Reece Preston and Scott Quilter return for an emergency broadcast to dive into the massive surprise announcement from OpenAI: the release of their new web browser, Atlas.

Is Atlas the revolutionary tool we've been waiting for? The hosts break down what makes this browser distinct from competitors like Perplexity’s Comet, noting that while Comet feels like a web browser with AI bolted on, Atlas feels like ChatGPT has had a browser added to it, featuring the integrated polish of the ChatGPT experience.

In this episode, you will learn about the key features of Atlas:

• Integrated ChatGPT: The full ChatGPT experience—including all your past conversations, projects, and Custom GPTs—is available in a companion sidebar right next to every window you open. This removes unnecessary steps like copying and pasting, allowing the assistant to see and act directly on the content in your browser.

• Custom GPT Potential: Discover the "huge amounts of potential" now that Custom GPTs and projects can interact with live web pages, potentially automating repeatable manual tasks on regular websites.

• Agent Mode Control: The updated Agent Mode now takes control of your actual browser, allowing it to bypass blocks and interact with secure pages (like your shopping cart). Importantly, you must physically turn Agent Mode on, addressing security concerns seen in other browsers.

• Browser Memory: Discussing the "equally amazing and terrifying" new feature where Atlas tracks and creates a narrative of your web browsing history. While this can be used as a productivity tool (e.g., analysing your day to suggest improvements), listeners should note that you can only archive, not delete, these browser memories.

Important Details & Warnings:

• Currently, Atlas is only available on Mac OS.

• The browser is built on Chromium, meaning your existing Chrome plugins should still work.

• For security and privacy, you can disable Chat features for individual websites, such as banking or accountancy portals.

• The hosts discuss the impact of Atlas's sudden launch, noting that Google's stock dropped 4% as a result, and they explore the potential risks and implications of OpenAI training its models on user browser history.

Tune in to hear Reece and Scott’s initial impressions, standout features (like the ability to transcribe images directly from a web), and why they are encouraging listeners to be curious and play with this powerful, yet brand new, tool

Welcome back to Prompt Fiction, Chapter 5!

Scott Quilter and Reece Preston admit they have "no idea what's going on," but they are ready to tumble headfirst into the weekly mess, mayhem, and technical genius.

This episode starts with Scott correcting a previous error regarding the Artisan/Atlassian confusion and noting the surreal presence of Meta AI on Mercedes F1 car.

We recap Scott’s experience at the GTIA event, featuring keynotes from Stephen Fry and Bob Geldof. They unpack how Fry compared the progression of AI to the initial development of the motor car, while Geldof used the decline of social media (from community "wonderment" to "isolation media") as a warning sign for the future of AI investment. They explore the concept of the "AI bubble bursting"—not that AI will stop, but that investors' need for returns will flood free tools with adverts, turning social media into a dangerous litmus test for AI.

In the News:

• OpenAI's Commercial Conquest: We discuss OpenAI’s major new partnership with Broadcom to create custom AI chips and its significant commercial integration with retailers like Walmart and Shopify. We also break down the launch of the complex new Agent Builder, designed to create specific, interconnected bots, though it currently requires substantial developer skill.

• The Claude Revolution: Reece announces he is officially switching allegiance after being blown away by Anthropic's new models, particularly the lightning-quick Claude Haiku 4.5. We explore the potential of the new Claude Skills, which are described as being "Custom GPTs turned up to 11" with cross-platform consistency.

• Google's Gas Pedal: Google’s AI suite, Gemini, has hit one billion monthly users. The new no-code app building platform Opal looks promising, but Google is also now taking actions to restrict third-party AIs by actively blocking ChatGPT Agent Mode logins.

• Safety and Chaos: We raise serious concerns about the new low hurdle to creating convincing deepfakes, noting that Sora 2 watermark removal tools make it trivial to produce fake videos of public figures. Conversely, we examine DigiLabs' attempt to bring stability with their Uncertainty Engine, providing confidence scores on critical AI outputs.

• Cautionary Tales and Glitches: Finally, we warn listeners about the "shady tactics" of the Otter AI note-taker that spreads like a virus by inviting itself to meetings. And in a moment of pure chaos, we demonstrate how a simple query—"Is there a seahorse emoji?"—can send an AI into an infinite emoji loop meltdown.

In this episode of 'Prompt Fiction', hosts Scott and Reece dive into the chaotic yet brilliant world of AI. They explore topics ranging from AI's role in Formula One sponsorships to the intriguing concept of wetware, a new frontier in computing. The conversation also touches on AI-generated music, copyright issues, and the potential of biocomputing. With humour and insight, Scott and Reece navigate the ever-evolving landscape of technology, offering listeners a blend of entertainment and education.

Keywords

AI, Formula One, wetware, biocomputing, AI music, copyright, technology

Takeaways

• AI is reshaping industries, from Formula One to music.
• Wetware represents a new frontier in computing.
• AI-generated music is both a creative tool and a copyright challenge.
• Biocomputing could revolutionise energy efficiency in AI.
• The ethics of AI training data remain a hot topic.
• Collaboration in AI is becoming more accessible.
• AI's impact on traditional media is profound and ongoing.
• The intersection of AI and human creativity is expanding.
• AI tools require careful oversight to avoid errors.
• The future of AI holds both promise and complexity.

Episode Summary

In Chapter 3 of Prompt Fiction with Reece & Scott, the chaos continues. From AI ministers to smart glasses, billionaires jousting for the top spot, and self-driving cars safer than humans — this episode dives into the weird, wonderful, and sometimes worrying intersections of tech, AI, and culture.

We cover everything from Albania’s AI government minister and Google’s Nano Banana-fuelled surge, to how Foo Fighters are having fun with AI, and even why ASDA and Sainsbury’s don’t want ChatGPT doing your weekly shop.

Expect tangents, banter, and a few eyebrow-raising truths about where AI is taking us next.

Episode Highlights & Links

• 🎬 Critterz – AI-powered animation project expanding into a full feature film. https://www.critterz.tv

• 💰 Larry Ellison overtakes Elon Musk (briefly) thanks to Oracle’s AI cloud boom.

• 🇦🇱 Albania appoints Diella, the world’s first AI government minister.

• 🕶️ Meta’s Ray-Ban & Oakley Smart Glasses – hands-on experience, features, bugs, and what’s next.

• 🍌 Google Gemini & Nano Banana – image generation frenzy pushes Gemini to the top of the App Store.

• 🔗 Google Gems – custom AI assistants you can now share like Google Docs.

• 🗂️ NotebookLM – new features: podcasts, videos, editable prompts, flashcards, quizzes, and mind maps.

• 👩‍💻 GPT-5 Codex – a specialised coding assistant inside VS Code.

• 📊 ChatGPT Usage Stats 2025 – practical use dominates, gender gaps narrow, and personal adoption outpaces workplace rollout. https://tsrs.uk/howpeopleusechatgpt2025

• 🛒 ASDA & Sainsbury’s block ChatGPT Agents – Cloudflare checks foil AI shopping hacks.

• 🚗 Waymo’s safety record – autonomous cars show 79% fewer crashes than human drivers. https://www.perplexity.ai/page/waymo-cars-have-79-fewer-crash-sygydzEiRVSLiOSGRfJhlQ

• 🔌 Perplexity Connectors & Perplexify Me – new integrations and a fun image tool. https://www.perplexity.ai/perplexify_me

• 💸 Big AI features moving behind big paywalls – what this means for accessibility.

• 🚨 Copy-Paste-Send of the Week – a procurement close-call where AI fakery almost slipped through.

• 🕵️ Internet detectives misuse AI – grainy photos, CSI zoom, and dangerous assumptions in the Charlie Kirk case.

• 🎵 Foo Fighters Welcome Back video – a playful AI-powered promo. https://www.youtube.com/watch?v=EfxUI_p6I6Y

Welcome back to Chapter 2 of Prompt Fiction! In this unscripted chapter, your hosts Scott and Reece plunge headfirst into the wild world of AI's latest developments, covering "the good, the bad, and the gloriously broken". 

This week, they celebrate vibe coding triumphs, from building entire websites (even on a phone!) to creating complex payment systems with AI, often outsourcing the heavy lifting to the machines. They also delve into ChatGPT's game-changing updates, including the ability to "branch" conversations for "waffleless" chats – perfect for cutting out tangents – and the new, clearer "ChatGPT for Business". 

But it's not all smooth sailing: the episode explores the impact of a global ChatGPT outage and critically examines the trend of companies like Tomato Plus and Renault slapping "AI badges" on everything from smart greenhouses to animated car mascots. Discover the surprising story of how the tiny island of Anguilla struck gold with the .ai domain boom, diversifying its economy in an unexpected way. 

Get ready for a laugh as they discuss South Park's hilarious take on "sycophantic" AI, which highlights the pitfalls of overly agreeable machines. Finally, they share some truly "glorious fails" in their "copy, paste, send" segment, including the infamously bizarre AI-generated crowd at Will Smith's Scarborough concert, and reveal some fantastic free access to top AI tools. Tune in for a chaotic, insightful, and often humorous look at AI! 

Welcome to Prompt Fiction, the podcast where stories aren't written in weeks or months; they're sparked and shaped in the moment by the power of a single prompt. 

Join hosts Scott and Reece in this unscripted experiment in creativity, as they dive into the latest AI news and explore new tools. Scott, with his background in the tech industry teaching AI, and Reece, a keen enthusiast who loves to play with AI, bring a unique perspective to the world of artificial intelligence. 

Each episode features: 

• Fresh, unexpected tales inspired by prompts. 

• Discussions on cutting-edge AI tools like Chat GPT-5 and Perplexity's Comet browser. 

• Insights into the challenges and responsibilities of AI development and use, including the backlash to GPT-5 and its impact on users. 

• Explorations of practical applications, such as "vibe coding" for non-coders and the use of AI in Excel with Copilot, alongside warnings about early adoption. 

• A look at the future of automated web browsers and image generation with tools like Nano Banana. 

• A special segment called "Copy Paste Send," dedicated to humorous and sometimes alarming examples of AI hallucinations and mistakes that occur when users don't check AI-generated content. 

Tune in for loose, engaging conversations that often go down rabbit holes, providing valuable insights and a touch of humour into the rapidly evolving world of AI.