Quantum Leaps and Billions: AI's Breakout Week

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. Mate, this week’s been wild—quantum leaps, robot brains, and even AI-powered batteries. Where do we even start?

Llew Jury

Yeah, Ollie, it’s been a cracker of a week. Let’s kick off with Rigetti Computing, hey? They’ve just hit 99.5% median 2-qubit gate fidelity on their 36-qubit quantum computer. Now, I know that sounds like a mouthful, but basically, it means their quantum system’s making half as many errors as before. That’s huge for AI, right?

Ollie Carter

Absolutely. I mean, quantum’s always been this sort of “someday” tech, but this is the kind of progress that actually moves the needle. Fewer errors means you can run bigger, gnarlier AI models—stuff that’d choke a regular supercomputer. It’s like, we’re not just talking theory anymore. This could be the backbone for the next gen of AI, especially as we keep pushing for more power-hungry models.

Llew Jury

Spot on. And speaking of brains, did you catch the Lp-Convolution research? It’s this biomimetic image processing model—basically, they’re copying how our brains process visuals. Instead of treating every pixel the same, it focuses on what’s important, like how you’d read a book versus watching footy. The wild bit? It’s outperforming AlexNet, and it’s robust even with blurry or dodgy images.

Ollie Carter

Yeah, and Dr. C. Justin Lee’s team behind it—they’re not just doing this in the lab. They’re showing real-world results, and it’s actually cheaper to run. I love when AI gets a bit more “human,” you know? Makes you wonder what else we can borrow from biology. Oh, and speaking of brains—Skild AI’s “universal robot brain” is a trip. Instead of building robots, they’re building the AI that could run any robot. Imagine if every robot, from your vacuum to a factory arm, could just download the same brain and get smarter overnight.

Llew Jury

That’s the dream, right? It’s like the Android OS for robots. And then, mate, NJIT’s work on AI-discovered battery materials—this is the kind of thing that gets me excited. They’re using AI to find new materials that could replace lithium-ion. It’s not just about making batteries last longer, it’s about sustainability, too. AI’s searching through chemical possibilities that’d take humans decades. It’s a bit sci-fi, but it’s happening now.

Ollie Carter

Totally. And, uh, speaking of robots, I’ve gotta admit—this all reminds me of robotics hackathons. I tried to make a “universal” robot brain, but, uh, let’s just say my bot spent more time spinning in circles than solving problems. Not quite Skild AI level, but hey, you learn more from the fails, right?

Llew Jury

Ha! We’ve all been there. If your robot didn’t catch fire, you’re already ahead of the curve. But seriously, these breakthroughs—quantum, biomimetic models, universal brains, new batteries—they’re all converging. It’s not just one field moving forward, it’s everything at once. That’s what makes this moment so electric.

Llew Jury

Alright, let’s shift gears to the model wars and some big leadership moves. China’s been making a lot of noise—Zhipu’s GLM-4.5 just dropped, and it’s aimed squarely at intelligent agent applications. It’s open-source, too, which is a big deal for the global dev community. We talked a few episodes back about China’s push into open-source AI, and this is them doubling down.

Ollie Carter

Yeah, and then there’s Alibaba’s new reasoning model—Qwen3. China’s not just catching up, they’re pushing boundaries. The only catch is, as The Economist pointed out, they’re still struggling to run these massive models at scale. Infrastructure’s the bottleneck, not the brains.

Llew Jury

Spot on. And Europe’s not sitting still either. ETH Zurich and EPFL are about to release a multilingual LLM trained on over 1,500 languages. It’s open, so anyone can build on it. Imagine the impact for translation, accessibility, all that. It’s a real step toward making AI global, not just English-first.

Ollie Carter

And then Cohere’s getting in on the action with Command A Vision—their first model that can handle both text and images. It’s like, everyone’s racing to be the next big multi-modal player. Reminds me of when we saw GPT-5’s reasoning jump last month—suddenly, everyone’s gotta keep up or get left behind.

Llew Jury

Exactly. But it’s not just the tech, it’s the people steering the ship. Greg Barbaccia’s appointment as the first US Chief AI Officer is massive. He’s already the Federal CIO, and now he’s leading the new Chief Artificial Intelligence Officer Council. It’s a sign the US is getting serious about AI coordination across government. We’ve seen this before in other sectors, but never with AI at this scale.

Ollie Carter

And in the private sector, Nextracker’s brought on a Chief AI and Robotics Officer. That’s a new role for them, and it shows how AI leadership’s spreading beyond just tech companies. Solar energy, of all places! It’s like, if you’re not hiring AI execs, are you even trying to innovate?

Llew Jury

And then there’s OpenAI, their headcount’s gone from 770 to over 3,500 in less than a year. That’s bonkers. I remember at my startups Alfresco and Reload, we tried to double the team in six months, and keeping up kept us very busy. Scaling that fast, you have to focus on your culture, your processes as everything gets stretched. I can’t imagine what it’s like at OpenAI right now.

Ollie Carter

Yeah, and it’s not just about hiring bodies, it’s about making sure everyone’s rowing in the same direction. Otherwise, you end up with a bunch of smart people building in silos. It’s a good problem to have, but still a problem. Makes you appreciate the challenge of leading in this space, not just building cool tech.

Ollie Carter

Alright, let’s talk money, policy, and a bit of AI fun. OpenAI’s latest round—8.3 billion dollars at a 300 billion valuation. That’s not just big, that’s “change the industry” big. And the round was five times oversubscribed! Some early investors even got squeezed out to make room for new ones. It’s wild how much capital is chasing AI right now.

Llew Jury

Yeah, and it’s not just OpenAI. Anaconda raised over 150 million, now valued at 1.5 billion. The National Science Foundation’s putting 100 million into new AI research institutes. And if you add up all the big tech spend—Meta alone doubled their capex to over 30 billion, and the total for AI infra this year is 155 billion. That’s a lot of GPUs Ollie.

Ollie Carter

It’s like the gold rush, but for compute. But with all this money flying around, the policy side’s heating up too. Trump’s administration just dropped their AI Action Plan—three pillars: accelerate innovation, build American AI infrastructure, and streamline regulation. They want to keep the US ahead, but also make it easier to get AI to market. Meanwhile, China’s countered with their own global AI governance proposal. It’s a real tug-of-war for who sets the rules.

Llew Jury

And here’s the kicker—AuditBoard’s research says only 25% of businesses have proper AI governance in place. That’s a bit scary, considering how fast everyone’s adopting this stuff. There’s a real gap between building AI and actually managing the risks. Even the SEC’s getting in on the act, setting up an AI Task Force to help their staff use AI tools. It’s not just about regulating others, it’s about modernizing themselves.

Ollie Carter

Yeah, and on the lighter side, AI’s getting more creative—and sometimes more ridiculous. There’s a new tutorial out showing how anyone can build a fun video game with AI, no coding experience needed. AI companion apps are going viral—people are making digital friends just by typing a description. And mobile AI art generators like Wombo Dream and YouCam AI Pro are everywhere. It’s like, if you’ve got a phone, you’re an artist now.

Llew Jury

And let’s not forget the AI fails. Cheezburger’s got a list of 25 that’ll make you laugh—AI still gets it wrong in the funniest ways. It’s a good reminder that, for all the hype, we’re still early days. Sometimes the robots just need a bit more training, or maybe a sense of humour.

Ollie Carter

Couldn’t agree more. Alright, that’s a wrap for this week’s episode. From quantum leaps to billion-dollar bets, and a few laughs along the way, AI’s moving faster than ever. Llew, always a pleasure mate.

Llew Jury

Cheers, Ollie. And before we go, a shout and thanks to the great team and our sponsor Advancer. Advancer are the AI experts providing real world AI and Automation strategy, AI products and customised AI solutions for your business. Check them out at Advancer.com.au! Catch you next time!

Ollie Carter

Thanks to everyone tuning in. We’ll be back next week with more on the funding boom, new models, and how these policy moves play out. Until then, keep your robots upright and your AI fails funny. See ya, everyone!