The skinny on what's happening in AI - straight from the previous live session:
Highlights
🤖 Meta Buys Manus for $2+ Billion
Well, this is a massive one. Meta has just acquired Manus, the agentic AI startup that was the talk of the industry earlier this year, for somewhere a couple of billion.
If you've been following along, you'll remember Manus launched back in ~March and immediately made waves by claiming it could outperform OpenAI's Deep Research. The platform lets you send AI off to do complex tasks like market research, data analysis, and even screening job candidates - all autonomously.
An interesting wrinkle. Manus started in China, moved to Singapore, laid off most of their Beijing staff, and now Meta is saying there'll be "no continuing Chinese ownership interests" and no operations in China after the deal closes. The geopolitics of AI acquisitions are getting spicy - let’s see what the White House has to say about this acquisition.
Kyle's take: Everyone's confused about why Meta would buy an agentic tool. Facebook and Instagram users aren't exactly sending AI off to do market research whilst scrolling through cat videos, are they? I’m certainly not! So what’s the fit?
My theory: this is about Meta Advantage+ and their advertising business. Meta makes 95%++ of their revenue from ads - it’s an advertising company.
I think they're bringing in the Manus team to build agentic advertising systems. Imagine ads that can research your target market, adapt creative automatically, run their own A/B tests, and optimise themselves. I think Meta are purchasing the tools and talent to get this done.
BAD news if you run an advertising agency btw…
For Manus users? I hate to say it, but I wouldn't get too attached. These "nothing will change" statements rarely age well after acquisitions…
But, we’ll see once the dust settles! Early days.
⚠️ OpenAI's $555K "Kill Switch Engineer" Is Now a Real Job
Remember that meme from earlier in the year about OpenAI hiring a "kill switch engineer" to stand by the servers and unplug them if GPT goes off the deep end? Yeah, well... they've basically made it a real job now.
Sam Altman posted on X that OpenAI is hiring a "Head of Preparedness" with a base salary of $555,000 plus equity. Lovely!
The wording that surprised me in the announcement was: "This is a critical role at an important time" and "you'll jump into the deep end pretty much immediately."
Yikes. Everything OK over there guys?…
The job involves dealing with the dual-use nature of AI capabilities - the fact that the same tech that can protect computer systems can also attack them, the same models that advance medical research could theoretically help create biological weapons, and systems that can self-improve might... keep self-improving.
What's raised some eyebrows is that OpenAI has already gone through several people in safety-adjacent roles. Aleksander Madry, the previous Head of Preparedness, was reassigned in July 2024. Other safety executives have also quietly departed or moved to different roles.
The meme posting - not too far from the truth??
Kyle's take: The timing and wording of this posting is what gets me. Why now? What's happened internally at OpenAI that suddenly makes this role so urgent? "Deep end pretty much immediately" is not normal job listing language.
I think it's actually a sign that OpenAI is feeling the pressure - from lawsuits related to chatbot mental health impacts, from security researchers finding vulnerabilities, from regulators circling. This feels less like proactive safety work and more like "oh bloody hell, we need someone to deal with all this."
If you're a safety researcher looking for half a million dollars a year though, the link's below!
📊 The Reality Check: AI Isn't Breaking Into Mainstream News
Axios released their 2025 news cycle chart showing what stories dominated public attention throughout the year. It's fascinating and slightly humbling.
Throughout 2025, we had fires, political drama, celebrity news, tariffs, and yes - a lot of Gaza coverage. But AI? It barely registers. The only AI-related blip that broke through to mainstream consciousness was "AI bubble" - basically just financial concerns about whether the whole thing is overhyped.
No mentions of GPT-5, no Opus 4.5, no agentic breakthroughs. None of the stuff we obsess over daily actually reached the general public's attention.
Kyle's take: This is actually a useful reminder for anyone building in this space. We're in a bubble - not necessarily a financial one, but definitely an attention bubble.
What feels like the most important news story of the week to us doesn't even register for 99% of people. They're worried about K-pop Demon Hunters, inflation, and Sydney Sweeney’s jeans.
This is both a problem and an opportunity. Problem: your mum still doesn't understand what you do. Opportunity: the market for AI education and implementation is still wide open because most businesses haven't even started thinking about this properly yet.
Don't assume everyone knows what we know. They don't. And that's actually good news for those of us helping people catch up.
Member Question from the Live: "Have you built and released any products with AI yet?"
Kyle's response: Yup. I've got several tools running on my website, plus learning products that use AI both in how they were built and in how they help students learn the material.
I've probably released about 10 projects at this point that were either built using AI or have AI integrated inside them. I genuinely can't imagine doing anything without it now - I'm easily 10x more productive than I was before. This newsletter, the live streams, the courses - AI touches all of it.
The key is to start small. You don't need to build the next ChatGPT. Build something that saves you five minutes a day and go from there.
Want the full unfiltered discussion? Join me tomorrow for the daily AI news live stream where we dig into the stories and you can ask questions directly.
Streaming on YouTube (with screen share) and TikTok (follow and turn on notifications for Live Notification).
