← Back to search

TECH GARAGERS: MANGOS, SpaceX, AI Agents, IPO, and Apple: who really commands the new era of AI?

Business Garage, Interviste podcast e news su Imprese, Affari, Tech, AI, StartUp, Marketing, Finanza · 2026-06-20 · 63 min
relevance 63 8744 words Episode page ↗ Audio ↗
Show full episode description
🔧 In questa nuova puntata di Garagisti Tech alziamo la serranda su una settimana densissima per il mondo tech, AI, startup e mercati. 🚀 Insieme a Vincenzo Cosenza, Luca Barboni e Gianluca Testa, commentiamo alcune delle notizie più interessanti del momento, cercando come sempre di andare oltre il titolo e portarci a casa qualche “bullone” utile da avvitare agli ingranaggi delle nostre aziende. 🥭 Arrivano i MANGOS Partiamo dai MANGOS, il nuovo acronimo che Wall Street usa per indicare i player più rilevanti della nuova era tecnologica: Meta, Anthropic, Nvidia, Google, OpenAI e SpaceX. Una sigla che segna simbolicamente il superamento dei vecchi FAANG e racconta molto bene il passaggio da un’economia dominata dalle piattaforme digitali a una guidata da AI, infrastrutture, spazio e potenza computazionale. Rimangono fuori Apple e Amazon. 🤖 Hermes e il lavoro agentico Parliamo poi di Hermes, un nuovo agente AI open e agnostico, utilizzabile anche in locale, che apre un tema sempre più centrale: come sta cambiando il nostro modo di lavorare con gli agenti? Dal terminale a Visual Studio Code, da Claude Code ad Antigravity, fino agli strumenti più accessibili per creare workflow e automazioni, la domanda è sempre la stessa: questi strumenti ci rendono davvero più produttivi o rischiano di illuderci di poter costruire qualunque cosa? 💸 Claude e la fine dell’AI “flat” Altro tema caldo: Claude cambia modello di pricing per alcuni utilizzi agentici. L’epoca dell’abbonamento flat inizia a scricchiolare e l’AI si avvicina sempre più a una logica “a consumo”, quasi come una bolletta dell’intelligenza artificiale. Una trasformazione che può avere un impatto enorme su aziende, sviluppatori, creator e consulenti. 🛑 Anthropic, modelli bloccati e geopolitica dell’AI Entriamo poi nella parte più geopolitica della puntata con il caso Fable 5 e Mythos 5, i modelli di Anthropic bloccati per motivi legati alla sicurezza nazionale e ai controlli sull’esportazione. Quando un modello AI diventa abbastanza potente, smette di essere solo software e diventa infrastruttura strategica, leva industriale e questione geopolitica. 🇪🇺 Mistral e la sfida europea Da qui il ragionamento si sposta sull’Europa: Mistral prova a raccogliere nuovi capitali per restare nella partita dell’AI generativa, ma la domanda resta aperta: l’Europa può ancora competere sui modelli di frontiera o deve giocare una partita diversa, più applicativa, verticale e industriale? 🚀 SpaceX, OpenAI, Anthropic e la stagione delle IPO Ampio spazio anche alla stagione delle grandi IPO tech: SpaceX, OpenAI, Anthropic e il possibile sbarco a Wall Street di Bending Spoons. Una nuova ondata che non riguarda solo la finanza, ma il modo in cui il mercato sta provando a dare un prezzo al futuro dell’intelligenza artificiale, dello spazio, delle infrastrutture e della distribuzione. 🍏 Apple Intelligence e il rischio di perdere il treno dell’AI Infine parliamo di Apple Intelligence, del ritardo di Apple sull’AI, del rapporto con Gemini, della possibile trasformazione dell’iPhone in un marketplace di intelligenze artificiali e dell’istruttoria AGCM legata al Digital Markets Act. Apple resta un love brand fortissimo, ma il punto è chiaro: nell’era dell’AI, la distribuzione non basta più se qualcuno riesce a cambiare il corso del fiume a monte. ⚙️ Una puntata ricca, concreta e molto Garage Meno hype, più bulloni. Come sempre, Garagisti Tech è il nostro antidoto contro lo stress di restare aggiornati su un mondo tech che cambia ogni settimana. 👉 Alziamo la serranda. Si parte. #GaragistiTech #MarketingGarage #IntelligenzaArtificiale #AI #Startup #TechNews #Innovazione #ArtificialIntelligence #GenerativeAI #AIagents #ClaudeAI #OpenAI #Anthropic #Nvidia #Google #Meta #SpaceX #AppleIntelligence #MistralAI #BendingSpoons #VentureCapital #IPO #GrowthMarketing #DigitalTransformation #BusinessInnovation #gianlucatesta #vincenzocosenza #lucabarboni
✨ Episode Outline — click any point to jump to it in the episode
Problem solved
Which players and AI agent tools really lead the new era of AI, and how marketers should pick a practical agent stack.
Benefits
  • Simple macOS desktop install for agent framework via Hermes
  • Research, scraping, synthesis to speed marketing workflows
  • Consult 500 sources at once instead of 5
  • Pre-configured LLM access avoids API key setup
  • Compare strengths/weaknesses across competing agents
Use cases
  • Vincenzo tested Hermes Agent on macOS, found great potential
  • Luca uses Claude Cowork/Claude Code for research, scraping, content drafts
  • Gumloop builds node and agent workflows with bundled LLM consumption
  • ChatGPT GPTs generate YouTube cover images
  • Gemini Spark beta-tested as Cowork-type agent in Google ecosystem
KPIs / results
  • Claude $20 plan limited to chats, moving to usage-based pricing
  • Anthropic overage cap defaulted to $10,000
  • Gumloop subscription $20-$30/month including LLM consumption
  • 13th episode of Tech Garagisti
Tools / build
0:00 / 0:00
🌐 This transcript was automatically translated to English from the original.
Welcome to a new episode of Garagisti Tech, the independent format for commenting on economic and tech news on AI and the world of startups. As always, we do it from a garage marketing perspective, so we try to get our hands dirty and give you useful and applicable content to our realities, to our companies. They are what we call the little bolts to put in our pocket at the end of each episode, which for us every little bolt is a bit of a starting point, an idea, something to reflect on. Today I am in the company of two exceptional guests because it is the first time to welcome Vincenzo Cosenza, therefore a very well-known face of the internet, AI marketing consultant, innovation and popularizer. Do you want to add something Vincenzo? No, look, that's fine. Thanks Gianluca for your hospitality. Edolbe is around Luca Barboni who was with us in the first debut episode of the tech garages, now we are at the thirteenth episode and therefore it is a welcome return. He is an intrapreneur, growth expert, now very vertical also on AI, right now, right Luca? Yes, yes, yes, actually thanks for the double invitation, happy to be back. Very good, then I would say that all we have to do is raise the garage door and leave. Which is Wall Street has come out with a new acronym called Mangos, which sounds a bit like a Netflix series, it sounds like something. In reality, it is this label by which the new successful players on Wall Street are identified. So basically M stands for meta by Mark Zuckerberg, A stands for Antropic by Daria Modei, N for Nvidia by Jensen Wang, G for Google by Sundae Pichar, and then there is the O of OpenAI by Sam Altman and the S, now at this moment Super Hype by SpaceX and Elon Musk. This is the new acronym for which Wall Street indicates the most successful players, the ones to keep an eye on. I thought this was very nice because there were all these images of mangoes with the faces of all these protagonists. I wanted to point out that, however, compared to the old acronym which was the famous FANG, when there was still Facebook, it was written with an F, there are a couple of excellent exits from this Wall Street basket. One is that of Netflix and this can be quite understandable, but there is also that of Apple and Amazon, which are a little heavier. This doesn't change anything in the sense of stock market value, it's simply an index, a way in which you look at the companies of the future with greater interest. Among other things, on Amazon I think that Jeff Bezos didn't take it very well for the fact that Antropic stole the AA of this mango from him, because then we'll see later, we'll link to another piece of news later which in my opinion is quite interesting from this point of view. I wanted to start with a technical thing already, because I saw Vincenzo that you released, just in recent days, a test that you did on this new agent called Hermes. Yes. So it is a new agent which, let's say the two words correctly, is an open agent, therefore, and can also be installed on the PC. I'll connect to this a bit, well, maybe if you talk to us about it briefly, then we'll think a bit about which agents the three of us are using most at the moment. Yes, yes, yes. Let's say that Hermes Agent is a framework for using agents, which is a bit of an answer to OpenClaw. Where OpenClaw was somewhat of the progenitor of these latest generation agent frameworks, but where OpenClaw is a bit more of a geek, of a technical developer, this Hermes Agent also has a desktop application, currently released for macOS, which makes, let's say, the installation very simple, because it is an application for Mac, it must be installed, and from there you have a graphical interface and you can start communicating with these people, then add other tools and then make it work. They're talking about it a lot in the United States, so I decided to do a test, a trial, and it's not bad, I have to say it has great, great potential. I would add that there is news in recent days that Cloud is limiting the flat plans for everything concerning the agency part, so essentially the 20 euro subscription will only become suitable for chats, so very little will be able to be done at the agency level and it is moving a bit towards the consumption part, therefore usage and consumption. So the reckoning is starting to arrive a bit, which we have been talking about for a long time with garage owners, that sooner or later once we get used to using it on a daily basis, then little by little the prices will probably tend to rise. Among other things, I give a small pill because there is the issue of excess use, so they then lead you to use an extra token. But I wanted to remind everyone, I don't know if everyone knows, that for example when I went to see my surplus it had a cap of 10,000 dollars, so you also have to be careful from this point of view, go inside, maybe go and modify this surplus limit based on your portfolio, because maybe someone has a 100 dollar plan, but then they have a 10,000 limit, an important bill could arrive from this point of view. Let's talk about agents for a moment, how we are perhaps using them at the moment, as a sample of course, then everyone uses them as they want. Luca, how are you using the agentic part of AI now? Yes, look, a few months ago, at the beginning of this year anyway, just before there was all the chaos, the Pentagon news, etc., I actually made a move from the GPT OpenAI stack, closed the account, removed the license, and went to Cloud, so I started messing around with Cowork, which let's say the part I use is Cowork, if I need a simpler routine, or Cloud Code, that's the part I use. And in general the most useful things are the use cases, on the marketing side, so there is research, scraping, synthesis, to then propose ideas, insights, speed up the search for information, or create semi-finished drafts, contents, strategies, so a bit of that front there. Which then is something that, let's say, I on the one hand out of necessity, in the sense that it is a non-technical background, on the other hand, I try for an almost approach, let's say that in any case it always puts the method and the process at the center, and then the tools afterwards. What I realize, even when talking to companies, myself included, is that in reality agents are something very powerful, but nine times out of ten you don't need a nuclear bomb if you have to open a nut. And most of the time, I have to open nuts, so I have to do things perhaps more efficiently, faster or in a smarter way. For example, instead of consulting 5 sources, as I might do when reading articles, I consult 500 at the same time. But the operation is relatively simple, the workflow is relatively simple, so certainly, let's say, I don't feel like I'm working magic with the agents, as I do with the reels on Instagram of those who build the animated visualization of the second brain, which is as beautiful as it is useless, that is, an artistic project, beautiful, I like it very much, I would never use it in real life during 8am, I would use it later, when the working hours are over, I would use it in the evening. But yes, let's say, my stack is the Cloud one and above all they are still marketing use cases, so, in reality, perhaps I don't even need the most advanced capabilities at the moment. Yes, yes, me, the use cases are more or less the same as Luca and therefore range from more in-depth searches up to the creation of semi-finished products for content. As a stack, let's say that, then, I also do a lot of consultancy work, including training, I necessarily have to test them all a bit, so I don't have a main stack, and therefore, let's say, preferential at this moment, I must say that I really like Codex by OpenAI, they have made excellent steps forward, both as, let's say, user interface which makes the tool less like a coder and more like a knowledge worker and so I'm using that a lot, then obviously for some time now I've been using a lot of co-work, cloud co-work or queues, depending on your needs, and obviously also Gemini. I'm a beta tester of Gemini, so I necessarily try out all the things that haven't been seen yet or haven't arrived in Italy yet and so I've now seen, I also made a video on this, the new Gemini Spark as a system, let's say, co-work type agent, which however is still far behind co-work, so it works above all on, let's say, applications from the Google ecosystem, and then they told us in Mountain View, there will also be the possibility of connecting it to MCP servers. In this regard, Google is much slower and more cautious than the new players who have to go faster, because they need to gain market shares, in short, quickly, also aimed at listing on the stock exchange. so I try a bit of everything also to understand which things are best, the strengths and weaknesses of each, but now we've reached a level where it's really difficult to say one is better than the other, they chase each other, one day one is better, the next day the other is better, and in short it's hard. At a certain point the normal user will have to choose, let's say, we who work there can perhaps afford to use more than one, but the normal user obviously has a theme of choice, it is precisely an economic theme, in short. And yes, in fact it is that, once one starts working on a platform and then it becomes all about the switch cost, it becomes complicated. I followed the path that we all followed a bit, from OpenAI, I moved to Antropic, then from ChatGPT I moved to Cloud, and I kept the ChatGPT subscription the same because I use the graphics part a lot, therefore the images part, because in my opinion they are still a step ahead, at least from this point of view. I have GPTs that create all my covers, for example content on YouTube that I struggle to recreate elsewhere. While on an agent level, I must say that I am very passionate about the whole Cloud Code discussion. I'm looking partly because I would really like to do a project, partly because I want to really understand what the limitations of this Vibe Coding are, in fact how effectively a project developed in this way can actually be implemented on the market, how many holes there are, if more time is wasted patching up the holes than designing. So I'm working a lot on it, I still have a bit of difficulty, a bit of revulsion, awe at using the terminal directly, so I started using, I used Visual Studio Code for editing, which is the one from Microsoft. Now I've switched to Antigravity and I have to say I'm quite happy with it, so I use it with Antigravity, I use Code. I'm doing quite interesting projects, that's it, so let's say that... Yes. I quote you, I quote you, another platform that I use a lot specifically for the creation of automations and agents, much more affordable, user-friendly than N8N, which has been talked about so much, which is called Gumloop, it is an American platform and it is very convenient because the subscription from 20 dollars to 30 dollars a month also includes the consumption of the various LLMs that we add as the brain of the agent or of the tools that we add. So there is no need to do the configurations like on the N8N, but everything is, let's say, already pre-configured. You use, let's say, not your credentials, but, let's say, they give them to you, you bypass, in short, all the difficulty of setting up an account perhaps with OpenAI, with Gemini, therefore the API keys, etc. So Gumloop is very interesting for creating both a node and agent workflow. Absolutely, it's a must try, yes, yes, it's very interesting. We cannot fail to talk today about the news of recent days, however, of the fact that Fable 5, Mythos 5, let's say, have been blocked by the American government due to security problems. It connects me to the discussion before the Mangos, that James Bezos was probably disappointed, that there, instead of Amazon, it ended up, in Antropic, because it seems that the tip-off about these jailbreak problems, of insecurities, came from Amazon itself. Here, by the way, is also very strange, because Amazon has invested in Antropic, so here we are in a world of parallel universes that are really difficult to fully understand. There is certainly also the position he had in Modena towards the government, so someone spoke out of spite, someone was therefore blocked. Dario Amodei himself, some say, cried wolf, wolf, wolf for a long time, we have a model so powerful that it could be dangerous and then in the end it was blocked. So there are many keys to understanding from this point of view, not least the fact that Antropic is part of that group of IPOs that are going on the market now, so even that is a small unknown to understand what will happen. Luca, how do you see this whole process? Are we facing a new dawn of much tighter regulation after an initial period in which everyone was a bit free? In general, perhaps it's something I said last time too, my approach towards default technology and therefore also AI is that of being an anarchist technologist, that is, free market and long live technology etc. However, in the case of artificial intelligence it is truly the only case in which I felt like regulation, that is, I am talking about before these events and in general it is one of those cases in which it is not, you know how to say no? AI is just a tool but it is not exactly the same as a hammer or a pair of glasses because for example for better or for worse because for example you can't ask a pair of glasses how I use you and AI you can ask it and it will even answer you and help you etc. etc. so it is something a little different and it is objectively potentially dangerous not fair for the people and for the global economy to let things go as they are because as much as a technology is the free market things will go towards abundance anyway and we will all be happy and happy it would be a bit like they say it comes to mind in Italian naive naive perhaps a little naive but at the same time this type of move going into specifics is difficult a bit like you were saying not to also see it as a tension of relations and geopolitics within America to the relations that exist between these companies and between the government the somewhat I don't know not choleric bipolar trend of this government in particular therefore it is difficult not to go towards that key of interpretation so let's say that of all the forms of regulation of which perhaps I felt the need for certainly this is not one of those because it almost seems to me yes if not it is a spite but still something very poorly reasoned and in fact let's say it is coherent like many things that America is doing which first gives someone a slap and then later invents the reason why he did it by improvising who knows what so I don't see a real reasoning in it I don't see a real vision of regulation as instead Europe is doing that at least in one thing we are getting a little more organised, that is the regulation between the other one I'll add something that is very nice that I heard that if it seems like they gave him 90 minutes to block access to all non-American people it's really ridiculous, right? Vincenzo what do you think? Yes but in fact it is certainly not the way to proceed if you have the security of the nation at heart let's say if someone perhaps at Amazon had identified a possible jailbreak they would certainly have had to notify Antropic maybe the government but in any case then the government should have notified Antropic to say before let's say let's burst the hole let's close this possible leak if it really was this then the way to proceed from a security perspective is not really that of saying block it doesn't make much sense certainly a bit let's say Modei was asking for it in the sense that this attack allupo it wasn't a great move so at a certain point you found someone who really believed in this alarm and then in short with the tools it has which are not exactly the most refined ones the most refined ones it certainly blocked you for issues such as political Antropic doesn't support this government it didn't finance it Open AI instead yes at the hands of its president in short there are certainly political motivations it's certainly a hard blow to Antropic's IPO and to its business anyway because Fable was supposed to be the model of peak and this is now, in short, it is no longer, let's say, a source of revenue, especially with a view to what you were talking about at the beginning of moving the business model to consumption and therefore making us pay for artificial intelligence like electricity with a consumption bill. I'll add one thing, some quick news because it seems that Mistral is looking for 3 billion, therefore negotiations, negotiations for and having a valuation of around 20 billion, given that the numbers are merciless compared to the numbers that are circulating in America, therefore at the level of investments but in your opinion there is still the possibility for Europe to remain in the game of generative AI or an attempt beyond the time limit now Luca but in general I think that in Europe perhaps I am too close to the Italian average but I will be he had to say we are we are late we have missed the train of digital transformation let alone AI that is we still have to explain to people what social media are what whatsapp is what whatsapp is they have all understood what social media are what online advertising is how these markets work how this world works in reality let's say we are still at a level of having to distribute and democratize a basic awareness that above all affects the public because rightly so if the public the public understood as governments, politics as well as people do not understand it is obvious that they do not invest it is obvious that they do not consider investing in AI infrastructure to be a socially relevant issue for a country, but instead talking about 600 other things from the highest sacrosanct to the most stupid of the small population who allow themselves to be distracted by what color what the name of that the scandals of the politician in question etc. etc. so I am still quite hopeful among other things I have known Mistral for a long time because two years ago I did the first local AI experiment with open source tools and the model I use taken from Gameface which it ran on my gaming computer on the Nvidia card it was actually one of the Mistral models and therefore in general I am still hopeful therefore I still believe in the possibility that a community could be born between private companies public movements which says given also the global geopolitical situation in which we find ourselves whereby Europe thought it thought it had this great best friend which was America and instead America said but whoever knows you I think of mine you think of yours so more than anything there is the risk that they will turn off the tap and we in fact we are out of the game, that's right, so I'm still hopeful that geopolitical pressure can move people from companies to governments to make projects from this perspective for a matter of survival. strategic because we are really talking about this saying, as he rightly says, the problem is precisely the infrastructure, that is, there is nothing comparable on which to run these things at the data center level at the cloud level, all of our own, understood as in Europe and the other thing is also understanding who enters into Mistral because if these 3 billion come from the Americans we are finished we have closed the circle well Microsoft is already in it so exactly exactly yes but I see Mistral in short a bit in the balance in the sense that it should be able to arrive at models truly frontier but beyond this perhaps it can only be saved if some national government imposes its use as has often been done for open source software in public administration but there are also other interesting subjects in Europe for example there is a laboratory called Black Forest Labs which makes a very advanced model called Flux a diffusion model therefore a model for generation they are Germans and they play it well they also have these models open weight so they are doing a good job on large language models in my opinion there is no match i.e. we Europeans should stay there more on the application field therefore also on robotics for example also yes applicative let's say in various senses therefore also agentive therefore I create agents specialized in the healthcare sector or other sectors here is the application field in my opinion a game like European companies can be played on the lowest part of the models I see it as difficult a question that I already asked also in the last episode of the techno garages because now that we have available for example cloud code I who have always had a SAS platform that manages sales along the distribution chains of manufacturing companies therefore I have always considered myself a decent analyst from this point of view but I have never been a developer now that I have cloud code in my hands I have the feeling of being different nted a demigod in fact then I don't know how true or not this feeling is so on the one hand you say whatever you want to do it seems like you can do it that it's there ready to be reached you just have to stretch out your hand and you can reach it on the other but I was also saying in the last episode everyone else has these conditions too so on the one hand the competition has grown disproportionately we also see the start ups which by now must have something other than MVP they must have already validated models they must already have traction so the competition is getting higher and higher and at the same time time these models suddenly change everything that is the scenario because what perhaps was going well today we are doing a project today tomorrow is already outdated so I ask you is it the era of opportunity or the era of illusions Luca I would tell you that it is the era of trust as it has never been in the sense that these are my take as a person who comes from the world of startup innovation and who therefore has been practicing the lean method for a decade technical complexity has never been the bottleneck of failure it has always been the market it has always been who pays you but there you have validated this thing you have verified that it matches with a real need of real people and companies so paradoxically the fact that we are all semi-gods of development does not solve the problem it accelerates it because before only the developers could make the mistake of saying I have an idea I invest three months in doing it and then I launch it nothing happens failure spent three months in nothingness now we can all make this mistake making Nvidia very happy and all the players on duty who sell us the shovels sell us the shovels to go and look for them and we believe we having solved the problem of life when in reality the problem of life has always been to have the customer know the market empathize with people truly solve group community problems in order to exchange value between turn it into wealth on both sides and therefore this is something that has dazzled a lot of people with vibe coding and in reality it is creating many more companies projects that fail so there are memes about here is the 38th tool created with vibe coding that no one will ever use and there is a reason this is happening and therefore on the other hand what is true however is that let's say the time from idea to roll out to the market is becoming zero because there are even those who heard an interview with an internal Antropic which said not only do I run or run different cloud instances in parallel but now I don't even do prompting anymore I have created a system that understands that I give it an objective and it understands what needs to be done to get there and passes the prompt to the other instance which then develops creates and then there is another one which revision and so on so the next step will be to design loops not design the prompt the request or the workflow and this however is true this is happening and so where is the difference if everyone can have a thought send a voice message to cloud code and cloud code the trust of the people does all the work the ability to create a community because we all have the idea we find them on the ground and we are all about to become capable of realizing them with a negligible monetary investment and therefore the difference is that when I launch something they trust something because they know me because there is a bond there is a pre-existing community and therefore in reality it is our game as those who do growth the real challenge shifts once again and even more so to distribution i.e. ok now there is the product and therefore zero users zero likes zero visits that we make that is the real challenge with a market that will be increasingly crowded with this stuff here so yes it will be increasingly more difficult and the lever for me will be trust that is the real currency of the era we are entering is trust and it also applies to LLMs nice that mistra the most powerful but I don't trust I know Claude I stay with Claude with Claude I feel calm I know the screens I don't want to change because I trust this one I don't trust the other one who doesn't have the constitutional AI which is a new brand well so in my opinion paradoxically everything will move to an even more human and empathetic dynamic of trust yes I agree with this I will however add another complementary thing which for me is also let's say the era of opportunities especially for those like us who do not have the ability to coding skills because then it is not necessarily not necessary that you have to create a product for the outside for the masses but I imagine you too have created products for yourselves that perhaps make your life easier. I have created a small software for you to follow the entire lead cycle or to make commercial proposals or to create carousels with a certain brand etc. in short, each of us today can create, with a little investment of time, personal software which perhaps even saves them the money that would be needed instead of buying this type of software on the market which obviously won't be a perfect software which will perhaps have some problems it will have to be maintained a little etc. but today in short all these things are done even more easily so from this point of view I believe it creates many opportunities for individuals especially for micro companies also for freelancers Luca you talked about envy of Jensen Wang and the fact of selling pickaxes just a few words made me think that he made the presentation where he presented this new super chip which therefore seems to want to also enter the hardware market so practically those who sold pickaxes now also go to the mine to dig, right? It seems like there are various theories on this too, someone says now the value of envy is so high that they must also find other ways to support this value and they will collaborate with the main computer manufacturers Lenovo Dell and a very high-end computer under the envy brand will probably also be released, this is also because there is the idea of ​​having AI on our personal computer on our computer and therefore there is also a discussion linked to these mega centers that are being created, right? to make AI run even the population itself, especially in America there are many protests from this point of view if they create a jig a jig nearby factors people are very worried and think that it is just a marketing move that of Jenson Wang who among other things is very good from this point of view he has now become with his leather jacket now it seems to me that he has taken the place of Steve Jobs by now so he has become very good and are they a marketing operation or is there a possible future in which AI will be actually hosted on our computers as far as we can know we of course do a bit of fantasy AI in general he too I feed a lot on memes about AI they are my favorite things he was good as you said he was very good at marketing in the sense that to make the comparison I recently saw an image that said as if the CEO of Oreos says in my opinion people should eat 50-100 Oreos a day yes this is what should be done thanks it's your company so it succeeded to pass on these messages from the series even expressing themselves on the tokens like ah if a developer doesn't consume at least these tokens a day he's an idiot the reseller is working behind them you're the one who makes the trillions but so how much are these statements worth but in the hype many have gone after the red light hasn't passed and having said that in my opinion it's definitely another area of interest for them maybe there's also let's say the vision of trying to understand how to expand horizontally also because in what they're doing they're already very far ahead I'd say established monopolists therefore it could be a way to understand other ways of entering other markets where there is not a strong precession of envy but they have all the prerequisites to have credibility on these others on these other fronts and in general I think that for all of these players the idea of inserting themselves into the hardware part is also a way of perhaps anticipating or undermining this whole movement that we were also talking about at the beginning of those who instead fiddle with it in an open source manner in a local way. I see it a bit like the fact that today I can't buy a phone anywhere where Facebook isn't already installed and therefore where my privacy permissions are disgraced by default at the moment in which I buy the phone and when at the hardware level they start to put models on you, it's not that you can turn off the operating system, it's not that you can but to say this, no, so it's as if they took away from you the choice of giving up your consumption of data and information and at that point it becomes really difficult to acquire hardware that isn't completely agnostic and atheist and clean and you decide what to play with whether to play with it etc. so in general let's say I wouldn't be surprised by this type of discussion from others because it would be a way of saying let me catch even those who want to be hackers who live in the woods at all costs and who don't use subscriptions but do everything locally, that is, they'd rather buy seven computer to make a slightly more powerful card to go around the place instead of using OpenAI I don't want to cheat them and I necessarily want to give them the paid model of my brand of my company so I see it a bit as a movement perhaps as the beginning of a movement of the Vincenzi kind in my opinion it's a move more than more than communication it's a truly industrial move of integration across the entire supply chain and therefore also of capturing a market that already exists i.e. the market of which today is the market for Apple's high-end desktops that market of people who need a super powerful computer at home, not necessarily to run AI stuff on it but still a super powerful computer, so there is an attempt to enter that niche and undermine Apple which perhaps from this point of view is a bit late with let's say a solution that at the moment leverages above all on the manufacturers who are already present on the market and on the Microsoft operating system and then in the future they already have a very powerful mini PC but there the pill is missing towards the consumer he has no idea of envy that is leaving aside us who are a bit of the interested sector the average consumer knows he doesn't know envy he knows Microsoft starface he knows Lenovo and the others and therefore that market in my opinion is a profitable market to control and then if we really go we will go towards that path that we said at the beginning where the models will be paid according to consumption the bills could become quite substantial and therefore at that point having a very powerful PC at home on which I run some local models could be the solution also think about when there was the launch of open claw the mac minis were sold like hotcakes and could no longer be found so in my opinion this market exists and it is good for Wang to preside over it among other things on the bill it came to mind sorry I don't remember the source maybe a hoax but I had read about a company that had left the area free IT for a year without giving him limits on using the clouds and at the end of the year he received the bill and it was 500 million yes but in Uber they had to put a ceiling of 1500 dollars per month per developer because surely there is the issue among other things there is also the issue where this cost is allocated on a service cost on a cost linked to the employee to the cost of labor therefore others are also triggered because in fact the employee becomes an enhanced employee but it is a whole series of discussions that are new for everyone so here it is we need to understand a little I conclude Jensen Wang's speech because it was very interesting he was very good because in the last 20 years he brought envy to being the most capitalized company in the world he was very good he was also lucky because he found the right moment at the crossroads ada when the right train passed with its cips which were created more for the world of videogames but then they were adapted to the whole AI discourse but it's nice because I was reading Alessandro Aresu's book China has won which among other things will be a guest in one of the next interviews that will be released on the Garage which said that Jensen Wang is Chinese in China Taiwanese in Taiwan and American in America so he has a very, very chameleon-like ability to be everything to everyone here we are talking about people with extraordinary ability and even with a big ego we cannot fail to mention Elon Musk who made one of the largest IPOs ever he launched SpaceX among other things we must remember that within SpaceX there is not only the company that launches rockets but there is also the whole issue of AI ex AI there is Starlink there is Grok so it is a company that has many elements within it which is said but Musk is selling the future on the stock exchange among other things I went to see it out of curiosity this morning before recording the episode it had been listed on 135 dollars I wanted to understand if after the first day it had a 20% instead it remained quite high because now it is worth 202 dollars so let's say that those who have invested up to now have brought home an interesting return those who have joined this IPO among other things are news that has gone a bit unnoticed and within this is the IPO price there is a 60 billion option to buy Cursor so in fact they bought it otherwise they would lose 10 billion so in fact it is already therefore also from this point of view Musk is also going after everything that is the Vibe Coding discussion here is how this IPO naturally has to be put together as we were saying before that of Antropic and also that of OpenAI and all three together reach almost 4 trillion so here we are at 3.70 trillion no of valuation it is something that has probably never happened before here in these in these terms here is how you see these this season of In such gigantic IPOs there really is the technology underneath there is a bit of value however it is higher than what it should actually be without talking about a bubble however the technology is actually there anyway here it is not a thing rather a small anecdote I went to a press launch with tech journalists and there was a journalist who was a bit more differently young who said but I haven't seen many of these things for me it will be one of the many passing things that will lead nowhere just wait here but you won't be so sure that's why life is changing a lot for me but here's what you think Luca in general I think that some figures are completely meaningless i.e. at the level of change impact transformation it is objective and indeed in some ways we are still at the beginning what at least to me makes it scream bubble or in any case it seems like it is when you compare these figures with for example the GDP of a country you have to say how is it possible that the entire economic output or this entire sector etc. is less than this thing how is it possible that if if you add them up I don't know the market of maybe relatives a to the most basic but more fundamental needs than having access to cagpt to AIs such as things to eat you don't get to these numbers so it's more worth using AIs than not eating they are not meaningless numbers in fact never seen no and therefore part of the hype of riding the wave then speaking of the on musk he has always been a master in manipulating expectations if we then talk about stock market prices because then when you go public on the stock exchange it is very much a game of expectations much less a game of turnover and objective metrics a thing of perception the only one for better or for worse he has always been good at moving in those in those logics not then I am not an expert in the IPO part of the stock market but I have read that yes that in short this beginning these days I have gone very well but it is also true that this type of listing with a kind of conglomerate of mix of companies is very particular historically these are not things that go well so we will see in a year or two years where they are at because in the hype Elon Musk is a master I would like to say Greg thank you that it went well knowing that he is very good at this let's see but it must be said that even if perhaps there is this strange negative side of the conglomerate it is also true that it is as if he had listed a closed loop because there is everything in there there is the infrastructure there is the digital part there is the distribution and also in relation to the other players who have listed themselves or to the one we have in Europe this concept has an edge so looking at it it could be a risky thing which then dismantles the hype dead or it could be a model of a group of AI companies which however does not only list the front end product let's say but also prices the infrastructure i.e. it is self-sufficient in a certain sense as a value chain seen in this way and so let's see it will be interesting also because actually he also has Starlink the entire satellite network However, SpaceX has launched these carriers into space so he actually also has something physical beyond the AI issue that Antropic and Open AI have, here you have AI alone, exactly, he certainly forced his hand a bit by creating this conglomerate, he wanted to put XAI inside at all costs and let's say some of the investors were not so in agreement and therefore with the entry into the stock market he actually wants to finance above all the AI part Musk absolutely wants to defeat Sam Altman so at all costs he is trying to build an AI company that you surpass Open AI in model computation infrastructures now even the purchase of cursors can certainly be read from this perspective then there is certainly the fact that having a frontier AI company you can certainly use it to optimize and improve all SpaceX's Starlink activities, this is so we definitely need an operation that is also useful industrially to strengthen what you physically have, however, in short, having seen Musk's moves in recent years, I also see it very much, let's say a move against Altman, in short, we'll see how it ends up, it's difficult for us to succeed because in short, Grock, there is a market that is currently so significant, indeed truly residual compared to the others, to conclude this part linked to these large IPOs we cannot fail to mention that Bending Spoons, so finally an Italian company has submitted the IPO application to the Nasdaq and it seems that we are talking about a valuation around 20 billion let's quote it because there are not many Italian companies that obtain these results even if they have a somewhat particular business so it would also be a bit to explain here from this point of view how you see it Bending Spoons quickly Luca and well as you were saying it's a it's a unicorn but in the sense not in the startup sense even if perhaps it becomes one if it makes this if it makes this listing but in the sense that it's really a very particular model that they managed to assemble and there are no comparable ones I would always say in Italy I say that they follow the same mix of tech companies but also a venture because they acquire and then they restructure they relaunch they manage so there is a logic also from portfolio but also from acquisition so it is truly a very particular thing so let's say on the one hand I say well done, hats off to you on the other I would say not so much to them but to those who mention them so often that it is difficult to say let's do like Benny Spoons because there is one from Benny Spoons in Italy therefore of things that you can really copy and apply to yourself and I don't know how much you can follow that model because it's really a white fly and good, really good yes yes no I don't have much to say either I agree with what Luca says they are certainly an appreciable company for the course that they have followed, it is also true that in these restructuring processes many users are, let's say, dissatisfied with the end of the applications they have acquired and then they have also restructured, often the restructurings are then accompanied by an increase in prices and subscriptions for users so in short there are various sides to consider I would like to quote a reflection which in my opinion is very interesting which I heard from Riccardo Out on the Actually podcast who said that now with the discussion that we have all become developers it is not so much the question of the app they acquire but that they say they are playing the game of distribution therefore of having lots of users to whom they can then possibly sell to them at a later time also sell them something else so this idea of distribution seems to me to be an interesting idea here to reflect on also to make people understand the potential value that Bending Spuns has here from this point of view therefore not only linked to the software but also linked to the people who are inside these packages here is about this software and I wanted to talk a bit about Apple about the chronic delay that is always there I have been an Apple user for 20 years so I I can even allow myself to speak badly of it every now and then so that's when the whole discussion about Apple Intelligence I found a bit cloying with the delay they have I have always criticized the European Union a lot when it blocks functions that we don't have and therefore we are always here late compared to the rest of the world in having some features so I have always been very critical from this point of view now Apple but even before that Apple Intelligence is sensationally late it will be released in the United States in September it already seems that it is only in English it is not available in other languages but it is already protesting that he can't spread it in the European Union because there is this blockade so in this case it really seemed a bit cloying to me here is Apple's controversy from this point of view given the chronic delay that in the end it had to rely on Google or in any case it will always do the usual even here we are talking about distribution they in fact hold the distribution of what the various products are because in any case they all pass by what is the iPhone I wonder and they are always there sitting there on the river waiting for the others to pass by they ask them for the donation to go on their product and that's fine so I wonder now at this moment if a moment will then come where perhaps a company like Anthropic a company like OpenAI, I'm not saying a wearable but also something that doesn't come out with something that actually replaces the iPhone upstream and therefore upstream which in the end is this river that doesn't divert the course of the river and sooner or later they find themselves having a small stream instead of this river that they've had up until now, how do you see Apple's position? I have always been an Androidian, not Apple for family values in the sense that my father works in IT and passed on to me the hatred for closed systems so almost ideologically Apple doesn't dismantle things modding hello Apple goodbye forever so therefore I didn't not follow in detail before I don't follow much now either in general but as a company as a phenomenon in the meantime already as you said at the beginning the paradigm shift apart from the listing on the stock exchange but in general in tech we are dealing with an era in which we were talking about the big four and Apple was there with these mega players so Open AI Antropic is added to the big four and all of them this AI ecosystem what actually remains a little behind evidently in an almost embarrassing way is Apple itself which seems to have no response to AI actually I remember that especially in these years where every day a news story comes out of Google has done this Microsoft has done this and Apple has the new colored case after which Open AI has done this Antropic has done this and Apple is doing a project on smart cars but what the fuck does it have to do with it It's possible that the topics of the meetings that the Apple executives are having are on the product, the case, the colors and the smart car and the iPhone camera so let's say that as a non-Apple fan I don't suffer from it personally because I don't use Vice Apple anyway and I don't see myself using it in the near future but I find it objectively very strange there's one thing that in the tech scene you notice and ask yourself what it is that is, this is the thing it's difficult to imagine that there is a strategic thought a secret bi plan that will surprise us all in five years' time it's difficult It's easier to imagine that they've missed the boat or have leadership problems so they haven't yet found a way to organize themselves and give a strong response to the topic of AI. However, not knowing it so well, maybe you have a different vision, they remain of the very strong hardware producers that's why computers are super valid and in the end then the iPhone here from this point but in fact let's say that their focus is on hardware especially consumer hardware so from their point of view let's say that artificial intelligence phenomenon has little impact on the demand of the average consumer who buys the iPhone if not perhaps certainly it would be desirable to have a Siri that responds to you in a way that is certainly closer to what ChatGPT Gemini does now etc. so that's why they probably haven't considered it that much so they're behind the topic is precisely, see if someone doesn't come up with how to make the consumer really perceive the need for AI in the phone or computer, as we say AI is a killer application it still isn't because let's say so I can live even without having AI in my Android and yet now I couldn't do it at least that is because I see the huge difference precisely with Siri so however this thing that I perceive because I am a power user the average user doesn't perceive it therefore he doesn't feel the need to have the intelligent assistant in the smartphone or computer but as you were probably saying the other risk could be not only from Google which perhaps will make an assistant so powerful as to make it a killer application the other theme is that of new types of device hardware which perhaps are wearable which make let's say the smartphone a little replaceable I won't say obsolete but replaceable and here we obviously know that OpenAI is working on it Meta has been working on it for some time it has validated the idea of the glasses they must certainly become more powerful yet they are still like saying a companion to the smartphone they are not let's say autonomous so in that area they could certainly be taken let's say taken aback even though they also seem to be working on smart glasses so it's still a love brand and hardly barefoot even if you miss the train of some phenomenon sooner or later I'll arrive the thing that I found a bit strange listening to the interviews of the various managers is this thing that it wasn't clear how they used the Gemini technology for their models because in short it's not that they use Gemini they used to create design improve their linguistic model so it wasn't clear what kind of use they made of it on the one hand maybe they can't they don't want to say that they relied too much on Gemini we don't know in short but surely, being behind, they are trying to catch up also in the area of linguistic models, but surely as you were saying, being a strong brand, they certainly have the famous trust, so your market community has very loyal customers, you don't lose that easily unless you make exaggerated blunders, but if you miss a train you really don't lose the trust of the users, I'm the first one who would buy my MacBook Pro again, so tomorrow if I had to change it, so I'm the first, God forbid, yes, but in fact also on the computer side, Microsoft has tried there too, but not has found the solution to how to use AI to improve the way in which a PC is used but someone will come and do it for example I also saw some demos of use of the latest model of OpenAI GPT real time which allows you to use the voice and therefore also allows you to use it to use the computer and then you tell it open the application do this do that so that is certainly a possible use new use new interface to use computers without perhaps the keyboard who will first be able to understand how to replace keyboards with mice artificial intelligence will probably conquer a large share of the market also because if we have a real assistant in which we can ask it to do things even the whole world of apps p could suddenly become obsolete, this is what we are used to, to the delight of Luca, I wanted to conclude by giving some news because Gcom has just opened the first investigation based on Digital Market Hutt on Apple precisely because it seems that the iPhone does not allow the transition from those who use the iPhone to Android in a complete way so it allows you to move only some elements but not all here is this resistance that is Apple in making you perceive more the switch cost under investigation so we will see therefore the first investigation of this Digital Market Hutt went against Apple for this reason here so it goes a bit to underline what you said before Luca oh yes yes well we've already passed the hour so I'd say it flew by it seemed like a super interesting episode in my opinion I hope you tell us too whoever sees us on YouTube who listens to us on Spotify what do you think a big thank you to Vincenzo Cosenza and Luca and Luca Barboni for taking part and if you're available then maybe after the summer I'll invite you again if you want to take part again I'm happy to say thank you hello everyone hello everyone so no we just have to lower the garage shutter bye