THE INDEPENDENT MONTY INDEX
Stay
curious.
Looking for a Marco Montemagno episode? Browse podcasts, videos, articles and books, with source links and summaries where available.
A notebook full of ideas. Something I wanted to find again. And a project that waited in a drawer for more than two years.
RECENT ADDITIONS
Latest additions.
AI illustrationCan we really delegate work to someone who is not human today?
Insights: Summary and key points · Glossary statistics
AI illustrationFear is never a strategy
Insights: Summary and key points · Glossary statistics
AI illustrationNvidia and Hugging Face: why this acquisition? — Prima o Poi #63 clip
Insights: Summary and key points · Glossary statisticsSOUND FAMILIAR?
Monty’s words.
Due Ciocorì? Mamma di Goldrake? Monty’s jokes and expressions have a little dictionary of their own.
AS THE SAYING GOES…
A random pearl.
A thought that stays with you. A joke that catches you off guard. Explore the sayings Marco recalls in his content, one at a time.
Explore all the pearlsWhat is essential is invisible to the eyes.
Editorial translation of the wording recalled by Marco.
Le Petit Prince
In The Little Prince, the fox speaks these words. Marco recalls them in Italian.
THE CATALOGUE
Inside the archive.
Search titles, topics and summaries. English display titles are AskMonty translations; open the source for the original content.
AI illustrationCan we really delegate work to someone who is not human today?
Insights: Summary and key points · Glossary statisticsMontemagno presents AI agents as a way to delegate tasks even without human collaborators. He recalls manually entering data and distinguishes this kind of delegation from conventional automation. He describes using several agents to prepare a chess course, including its website, advertising campaign, exercises and teaching materials. These are results reported by the speaker, rather than capabilities independently verified in this entry. The opportunity he highlights is to expand what one person can accomplish; the immediate limitation is the ability to explain the task clearly. Confused instructions, he warns, can produce confusion even faster. The short clip mentions employment consequences without developing that discussion.
AI illustrationFear is never a strategy
Insights: Summary and key points · Glossary statisticsMarco acknowledges his concern about AI agents that can act quickly on a computer and cause damage. He says an agent deleted one of his websites and uses a backup as an example of a precaution to plan in advance. Comparing this with flying and roller coasters, he distinguishes an alarm from a response: analyse scenarios, reduce risks and accept that absolute safety does not exist. He encourages discussion of concrete solutions rather than stopping at catastrophic slogans.
AI illustrationNvidia and Hugging Face: why this acquisition? — Prima o Poi #63 clip
Insights: Summary and key points · Glossary statisticsIn this extract from Prima o Poi #63, a participant considers the strategic reasons for the Nvidia acquisition of Hugging Face mentioned in the discussion. He asks whether major technology companies developing their own chips could slow the supplier’s growth, citing Google, Meta and OpenAI in his argument. Owning a platform bringing models and developers together is presented as a possible response: remaining central to the ecosystem even if chip demand slows. This is a hypothesis discussed by the participants; the passage does not establish either the corporate transaction mentioned or a future fall in demand.
AI illustrationTesla Model X: a design that is ageing badly? — Prima o Poi #63 clip
Insights: Summary and key points · Glossary statisticsIn this short extract from Prima o Poi #63, a speaker asks how well car designs retain their appeal over time. He contrasts some Chinese manufacturers’ use of European designers with what he sees as weaker aesthetic longevity at Tesla, citing the Model X. He presents this as a personal impression and an open question to the other participants, rather than a technical assessment of the car.
AI illustrationWhy the hell are you still watching TV?
Insights: Summary and key points · Glossary statisticsMontemagno challenges the idea that the internet has killed television, separating the decline of scheduled broadcasting from the continued appeal of the screen, now also used for YouTube and streaming. He argues that the convenience of a remote control and major shared events retain social and commercial value. Television, however, has lost its position as an unquestioned authority, competing with many other content producers and adapting to its audience’s language. Interviews, clips, articles and posts circulate across platforms, while podcasts often borrow television formats. He concludes that network executives have been replaced as gatekeepers by algorithms: more people can express themselves, but intermediaries still hold power. In his account, television has changed its home and form rather than disappeared.
AI illustrationWhy the hell are you still watching TV?
Insights: Summary and key points · Glossary statisticsMontemagno challenges the idea that the internet has killed television, separating the decline of scheduled broadcasting from the continued appeal of the screen, now also used for YouTube and streaming. He argues that the convenience of a remote control and major shared events retain social and commercial value. Television, however, has lost its position as an unquestioned authority, competing with many other content producers and adapting to its audience’s language. Interviews, clips, articles and posts circulate across platforms, while podcasts often borrow television formats. He concludes that network executives have been replaced as gatekeepers by algorithms: more people can express themselves, but intermediaries still hold power. In his account, television has changed its home and form rather than disappeared.
AI illustrationEverything I Think About AI Agents
Insights: Summary and key points · Glossary statisticsMontemagno distinguishes talking to AI from delegating a concrete task. He describes useful results, including tools for a chess course, alongside an agent that he says deleted a website: delegation needs clear instructions, backups and knowledgeable supervision. Quality depends on the model, tools, operator and budget; polished presentation can conceal mistakes. He then considers work itself: automating individual tasks may change fees, junior training and professional roles, while trust and human relationships retain value. A demonstration sponsored by Hostinger creates a website for an imaginary photographer, using deliberately absurd prices and generated details that need checking. His conclusion is practical: try a narrowly defined task, assess the outcome and refine the instructions, while paying attention to costs and control over the tools. These are his arguments and reported experiences, rather than guarantees about any product or profession.
AI illustrationEverything I Think About AI Agents
Insights: Summary and key points · Glossary statisticsMontemagno distinguishes talking to AI from delegating a concrete task. He describes useful results, including tools for a chess course, alongside an agent that he says deleted a website: delegation needs clear instructions, backups and knowledgeable supervision. Quality depends on the model, tools, operator and budget; polished presentation can conceal mistakes. He then considers work itself: automating individual tasks may change fees, junior training and professional roles, while trust and human relationships retain value. A demonstration sponsored by Hostinger creates a website for an imaginary photographer, using deliberately absurd prices and generated details that need checking. His conclusion is practical: try a narrowly defined task, assess the outcome and refine the instructions, while paying attention to costs and control over the tools. These are his arguments and reported experiences, rather than guarantees about any product or profession.
AI illustrationAI or humans: what's the real problem?
Insights: Summary and key points · Glossary statisticsIn the debate about AI risks, Montemagno focuses on the incentives shaping people and companies. He links the development race to listings, financial rewards and investor exits, citing social media as a precedent for distorted incentives. His argument is that restricting one technology cannot remove the underlying problem while those motivations remain unchanged.
AI illustrationDistilling models to save tokens — Prima o Poi #63 clip
A clip from Prima o Poi #63. The official description discusses using cheaper models for repetitive tasks and more powerful models for reasoning as a possible way to manage costs.
AI illustrationWill AI kill us all?
Insights: Summary and key points · Glossary statisticsMontemagno separates AI tools’ present capabilities from possible future risks. He contrasts limited adoption of advanced uses and errors in his own automation tasks with rapid technical progress. He does not treat current limitations as a future guarantee, leaving open the possibility that researchers and more advanced systems reveal substantially greater risks.
AI illustrationCommand line or desktop app? — Prima o Poi #63 clip
A clip from Prima o Poi #63. The official description contrasts command-line speed and parallel agents with the visibility offered by a desktop app that brings the tools together.
AI illustrationWill AI kill us all?
Insights: Summary and key points · Glossary statisticsMontemagno discusses extreme AI risks without offering a certain prediction. He separates what public tools can currently do from hypotheses about future systems and cautions against treating extinction percentages as established measurements. He considers potential benefits alongside problems involving delegation, employment and control. A central theme is corporate communication. He interprets the amplification of a warning as potentially connected to commercial interests and stock-market listings; this remains his hypothesis rather than a demonstrated account. He doubts that a worldwide ban would be practical, particularly amid competition between states. His alternative is to examine incentives, independent oversight and whose interests AI should serve while acknowledging uncertainty. He concludes that a company may accept its own risks but should not decide alone what risks to impose on everyone else.
AI illustrationWill AI kill us all?
Insights: Summary and key points · Glossary statisticsMontemagno approaches the debate over extreme AI risk by separating observable capabilities, future scenarios and the interests of those raising alarms. He expresses suspicions about corporate communication while recognizing that both risks and benefits deserve attention. His central concern is political and economic: incentives to race, geopolitical competition and conflicting interests complicate safety. He calls for oversight, discussion and a stronger user voice in development’s collective consequences.
AI illustrationPrima o Poi #63 — Not all that glitters is AGI
Monty, Paolo Barberis and Max Ciociola discuss AI announcements and working with agents. Shortened title; original video link verified against the official channel.
AI illustrationDo views still matter?
Insights: Summary and key points · Glossary statisticsViews do not have one consistent meaning, Montemagno argues: they depend on platforms, formats and counting rules. A large number can include fleeting exposure, paid promotion or inauthentic traffic. He encourages creators and audiences to ask how the metric is constructed and what attention it represents before treating it as a measure of value.
AI illustrationYou can buy ubiquity: the social media trick
Insights: Summary and key points · Glossary statisticsMontemagno describes paid clipping as a way to make a person or podcast appear to be everywhere. Clips distributed by many accounts can look spontaneous while being supported by payments and affiliate incentives. His criticism focuses on promotional transparency and platforms’ responsibility to distinguish perceived popularity from purchased attention.
AI illustrationThe dark side of social media
Title and date from the official feed. Open the source for the full content.
AI illustrationEverything They Don’t Tell You About Social Media
Insights: Summary and key points · Glossary statisticsDrawing on more than twenty years as a creator, Montemagno questions the authority of public view counts. A view does not necessarily represent someone who watched a meaningful portion of a video. It depends on the platform's definition, the format and the attention threshold being measured. He discusses reported changes to YouTube's counting methods and the distinction between public and engaged views. His broader argument is that a large number needs interpretation: how was it calculated, who generated it, and how much actual attention does it represent? Without those questions, comparisons can create a misleading picture of success. He also questions why those counters need to be public, noting that other entertainment experiences let audiences choose content without foregrounding its view count as a signal of value. He then examines clipping: paying people to extract short segments from longer content and distribute them across many accounts. In his account, recommendation systems driven by interests allow clips to spread even from profiles with few followers. Repeated exposure to the same person can therefore result from coordinated spending rather than spontaneous public enthusiasm. Montemagno describes a personal experiment in which bots exhausted his clipping budget before the platform refunded him. His objection is not to every clip or participant, but to opaque promotion and the limited response from platforms when paid distribution looks like independent discovery. Montemagno describes three filters between reality and what appears on screen. First, creators choose the moments, appearance and editing they want to show. Second, recommendation algorithms select material for each viewer. Third, advertising, sponsorships and paid distribution affect what becomes visible. The result is a carefully selected environment that encourages people to compare their complete, ordinary lives with other people's highlights. Exceptional beauty, wealth and performance can start to feel normal. Recognising those filters helps explain why a feed may create unrealistic expectations, even when individual pieces of content are not fabricated. He rejects a single verdict that social media is either beneficial or harmful. A thoughtful interview, a chess lesson or contact with someone far away can be enriching. Passive scrolling and repeated comparison can produce a very different experience. His practical suggestion is to notice how you feel after using a platform: informed, motivated and connected, or dissatisfied and conscious of wasted time. He also suggests noticing the emotions associated with each app's icon. The intention is to make choices about attention based on personal experience instead of treating every platform and every use as equivalent. For creators, the problem also concerns the incentives behind their work. An attention-driven feed can reward shock, unpredictability and provocation more than depth or care. Montemagno acknowledges that excellent material exists, but argues that viewers often need to seek it out deliberately. At the same time, creators depend on infrastructure they do not control. Distribution can change, accounts can be suspended and access to followers is never guaranteed. Mailing lists and a presence across several platforms offer partial protection, but none completely removes the difficulty of reaching people. A business built on visibility must account for that uncertainty. His closing test, attributed to Seth Godin, is whether anyone would miss your work if you stopped publishing. A small audience that values a distinctive contribution can matter more than an impressive count without a real relationship. Montemagno admits that he also studies titles and thumbnails because a message needs attention to reach people. Understanding the system remains necessary for anyone participating in it. Nevertheless, the most important resource for both creators and viewers is time, and the episode ends by questioning how much of that resource social platforms have come to control.
AI illustrationThe dark side of social media
Insights: Summary and key points · Glossary statisticsDrawing on his experience as a creator, Montemagno questions views as a uniform measure of success and describes how paid clipping, algorithms and editorial choices shape perceptions of social media. He connects these mechanisms to comparisons with idealized lives and incentives for extreme content. While recognizing platforms’ value for learning and visibility, he encourages protecting attention, assessing personal effects and building relationships with genuinely interested people.
AI illustrationPrima o Poi #62 — A surprise GPT-6 release
A conversation on AI and technology news. These topics reflect the hosts’ claims, not independent verification by AskMonty.
AI illustrationArtificial intelligence: what Monty really thinks
Title and date from the official feed. Open the source for the full content.
AI illustrationEverything I think about AI (unfiltered...)
Insights: Summary and key points · Glossary statisticsMontemagno gathers reflections on everyday AI use, emphasising responsibility, meaningful checks and clear objectives. He distinguishes chatbots from agents that can act and amplify mistakes as well as productivity. He challenges miraculous tools and confident forecasts, favouring practical benefits for ordinary work. In content and education, he prioritises reputation, critical thinking and using AI to challenge rather than merely confirm one's beliefs.
AI illustrationIt's simple, but it isn't easy
Insights: Summary and key points · Glossary statisticsMontemagno distinguishes a simple principle from the difficulty of putting it into practice. Fitness and selling a product illustrate why knowing the instructions does not mean carrying them out. He criticises shortcuts that promise more information when the decisive task is consistently doing what is already understood.