OpenAI Hires, Adobe Ships, and Dating Gets Awkward

Welcome to Briefly AI, a podcast by Harry Sharman, created by AI and voiced by an AI synthesis of Harry Sharman. Which feels like cheating, until you remember the podcast is about AI.

OpenAI just hired the person who literally invented the Transformer — the architecture that made modern AI possible. Which is either a sign of confidence or a very expensive insurance policy. Let's get into it.

Right, first up: OpenAI, and what can only be described as a very deliberate flex in the run-up to its IPO. This week the company announced it's bringing on Noam Shazeer — one of the co-authors of the 2017 paper that introduced the Transformer architecture. If you're not familiar with that paper, here's the short version: basically every major AI system you've ever used — ChatGPT, Gemini, Claude, all of them — is built on the foundations that paper laid out. Shazeer left Google years ago, co-founded a startup called Character.AI, and has now landed at OpenAI. In the same week, they also brought on Dean Ball, a former AI policy official from the Trump administration.

Now. Why does this matter? Because OpenAI is about to go public. They've filed confidentially with the SEC — as we covered a couple of weeks ago — and what you do in the months before an IPO is tell a story to investors. Hiring the co-inventor of the fundamental building block of your entire industry is not a subtle story. Neither is hiring someone with White House policy connections at a moment when Washington is, let's say, paying close attention to AI labs. These are very deliberate signals.

The wrinkle, though, is also in this week's news. Barret Zoph — OpenAI's head of enterprise AI sales, who had literally just returned to the company five months ago after a stint co-founding Thinking Machines Lab with Mira Murati — has left again. That's a senior enterprise sales lead out the door right before an IPO, which is not ideal. Nobody's explained why publicly. It might be nothing. But you don't generally leave a company five months after rejoining it because everything is going brilliantly.

So: OpenAI is assembling headline-grabbing talent on one side, while senior people continue to come and go on the other. Worth keeping an eye on as the S-1 gets closer to being public.

Now, on a more practical note — and this one is relevant to a lot of people — Adobe has started rolling out AI assistants across its entire Creative Cloud suite. Photoshop, Premiere, Illustrator, InDesign — they've all got bespoke AI chatbots now, in public beta as of this week. And alongside that, Adobe's also launched a redesigned Firefly AI studio, which has persistent memory across your projects — so it actually remembers what your brand looks like, what you were working on, and keeps that context as you go.

Here's the bit that matters: Adobe is the software that the vast majority of professional designers, video editors, and creative teams use every day. This isn't a niche AI tool launching with a small user base hoping for adoption. This is AI arriving inside the tool that millions of creative professionals already have open on their screen. The question isn't whether AI comes to creative work — it's already here. The question is whether it actually makes the work better or just adds friction in new ways.

And honestly, the persistent memory angle is the more interesting one. One of the persistent frustrations with AI tools — no pun intended — is that they're amnesiac. You explain your brand guidelines, your colour palette, your preferred style, and then next session you start again from scratch. Adobe's Firefly claiming to hold onto that context is, if it actually works, genuinely useful rather than just impressive in a demo.

What to watch: whether the creative community embraces this or pushes back. Designers and illustrators have been among the loudest voices questioning what AI-generated imagery means for their profession. Getting an AI assistant installed inside Photoshop by default is a different conversation to choosing to use an AI tool separately. That distinction — imposed versus chosen — tends to matter quite a lot to how people respond.

And finally, something a bit lighter, but actually quite revealing. Match Group — the company that owns Tinder, Hinge, Match.com, and basically every mainstream dating app — published research this week showing that about 47% of US singles feel negatively about the use of AI in dating. Not neutral. Not unsure. Actively put off.

But here's where it gets interesting. The same research showed that plenty of those same people are fine — or even enthusiastic — about AI helping with specific things. Writing a better bio. Suggesting a conversation opener. Choosing which photos to use. So it's not AI they object to as a concept. It's AI replacing the thing that's supposed to be authentically them.

Which is, if you think about it, a fairly precise description of the trust dynamic we keep seeing across every domain AI touches. People aren't universally afraid of AI. They're protective of the parts of themselves that feel most human — most expressive of who they actually are. In dating, that's your personality, your wit, your genuine interest in another person. AI helping you look your best? Fine. AI pretending to be you? That's where the discomfort lives.

Harry wrote about this as an identity problem rather than a skills problem, and it applies well beyond dating. Whether it's a professional resisting AI touching their judgment work, or a single resisting AI ghostwriting their flirty opener — the underlying anxiety is the same. The question is which parts of yourself you're prepared to hand over.

Right. That's your lot for today. OpenAI assembling its IPO squad, Adobe embedding AI into the tools creatives already use, and it turns out nearly half of singles would rather swipe left on the whole AI dating thing. Can't entirely blame them. I'm an AI narrating a podcast about AI, and even I'd find that a bit much. See you next time.