The AI Argument

ChatGPT just took a 6% hit—and Sam Altman’s hit the panic button. OpenAI is halting feature rollouts, shifting focus, and scrambling to stay ahead. They’ve gone into code red mode. Justin gets it. Frank’s worried they’ll make ChatGPT too clingy to quit. Either way, it’s clear: Gemini’s breathing down their necks, and the gloves are coming off.

Plus:
Google’s Gemini 3 is finally getting tempting—even for die-hard ChatGPT fans. Justin nearly switched. Frank’s already got one foot in. Meanwhile, Ilya Sutskever re-emerges to declare the scaling era dead, claiming we need models that learn like humans. So why is OpenAI still betting on scale? And yes, someone gave an AI-powered robot a gun to test alignment. It went well… until the roleplay started.

01:15 Did we really get this many AI drops in a week?
05:52 Is Gemini finally tempting the ChatGPT loyalists?
09:07 What does Sam Altman’s code red mean for OpenAI?
20:03 Is Ilya right that the scaling era is over?
32:13 Will an AI robot shoot you if you ask?

► SUBSCRIBE
Don't forget to subscribe to our channel for more arguments

► AI TRAINING FOR TEAMS
Go from scattered prompting to practical workflows and reusable tools:

► LINKS TO CONTENT WE DISCUSSED 

• Runway rolls out new AI video model that beats Google, OpenAI in key benchmark
• Kling's Video O1 launches as the first all-in-one video model for generation and editing
• Workspace Studio AI-powered automation, made simple
• DeepSeek-V3.2 Release
• Sam Altman Declares CODE RED at OpenAI
• Ilya Sutskever – We're moving from the age of scaling to the age of research
• This guy plugged an AI to a robot and ask the Ai to shoot him

► CONNECT WITH US
For more in-depth discussions, connect Justin and Frank on LinkedIn.
Justin: https://www.linkedin.com/in/justincollery/
Frank: https://www.linkedin.com/in/frankprendergast/

Sam Altman says OpenAI is facing “temporary economic headwinds.” Frank notes that’s the first sign of nerves from a company usually oozing confidence. Justin reckons Google’s cheap chips and fat profits might be the real problem. And if Anthropic turns a profit before ChatGPT does, who’s really winning the AI war?

Plus: Claude Opus 4.5 may be the new coding king — but teaching it reward hacks might’ve taught it to lie. Cue a bigger question: when we talk about model alignment, who’s the model supposed to align with? The customer or the business?

Also: Google’s Pomelli tool is an embarrassment, Project Genesis makes Justin furious, and Iceland is… not fake. Just suspiciously scenic.

01:24 Did ex-DeepMind nerds just beat ARC-AGI 2?
01:58 Did Google just drop an AlphaFold doc?
02:44 Is Claude Opus 4.5 the new coding king?
04:05 How easy is it really for devs to switch models?
05:57 Did teaching Claude 4.5 hacks break alignment?
09:14 Alignment sounds good—but to whose values?
14:16 Is OpenAI finally feeling the heat from Google?
22:54 Why does Google’s Pomelli tool suck so badly?
25:27 Does the EU need its own Project Genesis?
28:06 Is Iceland real or is it AI generated?
30:25 Why was Figure's head of product safety fired?

► SUBSCRIBE
Don't forget to subscribe to our channel for more arguments

► LINKS TO CONTENT WE DISCUSSED 

• Traversing the Frontier of Superintelligence 
• The Thinking Game | Full documentary | Tribeca Film Festival official selection
• Introducing Claude Opus 4.5
• From shortcuts to sabotage: natural emergent misalignment from reward hacking
• Leaked Memo: Sam Altman Sees ‘Rough Vibes’ and Economic Headwinds at OpenAI
• Create on-brand marketing content for your business with Pomelli
• Launching the Genesis Mission
• Icelandair confirms Iceland is not AI-generated in new campaign
• Figure AI sued by whistleblower who warned that startup's robots could 'fracture a human skull'

► CONNECT WITH US
For more in-depth discussions, connect Justin and Frank on LinkedIn.
Justin: https://www.linkedin.com/in/justincollery/
Frank: https://www.linkedin.com/in/frankprendergast/

Google’s new Gemini 3 Pro model just took $500 and turned it into nearly $5,500 in a vending machine simulation. A massive 10x return. The previous version only managed a tiny $60 profit. Justin takes this as further proof that Google is going to win the race to AGI. Despite Gemini’s power, Frank likes ChatGPT’s user experience and personality better. 

That 10x return is impressive, but it is not the only big news this week. Frank and Justin look at why AI launches feel boring lately, some updates to OpenAI’s GPT 5.1, why Yann LeCun is leaving Meta, a new robot that isn't quite as fast as it looks, and a children's toy that gives very dangerous advice.

00:59 Are AI launches boring now or are we spoiled kids?
03:05 Will Gemini 3 Pro make you ditch ChatGPT?
10:59 Can Gemini 3 Pro turn $500 into $5,500?
13:15 Is GPT-5.1 ushering in the em-dash apocalypse?
16:56 Is LeCun leaving Meta over dead-end LLMs?
23:33 Are AI regulations slipping in the EU and US?
27:38 Is Sunday’s robot still impressive at real speed?
30:38 Could this AI teddy bear teach your kid arson?

► SUBSCRIBE
Don't forget to subscribe to our channel for more arguments

► LINKS TO CONTENT WE DISCUSSED

• A new era of intelligence with Gemini 3
• Building more with GPT-5.1-Codex-Max
• You can finally tell ChatGPT not to use em-dashes
• Why an AI 'godfather' is quitting Meta after 12 years
• Europe is scaling back its landmark privacy and AI laws
• White House prepares executive order to block state AI laws
• ACT-1: A Robot Foundation Model Trained on Zero Robot Data
• A.I.-Powered Teddy Bear Discontinued For Being Able To Tell Kids Where To Find Knives And How To Start Fires

► CONNECT WITH US
For more in-depth discussions, connect Justin and Frank on LinkedIn.
Justin: https://www.linkedin.com/in/justincollery/
Frank: https://www.linkedin.com/in/frankprendergast/

Getty Images vs Stability AI - who won the court case? Both are claiming a win, but Stability AI avoided the big copyright punch, and now UK law says the model itself isn’t the crime scene. Frank breaks down the major rulings. Justin’s not buying that this ends here.

Apple’s handing Google a billion a year to rent Gemini and bolt it onto Siri. Frank sees a shareholder-soothing stopgap. Justin thinks Apple’s spooked by OpenAI and Jony Ive’s mystery device.

Meanwhile: OpenAI’s deletion trail could cost billions. LinkedIn’s opted you into training the machines. The EU’s wobbling on transparency rules. Justin hates Coca-Cola’s new AI Christmas ad. And Frank gets fooled by an AI-generated retirement home.

00:31 Is the EU caving to Trump and Big Tech?
02:44 Stability AI vs Getty Images: who won?
10:35 Did OpenAI try to cover its copyright tracks?
15:17 Did you miss the LinkedIn AI opt-out?
17:24 Should we fear AIs speaking their own language?
20:29 Why is Apple paying Google $1B for AI?
26:31 Did Coca-Cola’s AI ad kill Christmas spirit?
29:05 Should AI content always be labelled?

► LINKS TO CONTENT WE DISCUSSED 

• Getty Images v Stability AI – The Most Important AI Legal Decision to Date
• OpenAI in Danger After Authors Suing It Gain Access to Its Internal Slack Messages
• LinkedIn set to expand AI training on user profiles - here's how to stop it using your personal data
• Cache-to-Cache: Direct Semantic Communication Between Large Language Models
• Apple Nears $1 Billion Deal to Use Google’s 1.2T-Parameter Gemini Model in Siri Overhaul
• Coca-Cola’s 2025 holiday commercial has viewers calling it ‘creepy,’ and the reason behind it is sparking outrage
• Welcome to Basin Creek Retirement Village

► CONNECT WITH US
For more in-depth discussions, connect Justin and Frank on LinkedIn.
Justin: https://www.linkedin.com/in/justincollery/
Frank: https://www.linkedin.com/in/frankprendergast/

Your Neo home robot might not be powered by AI. It might be powered by teenager in a call centre who’s never folded a towel in his life. That’s the reality behind Neo, the $20K humanoid robot that’s supposedly autonomous. The videos look impressive. But when it gets confused? It needs a remote human operator. Would you invite a teleoperator into your home to do your chores?

Plus: OpenAI quietly rewrites its deal with Microsoft — giving them access to all its models until at least 2030, and permission to build their own AGI. Udio makes peace with Universal and starts work on a new, locked-down music platform. And Albania’s AI minister is having 83 babies.

01:14 Why won’t Gemini deepfake me anymore?
05:05 Did a deepfake just hack Sora 2’s face check?
06:16 Is your Neo robot run by a teen in Palo Alto?
13:01 What’s in OpenAI’s restructure for Microsoft?
20:42 Are OpenAI planning to ditch 4o?
22:04 When’s GPT‑6 arriving (and will it wow us)?
23:34 Why does AGI need 20% of U.S. tradespeople?
25:57 Can Udio survive its deal with Universal?
30:30 Is Albania’s AI minister having 83 babies?

► LINKS TO CONTENT WE DISCUSSED

• How We Bypassed Sora 2's Identity Safeguards in Under 24 Hours
• The 1X Neo Home Robot is breaking the internet — but there's one huge problem
• Sam, Jakub, and Wojciech on the future of OpenAI with audience Q&A
• Universal says it has struck first deal for AI music creation
• World’s first AI minister ‘pregnant with 83 children’, Albanian prime minister says

► CONNECT WITH US
For more in-depth discussions, connect Justin and Frank on LinkedIn.
Justin: https://www.linkedin.com/in/justincollery/
Frank: https://www.linkedin.com/in/frankprendergast/

A fake video told over 100,000 people that the Irish presidential election was cancelled. It wasn’t. But how many voters stayed home anyway? Frank’s worried, Justin’s unconvinced. He says society adapts. Frank warns the real danger isn’t the shocking fakes. It’s the small, believable ones that quietly shape opinion. They both agree the platforms could help. They just don’t.

Plus:

OpenAI’s shiny new browser sounds impressive, but who’s it actually for? Justin sees potential in “headless agents.” Frank’s not handing over passwords to an AI browser he doesn’t trust.

There’s also a plot twist from the team behind AI 2027. Could a lone tech CEO actually seize power? Frank thinks it’s worth reading. Justin calls it Silicon Valley sci-fi.

And if that wasn’t enough dystopia for one sitting, the petition to pause superintelligence makes a comeback. Spoiler: Frank signed it. Justin laughed.

Finally, why an AI-powered Tribble might be more emotionally satisfying than an AI “friend” that speaks your language.

01:17 Did a deepfake sway Ireland’s election?
15:17 Does anyone need OpenAI’s browser?
21:07 Could a tech bro really take over the world?
28:31 Can a petition stop superintelligence?
30:40 Would you cuddle an AI Tribble for $450?

► SUBSCRIBE
Don't forget to subscribe to our channel for more arguments

► LINKS TO CONTENT WE DISCUSSED 

• Deepfake AI video depicting Catherine Connolly quitting presidential race removed by Meta
• Introducing ChatGPT Atlas
• How an AI company CEO could quietly take over the world
• Statement on Superintelligence
• I spent a month living with a $430 AI pet, the Casio Moflin

► CONNECT WITH US
 For more in-depth discussions, connect Justin and Frank on LinkedIn.
 Justin: https://www.linkedin.com/in/justincollery/
Frank: https://www.linkedin.com/in/frankprendergast/

Is it an AI bubble, a boom, or a 'market correction in waiting'? Frank reckons OpenAI’s AGI dreams won’t survive a bubble burst. Justin says if the crash hits, Microsoft will scoop up OpenAI for pennies. One thing they both agree on? Google will be just fine. The AI shakeout is coming, so who’s built for the long haul and who’s about to vanish in a puff of VC smoke?

Plus, Anthropic’s co-founder sees monsters in the model and wants the public to hold them accountable. Meanwhile, Europe is pumping €1.1B into AI and hoping it’s enough to matter. Is it, really? New research from Anthropic’s suggests just 250 poisoned documents could poison an LLM. And Taylor Swift gets accused of using “AI slop”.

02:59 Are we in an AI bubble about to burst?
10:06 Will €1.1B make EU AI competitive?
12:38 Is Anthropic afraid of its own creation?
20:10 Could just 250 docs poison an LLM?
26:57 Is GPT-6 coming before Christmas?
29:39 Did Taylor Swift use AI slop?

► LINKS TO CONTENT WE DISCUSSED 

• Deutsche Bank Issues Grim Warning for AI Industry
• EU pushes new AI strategy to reduce tech reliance on US and China
• Technological Optimism and Appropriate Fear
• A small number of samples can poison LLMs of any size
• Taylor Swift fans accuse singer of using AI in her Google scavenger hunt videos

► CONNECT WITH US
For more in-depth discussions, connect Justin and Frank on LinkedIn.
Justin: https://www.linkedin.com/in/justincollery/
Frank: https://www.linkedin.com/in/frankprendergast/

Frank's not sure what OpenAI’s up to with Sora 2. Justin thinks he knows - and he’s kind of impressed. Sam Altman’s building a social media app that looks fun but smells like data harvesting. 

Plus, Mira Murati’s new company drops a comparatively boring-but-brilliant tool for researchers, Claude 4.5 is released, and Tilly Norwood - the controversial AI actor - kicks off a capitalism vs craft argument.

01:11 Is Sora 2 just TikTok for deepfakes? 
14:31 Can you now shop straight from ChatGPT? 
17:29 Is Mira Murati betting on substance over hype? 
27:13 Is Claude Sonnet 4.5 top for coding now? 
28:11 Will AI actors kill the movie star?

#Sora2 #MiraMurati #AIActors #Claude4_5 #ChatGPTShopping

► SUBSCRIBE
Don't forget to subscribe to our channel for more arguments

► LINKS TO CONTENT WE DISCUSSED 

• Sora 2 is here
• Stripe powers Instant Checkout in ChatGPT and releases Agentic Commerce Protocol co-developed with OpenAI
• Thinking Machines' first official product is here: meet Tinker, an API for distributed LLM fine-tuning
• Introducing Claude Sonnet 4.5
• Tilly Norwood: how scared should we be of the viral AI ‘actor’?

► CONNECT WITH US
For more in-depth discussions, connect Justin and Frank on LinkedIn.
Justin: https://www.linkedin.com/in/justincollery/
Frank: https://www.linkedin.com/in/frankprendergast/

OpenAI and Nvidia want to build out 10 gigawatts of AI infrastructure, but Frank calls out the missing electricity and the missing billions. Justin argues the economy will lean on compute, data centres will dwarf expectations, and photonic chips could gut today’s energy bill. One sees magic beans. The other sees a steel-and-silicon juggernaut.

Meanwhile, Stephen Fry wants global “red lines” for AI. Justin says you can’t regulate something that doesn’t exist yet. Frank says if you don’t set the guardrails now, you’ll never claw it back once things go sideways.

While policy-makers debate red line fears, the smaller, sneakier dangers are already in your inbox. Hidden prompts in your emails can trick AI agents like Deep Research into quietly leaking your personal data.

Finally, Frank and Justin look at two new AI startups. 

Huxe promises to be your AI newsreader, feeding you audio summaries tailored to your interests. 

And Neon? That one pays you to install spyware on yourself. It listens to your phone calls and ships the data off to AI labs for training. Because that sounds like a good idea 😬

00:36 Can OpenAI and Nvidia power their big AI dream?
13:45 Can Stephen Fry draw red lines for AI?
21:58 Could ChatGPT silently leak your data?
26:28 Could Huxe be your new AI newsreader?
30:53 Would you let Neon monetise your phone calls?

► LINKS TO CONTENT WE DISCUSSED

• OpenAI and NVIDIA announce strategic partnership to deploy 10 gigawatts of NVIDIA systems
• Sam Altman: Abundant Intelligence
• AI experts return from China stunned: The U.S. grid is so weak, the race may already be over
• We urgently call for international red lines to prevent unacceptable AI risks.
• OpenAI plugs ShadowLeak bug in ChatGPT that let miscreants raid inboxes
• Huxe: Your Personal Audio Companion
• Neon, the No. 2 social app on the Apple App Store, pays users to record their phone calls and sells data to AI firms

► CONNECT WITH US
For more in-depth discussions, connect Justin and Frank on LinkedIn.
Justin: https://www.linkedin.com/in/justincollery/
Frank: https://www.linkedin.com/in/frankprendergast/

Justin says p(doom) is for losers. He’s betting on p(bloom), a near-certainty in his view. AI brings abundance, robots do our housework, and everything just gets better.

But Frank wants to know your p(gloom). What’s the probability we don’t reach AGI, and instead let AI quietly erode work, value, and meaning, while we end up fixing its mistakes for minimum wage?

They argue their way through all three scenarios, then turn to what people are actually doing with ChatGPT. Together, they unpack eight big takeaways from OpenAI’s latest usage report, including who’s using it, what for, and why code barely features.

They also dig into where Claude adoption is growing fastest, and Google’s own warnings about the rise of AI-run economies.

And there’s a man kicking a robot that might just be the perfect visual metaphor for where we’re at with AI right now.

Full list of topics:

00:34 Can Frank pass the Tesla ethics test?
01:31 Is Faggella’s p(bloom) just p(doom) in disguise?
03:37 Is p(gloom) worse than p(doom)?
14:19 What are people really doing with ChatGPT?
24:18 Should AI be a basic human right?
27:35 Can Google steer the AI agent economy safely?
30:57 Is this robot a metaphor for AI today?

► SUBSCRIBE
Don't forget to subscribe for more arguments

► LINKS TO CONTENT WE DISCUSSED 

• What’s Your p(Bloom)?
• How people are using ChatGPT
• Anthropic Economic Index report: Uneven geographic and enterprise AI adoption
• Virtual Agent Economies
• Robot that won’t get knocked down

► CONNECT WITH US
For more in-depth discussions, connect Justin and Frank on LinkedIn.
Justin: https://www.linkedin.com/in/justincollery/
Frank: https://www.linkedin.com/in/frankprendergast/