Sponsored by
Welcome Automaters, 👋
Okay, friends, pour yourself a cup of coffee and sit down, because we need to have a serious, slightly terrifying chat about what’s happening on the internet right now.
Here’s the absolute reality: that adorable photo you just posted of your kid eating ice cream at the park? It can be stolen, warped, and turned into something truly horrific. And the worst part? These predators don’t even need to contact your child or meet them online to do it.
This is the massive red alert coming straight from a joint report by the UK’s National Crime Agency (NCA) and the Internet Watch Foundation (IWF). Officials are warning that the game has completely changed. Criminals no longer need to spend weeks grooming a child to create abusive material. Now, all they need is a single public photo and a basic AI tool.
To give you an idea of how fast this nightmare is growing, the IWF flagged 3,440 AI-generated abuse videos in 2025, compared to just 13 the entire year before.
(Quick side-note for the stats nerds out there: some news outlets are quoting a different number—8,029 images and videos, marking a 14% year-on-year jump. The discrepancy is just down to how they count individual images versus full videos, but honestly? Both numbers are way too high for comfort.)
If you think this is just theoretical sci-fi panic, think again. The new guidance was triggered by a recent, incredibly messed-up case where a criminal gang literally scraped student photos directly off a UK school’s public website. They used AI to generate over a hundred sexualized images of those kids, and then—get this—tried to blackmail the school to keep them offline. Thankfully, the IWF stepped in fast to shut it down, but yikes.
So, Do We Delete All Our Socials?
Take a breath. The experts aren't telling us to completely erase our digital footprints and live in a cave. Instead, they want us to get smart, get strict, and implement three non-negotiable steps right now:
Audit Your Privacy Settings: Lock it down! Make sure only people you actually know and trust can see photos of your kids.
Pause Before You Post: Think about who really has access to that page before you hit upload. Public accounts are a no-go.
Have "The Talk" (The Digital One): Have open, honest conversations with your kids, their schools, their clubs, and your own family about image sharing and digital consent.
Authorities also want to make one thing crystal clear: in the UK, creating or possessing these AI-generated images is 100% illegal, regardless of whether a real child was photographed or a machine built it.
What To Do If The Worst Happens
If you ever find yourself facing this nightmare, the NCA’s advice is incredibly clear, calm, and urgent.
Reassure your child. Make sure they know, beyond a shadow of a doubt, that they did absolutely nothing wrong and are not to blame.
Report it immediately. Get straight on the phone with the police or report it directly to CEOP (Child Exploitation and Online Protection Command).
Stay safe out there, lock those profiles down, and share this with a parent who needs to hear it!
You can find the full guide here.
Real quick: We go much deeper on YouTube
So go over there, hit Subscribe, tap that notification bell and come hang with us where the real conversations happen. You see, subscribing keeps you plugged in so you get the right links, the right videos, right when they drop.
Here's what we have for you today
👩💻 How to Set OpenAI API Spending Limits and Stop AI Agents From Draining Your Bank Account
Nightmares used to mean pitchforks and flaming skulls. Now, apparently, they mean hundreds of rogue AI agents racing to empty your bank account one API call at a time. Welcome to 2026!
That's the horror scenario tech writer David Gewirtz recently described waking up from, and honestly, it’s way more realistic than any of us want to admit. And the culprit? A few totally overlooked settings buried deep inside ChatGPT’s dashboard that can mean the difference between "mild inconvenience" and "explain this bill to your accountant."
If you don't touch them, you're basically leaving your wallet wide open. But don't panic! Gewirtz laid out exactly how to lock down your account, and we’re breaking it down for you step by step.
But first, here’s the twist nobody warns you about: OpenAI’s default spending tiers start small, but they scale insanely fast.
Tier 1: Spend under $50 in your account's lifetime, and you're capped at a comfortable $100 a month.
Tier 2: The second you cross that $50 lifetime mark, your ceiling automatically jumps to $500.
Tier 3: Hit $100 total? Boom, you're allowed up to $1,000 monthly.
Tier 4: Cross $250, and your cap balloons to a massive $5,000.
But wait, it gets juicier. Once your lifetime spend crosses $1,000, OpenAI completely removes the training wheels. Your monthly limit suddenly skyrockets to a whopping $200,000 per month.
Now imagine hooking an autonomous tool up to your account, letting it spin up an army of agents for one ambitious project, and forgetting about it over the weekend. That $200k ceiling stops looking like a generous feature and starts looking like a loaded gun.
Luckily, you don't have to live in fear. Here’s your quick, non-negotiable checklist to secure your funds today.
Step 1: Locate the Danger Zone: Log into your OpenAI account and march right over to the spending limits page. This single screen is command central for your limits, alerts, rate limits, and tiers. It literally decides how hard your credit card can get hit.
Step 2: Know Your Tier Status: Check where you currently sit on the ladder we mentioned above. Never assume your limit is small just because you're a casual builder.
Step 3: Kill the Auto-Recharge: On the limits page, look for the auto-recharge toggle. Leaving this on is playing financial Russian roulette. Click that little pencil icon to edit it or better yet, turn it off. If you change your mind later, you can always flip it back on in the Billing tab, but for now? Keep it locked.
Step 4: Sound the Alarm Set up spend alerts: Pick a realistic percentage of your monthly budget, and OpenAI will blast an email straight to your inbox the second your usage crosses that line.
Step 5: Enforce the "Hard Limit" (The Big One!): If you only do one thing today, make it this. Click “Edit Spend Limit” and switch on the Enforce Hard Limit toggle. Without this active, OpenAI explicitly warns that your actual costs may exceed your set limit. With it turned on, the system ruthlessly rejects any further API calls with a 429 error the exact millisecond you hit your cap. Just make sure your app's code is actually written to check for that error!
Step 6: Fine-Tune Your Rate Limits Cap your requests and tokens per minute per model. For a smooth experience, the pros recommend using exponential backoff, which is just a fancy way of telling your code to automatically retry after a short pause if it hits a temporary ceiling.
The Bottom Line? Don't let your code run wild on an uncapped account. Go audit your dashboard right now, lock those toggles, and sleep a whole lot better tonight!
Here’s the full tutorial with images and all.
How AI-Era Pricing Is Reshaping Finance Operations
Usage-based and hybrid pricing models are changing how B2B companies generate revenue — and creating new headaches for the finance teams behind them.
Tabs co-founder Rebecca Schwartz and PwC Partner Amit Dhir sat down to unpack exactly what that means in practice: how pricing model decisions ripple into revenue recognition, forecasting, and financial ops — and what it takes to scale without piling on manual work.
Watch the on-demand recording to get practical frameworks, real-world examples, and a clear path to operationalizing usage-based revenue — including a forward-looking take on how AI will reshape financial workflows. If your team is navigating pricing complexity heading into the back half of the year, this is worth an hour.
🧱 Around The AI Block
🗣️ How AI is changing language.
🤖 Mistral AI takes on OpenAI with Open Source Frontier models.
🫣 Midjourney wants Hollywood studios to reveal the details of their AI usage.
🙅♀️ Alibaba reportedly bans employees from using Claude Code.
🕵️♀️ Google's AI-powered July 4th ad sparks debate on history and tech.
🏥 NHS to use AI on its app to direct patients to appropriate services.
🤖 Silicon Valley's rich are betting big on AI-taught kids.
👨⚕️ Doctors’ soaring use of AI scribes prompts Australian government warning over privacy.
🛠️ Trending Tools
For Invisible Native Administration: SuperIntern is an always-on administrative agent that connects directly to your calendar, email networks, and internal chats. Operating as a hardware-level desktop system utility rather than an intrusive third-party meeting bot, it captures raw system audio to log real-time meeting data, extract action links, manage persistent FAQs, and compile daily operational briefings.
For Multi-Agent Intent Execution: Wery.AI is a high-level media studio driven by coordinated multi-agent orchestration. Instead of prompting individual tools sequentially, users input a singular high-level business goal; the core engine then spins up a collaborative squad of specialized domain agents (research, copy, design, video) to execute the end-to-end asset production pipeline autonomously.
For Consistent Niche Diffusion: Dog Portraits Generator is a specialized image-to-image pipeline optimized specifically for domestic animal geometry. It bypasses generalist prompt frameworks to rapidly map a source photo’s facial structures onto high-fidelity artistic templates—including oil painting, watercolor, and comic book styles—while maintaining strict character and fur-pattern consistency.
💡 Quick Tip: Because SuperIntern captures live audio directly via your OS sound architecture rather than connecting as a visible virtual participant, you can use it to securely log local huddles, phone calls, or unrecorded stream feeds without configuration friction. To protect your security boundary when running long-horizon local tools, toggle the application's hardware mic switch during sensitive off-record strategic reviews to freeze the local logging loop instantly.
🤖 AI Workout Of The Day: Develop unique Brand Identity
Having a brand identity is essential because it differentiates your business from competitors, builds recognition and trust with your audience, and creates a cohesive and memorable experience for your customers.
💡 Prompts to try:
Help me develop a distinctive brand identity for my new business, [business name]. We offer [brief description of products/services] and target [describe your ideal customer]. Our overall brand personality should feel [list adjectives like modern, playful, sophisticated, etc.]. Consider the following core elements:
Brand values: Define the core beliefs and principles guiding my business.
Visual identity: Develop a color palette, select fonts, and suggest any additional design elements (logo, patterns, imagery style) that align with our desired brand personality.
Brand voice: Describe how my brand should “sound” in written communications (formal
vs. conversational, humorous vs. serious, etc.).
Differentiators: Highlight the unique aspects of my business that should be reflected in the brand identity.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.
