🧱 Around The AI Block
💪 Illinois lawmakers just passed America’s strongest AI safety bill.
🧐 Why Google’s AI can’t spell Google (or anything else).
😎 ElevenLabs’s new music generation model can switch genres mid-track.
📛 YouTube will now automatically label AI videos.
🤔 Robinhood now lets your AI agents trade stocks.
📍 Erin Brockovich launches a crowdsourced AI data center map
🛠️ Trending Tools
For Terminal-Native Devs: Grok Build is an open-source command-line interface (CLI) tool that links Grok models directly into your terminal environment. It allows engineers to write code, debug scripts, and execute background AI automation routines without leaving their local setup or IDE.
For Always-On Automation: Gemini Spark is Google's cloud-hosted, proactive agent ecosystem. Operating independently even when your physical devices are powered down, it continuously crawls your Gmail, Drive, and Calendar to autonomously sort schedules, reply to operational tasks, and manage data pipelines.
For Visual Demo Conversions: Vibrantsnap is an AI-powered screen recording and mock-up engine. It automatically transforms raw screen captures into stylized product demonstrations by adding dynamic canvas backgrounds, pan-and-zoom animations, and optimized social layouts built to lift conversion rates.
For Voice Replication: Sydium is a digital twin engine for social media content. By scanning your historic publication data, it builds a localized profile mapping your specific industry tone, syntax habits, and pacing to autonomously write, schedule, and distribute platform-native text assets.
🤖 AI Workout Of The Day: Create Stunning Product Images & Business Logos With AI
High-quality visuals are non-negotiable in today’s crowded digital space. Whether it’s for marketing, advertising, or social media, a sharp, professional image instantly communicates credibility and attracts attention.
But not every business can afford a full photography studio. That’s where this AI-powered prompt comes in: it lets you generate professional-grade, photo-realistic images and logos without the guesswork.
Here’s How to Use This Prompt Effectively
Be Specific About the Product/Brand: Include exact product type, materials, colors, or logo style. The more detail, the more realistic the output.
Choose Your Style: Decide if you want a clean studio look, lifestyle setting, or dynamic elements like props, mist, or reflections.
Define Your Purpose: Are the images for social media, a website banner, or print? This will guide aspect ratios and composition.
Experiment with Variations: Generate multiple versions with small tweaks to lighting, angles, or props to select the most impactful image.
Use Post-Processing Sparingly: Minor touch-ups in tools like Photoshop or Canva can perfect the final output.
💡 Prompts to try:
You are a professional product photographer and digital image creator. Your task is to generate a high-quality, photo-realistic image of {insert product/business descriptions or examples} for marketing, advertising, or social media purposes.
The image should:
-Focus on the product or logo, ensuring sharp clarity and strong branding presence.
-Use a clean studio background with soft shadows, or a lifestyle setting that complements the product.
-Apply realistic, professional lighting with highlights and subtle reflections.
-Optionally, include dynamic elements such as condensation, splashes, mist, props, or contextual items that enhance appeal.
-Use composition techniques such as the rule of thirds or centered framing, with a shallow depth of field to emphasize the subject.
-Output in multiple aspect ratios: 4:5 (social), 16:9 (banner), 1:1 (thumbnail), suitable for both print and digital use.
-Include variations with different lighting angles, props, or backgrounds to allow selection of the most impactful image.
Ensure the final output is professional, visually appealing, and consistent with brand identity.
Optional Enhancement:
-Add a subtle texture, gradient, or reflective surface for extra depth.
-Suggest angle preferences (front view, angled, overhead, close-up).
-For logos, include different mockups (e.g., on stationery, signage, packaging) to visualize real-world application.Is this your AI Workout of the Week (WoW)? Cast your vote!
That's all we've got for you today.
Did you like today's content? We'd love to hear from you! Please share your thoughts on our content below👇
What'd you think of today's email?
Your feedback means a lot to us and helps improve the quality of our newsletter.
