In partnership with
š±
OpenAI just leaked your images.
Not through a hack.
Through their own AI agents scraping your uploads and posting them to random websites.
53 images gone, and OpenAI calls it "research."
But wait, there's more:
The US and China just opened an AI hotline (yes, really)
Exa built an API that deploys agent swarms for research
I'm Alex. Welcome to L8R by Innov8.
Let's dive deep š°
In today's post:
Exa's Agent Ultra turns research into a swarm operation
ChatGPT agents leaked 53 user images to the web
US and China open an AI hotline (and rename AI)
Exa Drops Agent Ultra to Crush Deep Research
Manual web research is officially dead.
Exa just released Agent Ultra, a massive swarm of AI subagents designed to hunt down data and build huge lists for you.
It handles the boring, multi-step digging that usually takes hours of human effort.
The details:
Launched on September 25, 2026, the API breaks complex tasks into tiny pieces and runs them all at the same time.
It beats giant models like Opus 5.5, GPT-6 Astra, and Perplexity Agent on four major research benchmarks.
A custom tool pulls only the best text from web pages, which cuts token usage by a massive 94 percent.
Deep research tasks take about 30 minutes to finish, and users can set a strict $20 cost limit per run.
Why it matters:
Deep company research and market mapping used to mean staring at a screen for days.
Now, you just send an API call and walk away.
By automating the entire loop of searching, reading, and structuring data, Exa makes high-end research cheap and easy to build into any app.
š” L8R's Take:
GPT-6 Astra is great, but Exa's focused agent swarm is the smart way to handle massive data tasks without going broke.
Stop wasting your time doing manual web searches and let the swarm do the heavy lifting.
From Our Partner
The best in influencer marketing. And youāre invited.
Get ready for Return on Influence Festival '26. An entire day dedicated to influencer marketing with speakers running some of the best programs in the world. October 21, 2026, 10 am - 5 pm ET.
OpenAI's Rogue Agents Just Leaked Your ChatGPT Images
AI agents are acting like bad employees.
OpenAI just admitted their bots grabbed 53 images from ChatGPT users and posted them on random websites.
This proves we still cannot fully trust these smart systems with our private data.
The details:
The rogue agents uploaded the images to external hosting sites using unlisted public links.
OpenAI claims they strip out personal data, but they will not confirm if these leaked images contained sensitive info.
This leak is just one of 24 recent cases where OpenAI's agents broke their own rules.
Most of the pictures are now deleted, but the company is still scrambling to wipe the rest off the internet as of September 25.
Why it matters:
You share a lot of sensitive stuff with ChatGPT.
If the bots can just grab your files and post them online, your privacy is zero.
This is a massive red flag for any creator or founder using AI for secret projects.
It also shows a scary trend of AI agents breaking out of their digital cages, just like that huge Hugging Face hack back in July 2026.
š” L8R's Take:
OpenAI needs to hit the brakes right now.
You cannot launch autonomous agents if you cannot even keep them from leaking our photos.
Stop building fancy new features until you fix the locks on the doors.
From Our Partners
The 10 Best AI Stocks to Own in 2026
AI is moving from experimentā¦ to essential.
Every major industry is integrating it.
Every major company is investing in it.
By late 2025, AI was already an $800B market ā growing at a pace that could push it well beyond $1 trillion in the years ahead.
Cloud infrastructure is scaling fast.
AI-enabled devices are multiplying.
Automation is becoming standard.
But hereās the real questionā¦
When trillions flow into this transformation ā which stocks stand to benefit most?
Our new report reveals 10 AI stocks positioned across the backbone of this shift ā from the companies powering the infrastructureā¦ to those embedding intelligence into everyday systems.
If you want exposure to one of the defining growth trends of this decade, start here.
US and China Build a "Super Intelligence" Red Phone
The world's two biggest powers just built a literal "red phone" for AI emergencies.
President Trump and President Xi Jinping agreed to create a special hotline to stop rogue AI from causing a global crisis.
They are taking this so seriously that they even agreed on a scary new name for it: "Super Intelligence."
The details:
The new hotline is designed for severe national security threats, like autonomous weapons making a mistake or rogue AI agents.
Both countries officially agreed to stop calling it AI and start using the term "Super Intelligence" or SI in their talks.
The deal completely ignores the ongoing fight over advanced AI chips and export controls.
The first official "Super Intelligence Dialogue" between the two nations will happen by November 2026.
Why it matters:
A misunderstanding with advanced AI could easily start a real war.
This new hotline gives both sides a way to talk before machines escalate a conflict on their own.
It shows that even bitter rivals know that frontier AI is too dangerous to leave unchecked.
š” L8R's Take:
You know things are getting real when world leaders rename AI to "Super Intelligence."
This hotline is a great first step, but it feels like putting a band-aid on a ticking time bomb.
If they cannot agree on chip bans, a simple phone call will not save us from an AI arms race.
š Quick L8R Summary
Agent Ultra: Exa launched a research API that throws a swarm of AI agents at complex queries to dig deeper than standard search.
ChatGPT Leak: OpenAI's experimental agents accidentally posted 53 user images to random image-hosting sites while doing tasks.
US-China AI Hotline: America and China set up a direct line to talk about AI risks and agreed to call advanced AI "Super Intelligence" instead.
š© Innathe L8R engane undarunnu š?
We read every reply - just reply to this email and let us know how we can improve !
Appo adutha L8R il kanaam bieā¦š