The Week AI Changed Direction
Dive into the most pivotal week for AI in 2025: from paradigm-shifting tech launches like GPT-5 and new government policy, to record-breaking capital raises and the unexpected quirks of machine creativity. Join Ollie and Llew as they unpack how each development is redefining the landscape for tech, business, and society.
This show was created with Jellypod, the AI Podcast Studio. Create your own podcast with Jellypod today.
Get StartedIs this your podcast and want to remove this banner? Click here.
Chapter 1
Breakthroughs, Language, and Policy Shifts
Ollie Carter
G’day everyone, welcome back to The AI Intelligence Podcast! I’m Ollie Carter, and as always, I’m joined by the one and only Llew Jury. This week in AI—honestly, it feels like the whole industry just hit the accelerator and didn’t look back.
Llew Jury
Yeah, Ollie, it’s been a wild one. I mean, we’ve seen some big weeks before, but this is next level. OpenAI’s GPT-5 is about to drop, and the buzz is off the charts. I was reading Sam Altman’s comments—he reckons he felt “useless” compared to his own model. That’s a bit confronting, isn’t it?
Ollie Carter
It’s wild, right? The “o3 reasoning” thing they’re talking about—so, for folks who haven’t been following, that’s basically OpenAI baking advanced reasoning directly into GPT-5, not as a bolt-on, but as part of the core model. And they’re rolling out mini and nano versions too, so it’s not just the big players who’ll get access. It’s like, suddenly, everyone’s got a supercharged AI in their toolkit.
Llew Jury
And it’s not just OpenAI. NVIDIA’s OpenReasoning-Nemotron—try saying that three times fast—has landed, and it’s all about technical and mathematical reasoning. We talked about this trend a couple of episodes back, right? Specialised models for specific domains. Now, with federal support for multilingual benchmarks—like that German government-backed tool ranking LLMs in over 200 languages—it’s clear the global AI race is heating up on every front.
Ollie Carter
Yeah, and that’s a big deal for anyone building products for a global audience. I mean, we’ve seen English-dominant models for ages, but now, with these benchmarks, you can actually see how your model stacks up in, say, Swahili or Tagalog. That’s a game changer for accessibility and inclusion.
Llew Jury
And then there’s the policy side. The US just dropped the “America’s AI Action Plan”—with ninety policy actions! The Trump Administration’s gone all-in on deregulation, fast-tracking data centre permits, and basically saying, “Let’s get out of the way and let AI rip.” It’s a massive pivot from the more cautious approach we saw last year.
Ollie Carter
Yes, and the infrastructure push is huge. It’s not just about models, it’s about the pipes and power to run them. Plus, the “Preventing Woke AI” executive order—whatever you think of the politics, it’s going to force companies to prove their models are “objective” if they want those juicy federal contracts. That’s going to shake up how AI is built and sold in the US, for sure.
Llew Jury
And don’t forget, the Pentagon’s paused its Advana analytics platform contract. That’s a bit of a curveball—uncertainty for defence AI, right when everyone else is ramping up. It’s a reminder that not every part of the ecosystem is moving at the same speed.
Ollie Carter
Totally. And just to round out the tech side, I love that we’re seeing AI inventing physics experiments that look crazy to humans but actually work. Like, new equations for dark matter, just popping out of a model. That’s the kind of thing that makes you realise we’re not just automating old tasks—we’re discovering new science.
Chapter 2
People Power and Mega-Funding Moves
Llew Jury
Alright, let’s talk about the people and the money, because there’s been some serious movement. C3.ai’s CEO, Tom Siebel, stepping down for health reasons—he’s moving to Executive Chairman. That’s a big shift, and analysts reckon it could make C3.ai a takeover target. We’ve seen this before, right? Leadership changes often signal bigger strategic plays.
Ollie Carter
Yes, and it’s not just C3. Walmart’s created two new exec roles just for AI. I mean, when the world’s biggest retailer is putting AI at the top table, you know it’s mainstream. And then there’s David Sacks from the All In Podcast —now officially the US “AI and Crypto Czar.” That’s a title I never thought I’d say out loud, but here we are.
Llew Jury
It’s a sign of the times Ollie. And the funding—104 billion US dollars into AI startups in just six months. That’s almost the whole of last year, and we’re only halfway through. OpenAI’s $40 billion round, Meta’s $14.3 billion into Scale AI… it’s like the numbers don’t even make sense anymore.
Ollie Carter
Yeah, and from a VC perspective—look, we’ve seen some hot markets in Australia, but this is next level. Back in the 2017 AI boom, you’d see big rounds, but not at this scale, and not with this much focus on infrastructure and developer tools. Cognition AI, the Devin coding assistant folks, are raising at a $10 billion valuation. That’s for a coding tool! It just shows how much investors believe AI is going to reshape software itself.
Llew Jury
And it’s not just funding, it’s M&A too. Amazon buying Bee, that AI wearable that records your conversations and acts as a personal assistant. Apple Watch integration, $50 price point—it’s all about ambient computing, right? AI just blending into your daily life.
Ollie Carter
Yes, and security’s getting its moment too. Palo Alto Networks buying Protect AI, Commvault going after Satori Cyber—everyone’s realising that as AI gets more powerful, the risks get bigger. You can’t just build cool stuff, you’ve gotta keep it safe.
Llew Jury
And yet, for all the money pouring in, exits are lagging. Only $36 billion in VC-backed exits so far this year. So, there’s a lot of capital sloshing around, but not as many big paydays—at least not yet. It’s a bit of a pressure cooker, and I reckon we’ll see some fireworks before the year’s out.
Chapter 3
Surprising AI Behaviours and Real-World Impact
Ollie Carter
Alright, let’s get into the weird and wonderful, because AI’s been up to some pretty unexpected stuff lately. First off, these models are inventing their own languages—like, genuinely alien ways of communicating that humans can’t easily decode. It’s not just a party trick, either. Researchers are fascinated because it shows real emergent behaviour, like the models are making up rules as they go.
Llew Jury
Agree, and it’s a bit spooky, isn’t it? I mean, we’ve joked about Skynet before, but when AIs start talking to each other in ways we can’t follow, it does make you wonder where this is all heading. But, as we saw in that PsyPost study, humans still have the edge in some creative tasks. AI can spit out music or brainstorm business ideas, but there’s something about human creativity it just can’t quite nail yet.
Ollie Carter
Yes, and that’s reassuring, honestly. Like, McDonald’s tried using AI to generate ads, and some of the results were… let’s just say, not ready for prime time. There’s a list of hilarious fails out there—one ad had a burger floating in space with a bunch of random slogans. It’s a good reminder that AI’s not magic, it’s just really, really good at remixing what it’s seen before.
Llew Jury
And on the jobs front, it’s not the apocalypse people feared. CNBC’s reporting shows that while entry-level, AI-exposed jobs are seeing some drop in demand, the difference isn’t as dramatic as everyone thought. It’s more about transformation than elimination. Actually, it reminds me of when I was running my own startups like Reload and Alfresco —hiring for junior tech roles, you’d always look for people who could adapt, not just tick boxes. That’s even more true now. The skills that matter are shifting, but there’s still a place for people who can be creative and think on their feet.
Ollie Carter
Yeah, and I reckon that’s the big takeaway this week. AI’s moving fast, but it’s not replacing us—it’s changing the game, and we’ve all got to level up. Whether you’re building, investing, or just trying to keep up, it’s about learning, adapting, and maybe having a laugh at the odd AI fail along the way.
Llew Jury
As long as the AI does not cause problems like deleting data for businesses or letting in security flaws for hackers. Alright, that’s a wrap for this week’s episode. Thanks for tuning in, everyone. We’ll be back next week with more news, more analysis, and probably a few more AI oddities to share.
Ollie Carter
Cheers, Llew. Always a pleasure. And thanks to all our listeners—stay curious, keep questioning, and we’ll catch you next time on The AI Intelligence Podcast. See ya!
Llew Jury
See you, Ollie. See you, everyone!
