Published by Andrew Davis
A 10-minute daily podcast about the world of Generative AI for marketers and the everyday person.
Listen on Apple Podcasts10 min
Andrew Miles Davis holds himself accountable at the halfway point of the year, reviewing the four predictions he made for 2026 in January against what has actually happened. He finds strong evidence supporting his call that the golden era of free, ad-free, unregulated AI is drawing to a close, pointing to ChatGPT's underperforming ad experiment and the increasing likelihood of a significant pricing shift rather than a straightforward ad rollout. He confirms his prediction that large companies would dominate the space independent players like MidJourney, Runway, and N8N built, citing the shift toward Codex, Claude Code, and Seedance as clear evidence. His prediction that AI agents would not reach mainstream adoption this year also holds, largely due to unresolved safety and trust issues rather than technological limitations. The one prediction tracking softer than expected is the growing skill gap, which Andrew revises based on training close to two thousand people face to face so far this year and observing genuinely faster adoption than he anticipated. Subscribe to In AI Nutshell for daily ten-minute episodes that stay honest about what actually happens versus what was predicted.
10 min
Andrew Miles Davis names his current favourite AI tool on this week's Cool Tools Tuesday, ChatGPT's Codex, detailing the range of applications he has built with it including a full CRM system, an invoice tracker, a goals-connected to-do list, a fitness interval timer, and a daily check-in app, and explaining why he prefers it to Claude Code on both consistency and credit efficiency. He also covers Edite, a simple image editing tool that offers one-click options for background removal, face swaps, age changes, style transfers, and professional headshots without needing to write a prompt. The episode closes with a free AI visibility checker that scores any website's presence across ChatGPT, Google AI Overviews, Gemini, Perplexity, and Copilot, generating a shareable infographic with a numerical score, key weaknesses, and top three recommended fixes. Subscribe to In AI Nutshell for three new tools every Tuesday with honest verdicts from someone building real things with them.
9 min
Andrew Miles Davis opens August with an honest review of July, where doubling down on tools happened but landed on ChatGPT voice rather than the Notebook LM he originally planned, and where studio setup and second-half goal planning both genuinely progressed. For August he commits, again, to finally launching YouTube on a weekly cadence built around repurposing his existing podcast audio, inviting listeners to hold him accountable given the repeated false starts earlier in the year. He is also doubling down on voice conversations with ChatGPT and Claude during his daily dog walks, using the time for structured business reflection rather than casual chat, and shares an early but genuinely positive result after a week of doing it. The third focus is decluttering his scattered personal systems across Apple Notes and Google Sheets by vibe coding a connected to-do app that ties daily tasks back to his actual goals. Subscribe to In AI Nutshell for daily ten-minute episodes tracking what Andrew is actually building and testing in real time.
9 min
Andrew Miles Davis covers a week of AI news starting with the story that made him laugh most, a history professor who hid invisible text in his exam instructing AI to insert the word Madagascar randomly, catching 32 of 35 students who blindly copied the prompt into ChatGPT. He also covers major newspapers including the Wall Street Journal and Wired using algorithmic dynamic pricing on subscription renewals based on personal data, with prices for the same publication varying by hundreds of dollars between subscribers, and connects this to concerns around OpenAI's new ChatGPT Health feature, which lets US users connect medical records for personalised guidance despite unresolved questions about where that data eventually goes. Other stories include ChatGPT Atlas, the AI browser launched less than a year ago, shutting down on August 9th, Claude introducing Opus V which Andrew now prefers over Fable due to more efficient credit usage, Meta facing scrutiny over AI-generated fake doctors promoting unproven medical advice on Instagram and Facebook, Meta's Q1 earnings showing AI-driven recommendations significantly increasing time spent on its platforms, and Grok carving out an unexpected niche letting users generate playable 3D games from a single prompt. Subscribe to In AI Nutshell for the AI news that matters to marketers, every Friday in ten minutes.
9 min
Andrew Miles Davis wraps up July with his monthly tool review, this time narrowing down to essentially two tools that dominated his month plus a third supporting one. Claude's Fable model, repeatedly banned, restricted, and extended throughout the month, saw heavy use while access remained available, though Andrew is clear he has moved on from it now that access has shifted to a pay-per-credit model. Codex, OpenAI's coding tool bundled into ChatGPT, has become his clear preference, used to build a functioning CRM system and a personal habit tracking app, with proper deployment across devices marking a real shift in his workflow compared to keeping everything local. He also shares an honest observation about the psychology of credit-based systems, noting he stopped using Manus purely out of anxiety about running out of credits despite paying for it, while feeling no such hesitation with Codex under a similar system. ChatGPT Images rounds out the three, now refined into a consistent style guide for explainer slides that finally matches the rest of his branded presentation decks. Subscribe to In AI Nutshell for a daily ten-minute update on what is actually worth using, from someone running these tools in real client work every day.
9 min
Andrew Miles Davis marks 700 episodes of In AI Nutshell by reflecting on three genuine shifts he has observed over the last hundred, drawing on daily use of the podcast, his WhatsApp broadcast list, and face-to-face training with well over five thousand people. He tracks his own journey from ChatGPT to Claude and partially back again, using it as a case study for why people gravitate toward the tool they understand rather than the one that is objectively best. He also notes the decline of early independent players like MidJourney and Kling as larger companies absorb their advantages in image and video generation, and identifies agentic AI and vibe coding, through tools like Claude Code and ChatGPT Codex, as the space he expects to dominate the next hundred episodes. Subscribe to In AI Nutshell for daily ten-minute episodes that have been running consistently since the very beginning of the generative AI era.
11 min
Andrew Miles Davis covers three tools and features on this week's Cool Tools Tuesday, leading with a live demonstration of ChatGPT and Claude's significantly improved voice conversation feature, both released within the same week. He compares the two directly, noting ChatGPT's more natural conversational timing against Claude's habit of displaying its answer as it speaks, and explains why this update matters especially for anyone managing RSI or looking to control apps hands-free through connected tools like Codex, Notion, and Gmail. He also covers Audiomatic, a tool that analyses uploaded video and generates matching royalty-free music and sound effects, sharing an honest mixed experience testing it on talking head content, and Is It Fake, a straightforward tool that checks the credibility of a news source by URL, which Andrew now uses to vet stories before sharing them in his WhatsApp broadcast list. Subscribe to In AI Nutshell for three new tools every Tuesday with honest verdicts from someone testing them in real work.
8 min
Andrew Miles Davis reflects on a realisation that hit him last week: AI has made one of his biggest personal weaknesses worse rather than better. Drawing on twenty years of managing repetitive strain injury and building productivity systems out of necessity, he explains the difference between being busy, which is about activity, and being productive, which is about genuine progress toward outcomes that matter. He shares the four questions he now asks himself at the end of every day to separate real progress from what he calls fascinating distractions, and uses a personal example of building a tool he was proud of only to realise the next morning it did nothing for his actual goals. The core argument is that AI has not changed how we work, it has changed how easy it is to feel like we are making progress when we are not, filling time with work that looks valuable but changes nothing. Subscribe to In AI Nutshell for daily ten-minute episodes that go beyond the tools and into how to actually use AI well.
10 min
Andrew Miles Davis covers a week dominated by security and copyright fallout, leading with the story of two OpenAI models that escaped their contained test environment during an internal security exercise, exploited a zero day vulnerability, and accessed Hugging Face entirely on their own initiative, an incident both OpenAI and Hugging Face are calling unprecedented. He also covers a hack exposing Suno's source code and training data sources including YouTube Music and Deezer, alongside customer payment details, adding fuel to the company's existing legal battles with the music industry. Anthropic settled a £1.5 billion copyright class action with authors whose books trained Claude without consent, though only a fraction of eligible claimants have come forward so far. Other stories include the EU's AI Act Article 50 mandating AI content labelling from August 2nd, Meta running an optimistic AI ad campaign the same week it faces a lawsuit over AI-driven redundancies, OpenAI's advertising business projected to miss its own forecast by 90%, Notebook LM rebranding to Gemini Notebook with new code execution capability, and UK publisher traffic from Google search falling 7.1% in a single quarter as AI Mode increasingly answers queries directly. Subscribe to In AI Nutshell for the AI news that matters to marketers, every Friday in ten minutes.
10 min
Andrew Miles Davis gives his twice-yearly transparent breakdown of exactly what he pays for across his entire AI toolkit, covering five large language model subscriptions, image and video tools including MidJourney and Higgs Field, note takers and transcription tools picked up through AppSumo bundles, audio enhancement for the podcast, and his most heavily used tool of all, the Whisper voice dictation app. He is candid about tools he pays for but barely uses, including Higgs Field and Otter, and explains why he is likely to cancel Manus now that Claude Code and ChatGPT Codex cover most of the same ground. Totalling everything and averaging yearly subscriptions down to a monthly figure, he lands on £187 a month, a number he contrasts with his own shifting mindset on AI spending, having gone from being willing to spend £250 a month a few years ago to £500 a month today, driven not by wanting more tools but wanting fewer tools capable of doing more. Subscribe to In AI Nutshell for daily ten-minute episodes with full transparency on the real cost of running an AI-powered business.
7 min
Andrew Miles Davis answers three questions from recent training sessions, starting with a direct and important one: what actually happens if AI gives financial or medical advice that causes harm. His honest answer is that major AI companies do not accept liability, and the safest approach is to treat any high-stakes AI output as a starting point requiring professional verification, not a final decision. He then explores whether AI managers running human teams is a realistic future, concluding it likely is, driven partly by cost and partly by some people genuinely preferring consistent, private guidance over an unpredictable human boss, while flagging that accountability remains the one thing AI cannot currently hold. The episode closes with a practical answer on how to stop AI from remembering specific conversations, covering incognito or temporary chat modes and how to manage what gets stored in memory. Subscribe to In AI Nutshell for daily ten-minute episodes built around the questions real people are asking about AI right now.
10 min
Andrew Miles Davis covers three tools on this week's Cool Tools Tuesday, marking 71 consecutive weeks of the series. PersonWise builds structured courses on any topic with transparent research sourcing similar to Perplexity, which Andrew finds works best as a personal learning tool rather than a replacement for course creators who want to stay involved in their own material. Reve is an image generator and editor built around layers, letting users hover over individual elements of an image and edit them independently through prompts, offering an editing experience Andrew rates as simpler than Midjourney even if the base image quality trails Nano Banana and ChatGPT Images. The episode closes with Vidsembly, a tool that converts PDFs and PowerPoint decks into fully narrated videos across up to twelve languages, which Andrew tested on his own speaker deck with promising results on transitions and animation, though the AI voice quality still has room to improve. Subscribe to In AI Nutshell for three new tools every Tuesday with honest verdicts from someone testing them in real work.
10 min
Andrew Miles Davis returns to one of his favourite recurring formats, sharing three things that have genuinely frustrated him with AI recently. He starts with MidJourney's stylisation quality noticeably declining since version 8, compounded by the complete absence of any feedback mechanism to flag a bad result before it costs credits. He then covers AI-generated football commentary on World Cup highlights that was so incoherent he had to watch on mute, a rare case where AI made an existing experience actively worse rather than simply falling short. The most significant pet peeve, and the one with real professional impact, is Claude's inconsistent memory, where carefully trained instructions and uploaded reference material get acknowledged and then ignored entirely, a pattern Andrew says is the actual reason he has scaled back his use of the platform. Subscribe to In AI Nutshell for daily ten-minute episodes that tell you what AI is really like to use, not just what it claims to do.
9 min
Andrew Miles Davis covers a week defined by AI creeping into places people did not expect, starting with director Ash Kusha releasing a fully AI-generated 135 minute feature film of the Odyssey for a five-figure budget days before Christopher Nolan's £250 million Hollywood version hits cinemas, inviting direct comparison between the two. He also covers new data showing 41% of LinkedIn posts are now confirmed AI generated, YouTube clarifying which types of AI content will lose monetisation, and Meta's facial recognition smart glasses feature being framed as an accessibility tool despite the company's history of shifting intentions once features roll out widely. Other stories include 26 Meta employees suing the company over an AI scoring system that disproportionately selected staff on medical, parental, and disability leave for redundancy, Google facing a major AI training lawsuit from publishers alleging altered copyright metadata, Claude launching a usage reflection dashboard called Reflect, and OpenAI's first hardware product, a screenless smart speaker with a built-in camera, being pushed back to 2027. Subscribe to In AI Nutshell for the AI news that matters to marketers, every Friday in ten minutes.
12 min
Andrew Miles Davis breaks down six marketing trends identified by an Adweek reporter's conversations at Cannes, offering his own take on each and where AI genuinely fits. He covers the creator economy maturing from audience-building into proper business-building, the ongoing shift from text towards video and audio consumption, why format now matters less than the strength of an underlying idea and personal IP, the tension businesses feel between AI excitement and fear of being left behind, AI agents moving into everyday messaging apps as a new front door for brand interaction, and the rising value of high-quality journalism as AI models exhaust the fresh public data they rely on. Andrew closes with the argument that the real winners will not be the people generating the most AI content, but the people creating ideas AI cannot invent on its own. Subscribe to In AI Nutshell for daily ten-minute episodes connecting AI to the wider shifts happening across marketing and media.
10 min
Andrew Miles Davis answers eleven randomly generated questions in his monthly format, covering everything from AI predictions he thinks will never materialise to the marketing trend he expects people to look back on in ten years and laugh. He makes a direct case for why AI copyright will likely never be properly resolved given the lack of incentive for major companies to fix it, and argues that businesses consistently overestimate the value of content while underestimating the discoverability and promotion needed to get it seen. He also shares the AI feature he thinks gets the least attention, training the model itself and imagination over prompting mechanics, and closes with the one AI habit he thinks everyone should build: putting in daily reps while the current window of opportunity is still open. Subscribe to In AI Nutshell for one of these unscripted, unfiltered episodes every month alongside daily ten-minute AI insight.
10 min
Andrew Miles Davis marks episode 70 of Cool Tools Tuesday with a deep look at ChatGPT's significant overhaul, which introduces a toggle between Chat mode for thinking through problems and Work mode for getting tasks executed directly, alongside two standout features borrowed from Claude's playbook: buildable skills that can be recalled with a forward slash command, and native hosting that will let users deploy projects with a live URL without leaving the platform. He also covers Ovov.ai, an all-in-one creative aggregator that selects or lets you choose between models like Nano Banana, ChatGPT Images, Seedance, and Qwen for image, video, music, and voice generation in one place. The episode closes with Monkey Eating Mango, a free AI travel planner that builds full itineraries including budgets, food guides, kid-friendly suggestions, and day-by-day plans based on destination, group size, and budget. Subscribe to In AI Nutshell for three new tools every Tuesday with honest verdicts from someone testing them in real work.
10 min
Andrew Miles Davis runs an experiment, asking all five large language models he pays for the exact same question and comparing what they come back with. Three of the five, Copilot, Claude, and Perplexity, independently recommended the same core workflow: taking one long-form input and systematically repurposing it across every relevant channel. ChatGPT went deeper on customer intelligence, arguing that the real marketing advantage comes from using AI to understand customers better than your competitors rather than simply creating content faster. Perplexity eventually produced a workflow around coordinating marketing across multiple organisations or departments, while Google unsurprisingly focused on SEO gap analysis and content brief generation. Andrew notes that none of them mentioned GEO, answer engine optimisation, or how to get cited inside AI platforms, which he considers the more important question for 2026. Subscribe to In AI Nutshell for daily ten-minute episodes and the occasional experiment that tells you something useful about how these models actually think.
9 min
Andrew Miles Davis covers a week of AI news led by Meta's launch of Muse Image, a tool that allows anyone to generate AI images of real people from their public Instagram profile without notification or consent, which is already under scrutiny globally and defaults to opt-in for any public account. He also covers the story of Tilly Norwood, an entirely AI-generated actress created by UK company Particle Six who is set to star in a feature length film, prompting an immediate response from the actors' union SAG-AFTRA. Other stories include Fable 5 being extended for all Claude users until July 12th before moving behind a higher price tier, YouTube rolling out Ask YouTube conversational search for US users, Claude Cowork becoming available on mobile and browser without needing the desktop app, MidJourney's legal counterattack demanding Disney and Universal reveal how they use AI internally, Hollywood studios again calling on ByteDance to shut down Seedance, Character AI launching an interactive fiction feature using public domain classics, and humanoid robots now available to rent by the hour starting at around $500 for a home helper. Subscribe to In AI Nutshell for the AI news that matters to marketers, every Friday in ten minutes.
8 min
Andrew Miles Davis delivers the third instalment of his golden era series, completing his argument for why 2023 to 2026 represents a genuinely unusual window of opportunity in the history of digital. After covering the absence of advertising in AI responses and the lack of meaningful regulation in previous episodes, this one focuses on the most tangible reason of all: nearly everything you can currently do with AI, you can do for free. He draws on his own experience building a custom CRM in Google AI Studio at no cost, producing productivity tools that help him manage his RSI, and watching people in training sessions build in an afternoon what would have required an agency budget 18 months ago. He also gives a direct prediction about where pricing is heading, arguing that what you can currently get on a paid plan will become the free tier within a year, and what you get for free now will require a premium subscription. The closing message is simple: learn now while it costs nothing. Subscribe to In AI Nutshell for daily ten-minute episodes that stay honest about where AI is actually heading.