- il y a 2 jours
Découvrez le code qui vaut 9 milliards et comment vous pouvez l'utiliser avec les outils que vous avez déjà. Dans cette vidéo, nous allons explorer les capacités de Claude Cowork, une plateforme d'automation basée sur l'intelligence artificielle qui peut révolutionner votre façon de travailler. Nous allons vous montrer comment utiliser Claude Cowork pour automatiser des tâches répétitives, créer des workflows personnalisés et augmenter votre productivité. Vous apprendrez également comment utiliser les plugins de Claude Cowork pour étendre ses capacités et comment utiliser Perplexity Computer pour prendre des décisions éclairées. Nous allons également discuter des avantages de l'utilisation de Lovable AI et de Cursor AI pour améliorer vos processus de travail. Si vous êtes intéressé par l'automatisation et l'intelligence artificielle, cette vidéo est faite pour vous. Alors, découvrez comment utiliser Claude Cowork et ses outils associés pour améliorer votre productivité et votre efficacité.
🚀 N°1 FORMATION IA & BUSINESS
👉 https://parlonsia.teachizy.fr
📩 Email : formation.ai87@gmail.com
🧠 Ce que je propose
Formations IA pour automatiser ton business
Création et mise en place d’agents IA pour ton activité
Accompagnement pour utiliser ChatGPT, Gemini & co sur tes vrais dossiers
🌐 Site & contenus
Site & formations : https://parlonsia.teachizy.fr
Chaîne YouTube : https://www.youtube.com/@Parlons.IA.2
Chaîne Dailymotion : https://dailymotion.com/formation.ai87
Podcast : https://spoti.fi/4dqZ3uO
Blog : https://medium.com/@flma1349/
📲 Réseaux & contact
X / Twitter : https://x.com/ParlonsIAx
TikTok : https://www.tiktok.com/@parlonsia
🧪 Outils IA à tester
Coder Agent IA : https://bit.ly/Coder_agentiA
Short AI : http://bit.ly/4lzE782
SEO Agent IA : https://urlr.me/P8AS5N
— 🎁 -25 % avec le code PARLONSIA25
Découvrez le code qui vaut 9 milliards et comment vous pouvez l'utiliser avec les outils que vous avez déjà. Dans cette vidéo, nous allons explorer les possibilités offertes par les agents d'intelligence artificielle, tels que Claude Cowork, et comment ils peuvent être utilisés pour automatiser des tâches et améliorer la productivité. Nous allons également discuter des plugins et des extensions disponibles pour Claude Cowork, ainsi que des cas d'utilisation pour lesquels il peut être utilisé. Si vous êtes intéressé par l'apprentissage de l'utilisation de Claude Cowork et de ses capacités, cette vidéo est pour vous. Nous allons vous montrer comment utiliser Claude Cowork pour créer des tâches automatisées, gérer des projets et même utiliser des curseurs pour interagir avec votre ordinateur. Nous allons également aborder les concepts de perplexité et de bolt, ainsi que leur relation avec l'intelligence artificielle. Rejoignez-nous pour découvrir comment vous pouvez utiliser le code qui vaut 9 milliards pour améliorer votre productivité et votre workflow avec l'aide de l'agent d'intelligence artificielle lovable, Claude Cowork.
🚀 N°1 FORMATION IA & BUSINESS
👉 https://parlonsia.teachizy.fr
📩 Email : formation.ai87@gmail.com
🧠 Ce que je propose
Formations IA pour automatiser ton business
Création et mise en place d’agents IA pour ton activité
Accompagnement pour utiliser ChatGPT, Gemini & co sur tes vrais dossiers
🌐 Site & contenus
Site & formations : https://parlonsia.teachizy.fr
Chaîne YouTube : https://www.youtube.com/@Parlons.IA.2
Chaîne Dailymotion : https://dailymotion.com/formation.ai87
Podcast : https://spoti.fi/4dqZ3uO
Blog : https://medium.com/@flma1349/
📲 Réseaux & contact
X / Twitter : https://x.com/ParlonsIAx
TikTok : https://www.tiktok.com/@parlonsia
🧪 Outils IA à tester
Coder Agent IA : https://bit.ly/Coder_agentiA
Short AI : http://bit.ly/4lzE782
SEO Agent IA : https://urlr.me/P8AS5N
— 🎁 -25 % avec le code PARLONSIA25
Découvrez le code qui vaut 9 milliards et comment vous pouvez l'utiliser avec les outils que vous avez déjà. Dans cette vidéo, nous allons explorer les possibilités offertes par les agents d'intelligence artificielle, tels que Claude Cowork, et comment ils peuvent être utilisés pour automatiser des tâches et améliorer la productivité. Nous allons également discuter des plugins et des extensions disponibles pour Claude Cowork, ainsi que des cas d'utilisation pour lesquels il peut être utilisé. Si vous êtes intéressé par l'apprentissage de l'utilisation de Claude Cowork et de ses capacités, cette vidéo est pour vous. Nous allons vous montrer comment utiliser Claude Cowork pour créer des tâches automatisées, gérer des projets et même utiliser des curseurs pour interagir avec votre ordinateur. Nous allons également aborder les concepts de perplexité et de bolt, ainsi que leur relation avec l'intelligence artificielle. Rejoignez-nous pour découvrir comment vous pouvez utiliser le code qui vaut 9 milliards pour améliorer votre productivité et votre workflow avec l'aide de l'agent d'intelligence artificielle lovable, Claude Cowork.
Catégorie
🤖
TechnologieTranscription
00:00Look how it works. The architecture of agentic systems today is still the same. Here's how it works.
00:05walk.
00:05When you ask a question, what happens is that your prompt will be understood by what
00:10'We call an orchestrating agent.
00:14Do you think you are able to use the perfect prompts given by these influencers to
00:20How can we configure these types of systems?
00:22And that's exactly what we're going to talk about today in this video.
00:26Amazon has just laid off 14,000 executives. Laid off in a single announcement.
00:31Microsoft, Google, and Salesforce did exactly the same thing.
00:36And do you know what they officially communicated in their letter? Two letters. I.A.
00:41No economic motive, no reason for bankruptcy, simply artificial intelligence.
00:48But here's the thing, what nobody tells you is that the people who did this, those who
00:53builds these systems that replace the frameworks,
00:56These are not Google engineers with 15 years of experience.
01:00These are the people who understood one thing, one single thing, the ones we're going to talk about today.
01:07in this video.
01:08Let me give you the raw numbers, not the influencer figures.
01:12In 2025, Amazon announced 14,000 corporate-type job cuts.
01:18Even managers are being laid off these days.
01:20In the United States, the permanent office job is collapsing.
01:23Mass layoffs, frozen recruitment, radio silence after interviews.
01:28Even those with 5 to 6-figure salaries are no longer exempt.
01:31The cause? An explosive cocktail.
01:34I.A.
01:35Profitability.
01:35Salary too high.
01:37And impatient investors.
01:38Amazon is eliminating 14,000 management positions.
01:41UPC has already fired that many people.
01:43Tens of thousands of office workers find themselves out on the streets.
01:47Without visibility.
01:48Without a plan B.
01:49Even recent graduates no longer have an entry ticket.
01:52And what about those who remain?
01:53More pressure.
01:54Less time.
01:55Reuters confirmed that the initial target was double the planned amount.
01:5930,000.
02:00The official reason, I quote,
02:03"The transformative growth of generative AI."
02:06Translation, Chad GPT, Claude.
02:08CNBC ran an article about it with the headline
02:12"Artificial intelligence is wiping out jobs."
02:16Do you know the overall result?
02:18We are at more than 1 million jobs by 2025.
02:211 million.
02:22Just for the United States.
02:23Okay then.
02:24You might say, it's the United States, it doesn't affect me.
02:27So, allow me to correct you.
02:29What is happening in the United States today,
02:31This is the laboratory for what will happen here in 6 months.
02:36It has been the advertising laboratory since 1988.
02:38I've worked for many, many big brands.
02:41And now, lately,
02:44even if I had requests,
02:45In the end, it doesn't work out because I'm told,
02:47Okay, in the end, we're going to do it internally with AI.
02:494 months.
02:50If we haven't already started there.
02:52So, it's always like that.
02:53It starts at their place.
02:55And once the laboratory is running at full speed,
02:57This will impact us.
02:58For what ?
02:59Because companies there will produce faster,
03:02Cheaper and better.
03:03While we are still asking ourselves the question
03:06how are we going to pay back
03:08France's 3 trillion euros of debt.
03:10But what becomes really interesting,
03:13The thing is, there are actually two ways to read this information.
03:16The first is fear.
03:17The retreat.
03:18I'm going to wait and see what happens.
03:20The second is to build systems
03:23which replace positions.
03:24To become indispensable to this economy.
03:27And how can I be part of this team?
03:30And that's exactly what I'm going to explain in this video.
03:33You know what struck me the most?
03:35It's not so much the number of layoffs.
03:37One million.
03:38But it's more about speed.
03:40Cursor.
03:41Are you familiar with Cursor?
03:42The coding tool.
03:43It was founded just a few years ago,
03:45by 4 MIT students.
03:48Do you know how much it was worth after 3 years?
03:51$2 billion.
03:51Do you know what Cursor's valuation is today?
03:55$29.3 billion.
03:58We are currently in discussions on the markets for 50 billion.
04:03I'm going to ask you a question.
04:05What have they done that you haven't?
04:08And I think the answer will surprise you.
04:10What has Cursor, Lovable, or Bolt done that you haven't?
04:14Here's the effect.
04:15Cursor is worth 29.3 billion.
04:18Lovable, 6.6 billion.
04:20300 million raised in December.
04:23Bolt New is worth 105 million.
04:25And it's currently on its way to a billion.
04:28Do you know what they have in common?
04:29All three of them.
04:30Same sector?
04:31No, that's not it.
04:31Cursor is doing coding.
04:33Lovable does vibe coding.
04:34And Bolt does rather rapid prototyping.
04:37These companies are collectively worth $35 billion.
04:42And that's what's behind it.
04:44ChatGPT or even this, Cloud.
04:46And none of these companies, neither Cursor nor Lovable, have learned from ChatGPT or Cloud,
04:53neither to speak, nor to function, nor to create a system from scratch from a neural system
04:59starting from scratch.
05:00No, not at all.
05:01Do you know what they did?
05:02They didn't invent AI.
05:03They learned to code system instructions.
05:07Exactly what I teach you in the training courses.
05:11Understanding how system instructions are structured, how AI agents and LLMs are converted to respond to
05:19business and societal problems and how to value them.
05:23What you see before you is the prompt 9 billion system.
05:28This is Cursor's prompt system.
05:30People don't realize that, because of the prompts you get from social media, the
05:35influencers,
05:36The number one skill today is having AI engineers who are capable of changing and configuring
05:43neural systems using prompt engineering,
05:47by having knowledge of code to structure system instructions.
05:51And that's valued at 9 billion.
05:53So, I have a question for you.
05:54Why do you have this type of interface today?
05:57Simply because they are orchestration layer interfaces.
06:02Behind, you have exactly the same thing.
06:05And these overlays are now obsolete.
06:08With the arrival of Claude and ChatGPT in SaaS mode,
06:11The SaaS investment world has fallen by more than 15% in less than 24 hours.
06:17This is a trend that continues.
06:19Software sales are collapsing.
06:22For a very simple reason, we no longer need to create interface overlays today.
06:28You have companies that have only one need: to access their data locally.
06:33configure systems, give instructions and make them work in a business.
06:39Whether you know how to use Claude, Codex, or any other type of AI today that is capable of
06:45to do this type of work,
06:46You no longer need to code an interface.
06:49You no longer even need to deploy over the internet.
06:52You transfer skills to companies that need reliability, security, and personalization.
07:00You have in your hands exactly the same tools as those who created Bolt, Lovable, Cursor.
07:06You have ChatGPT and you have Claude.
07:09The question today is, do you have the skills to be able to do it?
07:13If today you are using the methods of perfect promps, those given to you by influencers,
07:18Do you think you are able to use the perfect prompts given by these influencers to
07:24How can we configure these types of systems?
07:26Tell me in the comments, honestly, do you think these prompting methods will allow us to configure these
07:32systems?
07:32A mysterious hacker extracted and leaked a very large part of the workings of Perplexity Computer this weekend.
07:39And this mysterious hacker put all this information into the Perplexity Agent AI information, how to transform Perplexity into an Agent
07:48AI.
07:48Perplexity is a system that is valued at around one billion.
07:52Giving access to the internal system makes you aware of the power of system instructions and how they work.
08:00Prompt engineering in the professional field, not the kind you find on social media.
08:04There are two approaches today that need to be examined with AI, and why there are two
08:08completely different orientations.
08:09You have Perplexity, so classic AI, in which you ask a question.
08:13What is the latest information on the leak of the new anthropomorphic model Claude Mythos?
08:19In this context, you are on what is called a retrivial system.
08:23And the way it works, as you will see, is essentially the same.
08:26The model identifies the key points you want to search for; it will perform a query on an internet request; it
08:33It retrieves information, consults it, extracts it, and then compiles a response.
08:38But beyond that part, you won't be able to do much more.
08:42Everything changes when you enter a computer system.
08:44See how it works.
08:45The architecture of agentic systems today is still the same.
08:48Here's how it works.
08:50When you ask a question, what happens is that your prompt will be understood by what
08:55'We call an orchestrating agent.
08:57You will have a primary agent who receives the information.
08:59From there, he will understand your request and make an inventory of what he needs to do.
09:05And so, he's going to make what's called a plan.
09:07And you'll see that it will define who I need, which connector, tools, MCPs,
09:12What interface do I have available, what tools do I have to choose?
09:16What skills do I need?
09:18And then, it will send parallel executions.
09:21What does that mean?
09:22AI has changed dimensions in terms of scaling.
09:25We're no longer dealing with a system where you ask a question, the element goes online, it retrieves the information, and then...
09:29Give an answer, that's it, it's over, you have your own perfect.
09:32No.
09:32You have switched to a system that is currently understanding your request, defining the set of tools, which
09:37will deploy AI agents, ChatGPT and Claude to virtual employees.
09:43In fact, each time, you have to understand that this is Claude, this is ChatGPT, this is ChatGPT,
09:47and the system self-evaluates, exchanges information and creates an autonomous system capable of performing a task.
09:54And that's simply incredible.
09:57The question today is whether you are capable of moving from that system to something else.
10:01something that you will be able to sell, use in business, or offer as a service to businesses?
10:06I'm going to explain what skills you need to develop in order to do that.
10:09The processing is always the same.
10:11You have a Head Agent system that will retrieve information, define how many sub-agents, and therefore how many employees
10:19AI, AI agents, he's going to need.
10:21For these employees, he will have to define their roles and what type of contract he will assign them.
10:25In terms of context, what type of specialization will he give them?
10:29And inside, they will need to know what the MCP functions are, what the web searches are, the code,
10:34the API functions they can use, and how they will manage the exchange of information.
10:38Because at some point, there will be memories to exchange between them, to know who is doing it
10:42what and how they discuss.
10:44So the architecture of this system today is based on the skills that companies are looking for, and which
10:49allows you to move from simply an interface to a front-end,
10:53where behind it there was ChatGPT or Cloud, to an application where you no longer have a front-end.
10:58It is directly ChatGPT, Codex, Cloud, Code, Cowork or even Perplexity Computer that gives you the front-end.
11:06And now, what skills will you put in your AI agents? What skills will your agent have?
11:12will have,
11:13how you will model your company's processes, and what connectors you will use.
11:18Today, if you go into that box, you have an entire company that can make that system work.
11:24You have everything: Google Drive, Notion, Dropbox, Asana, CRM, HubSpot, Slack, Teams, Vercel, everything, Stripe payment method, Appify.
11:35We even do analysis for those who do SEO with YouTube Analytics or Google Ads, Paypal,
11:41Mailing.
11:42You have an infinite number of connectors to make the AI work in autonomous mode.
11:47However, you won't be able to tell the AI, "You're a specialist with 15 years of experience."
11:51Use all your skills, do my work for me, because I saw it in the tutorial on
11:57YouTube.
11:57That won't work.
11:58The three skills to develop immediately to be among those who have understood the opportunity of the revolution
12:05AI agents and agentic systems in 2026,
12:09It's about understanding the prompt engineering of advanced systems, the kind used for Cloud Code, Cloud Cowork, and ChatGPT.
12:15Agents,
12:16or even Cloud CLI, Perplexity Computer, in short.
12:20Agentic systems are coded in this way.
12:23The second essential skill to have this year in 2026 is knowing how to model skill functions.
12:30The skill system is the universal system.
12:33All AI agents today operate on skill systems.
12:36I've included some videos in the description to help you understand what it's for and how to use them.
12:41structure.
12:41You have a skill, you are capable of developing that skill for an entire team within your team.
12:47to standardize, transmit and reuse.
12:50And with this skill, you'll be able to monetize your expertise by transferring it, selling it to companies, by
12:56customizing it within the company.
12:58You no longer need to waste time with a front-end.
13:01You transfer adaptive skills to the company.
13:04The third skill to have is understanding MCPs, how they are structured, and how to define MCP code lines.
13:09and how to coordinate them in a prompt system.
13:12The fourth skill to develop is how you will write prompt systems to coordinate your AI agents.
13:19the memory, MCP functions and tools inside this loop.
13:23And that's exactly what we do with level 2, mastering the best of AI.
13:27You will learn to automate what exhausts you.
13:30You will be officially certified because I give you all the methods to pass the diploma exams.
13:36You will master your data and you will know, in less than 15 days,
13:41scale your business with skills that no one else on the market has today.
13:45I'm not giving you a tutorial.
13:47I'm giving you an artificial intelligence ecosystem for the professional world.
13:52Agentic systems, compared to a classic chat system,
13:55It's a system that does what's called a gathering context.
13:58It retrieves information, whether it's an internet search or a prompt,
14:02and he will have to choose between two possibilities.
14:04Does he need to take action, that is, write a text?
14:07Or does it need to trigger a tool?
14:09And that part, that choice, depends on your instructions.
14:12depends on how you write the information.
14:15The additional step is whether the model is capable of verifying what it comes from itself.
14:19to create?
14:20And so we add an iteration function.
14:22Am I capable of analyzing the result I was asked to provide?
14:25The agentic system must be viewed with three pillars.
14:28The gathering context, how I retrieve the information, on what basis,
14:32What type of decision is made regarding the model, and on what basis?
14:35and how he will choose to check the work he has done.
14:38That's the overall loop of AI systems today.
14:42which must be mastered through the tools,
14:45Prompt Engineering Advanced in the field of AI prompt systems,
14:49MCPs, skills,
14:51with the integration of functions known as HITL,
14:55that is to say, the possibility of human intervention in the functioning of the model.
14:59In the training courses, you will learn to define the tools,
15:02to be able to integrate multi-step processes into an AI system,
15:07to understand how you will use the functions in an agentic system,
15:12to code the agent loops themselves,
15:14to create employees,
15:16and to understand how these models work,
15:18and how they differ from one another,
15:21so that you can optimize both the cost,
15:24the auditable aspect
15:25to produce content for a company,
15:27It's not the same as making a video on the internet.
15:29You need to know how these models work,
15:32so that you have reliability in the LLM responses.
15:36And that's exactly what we do within the training programs.
15:39If you enjoyed this video,
15:40Feel free to put it in the description or comments.
15:43Those who want to go further,
15:44to learn how to use AI,
15:45and to scale your business,
15:47to take control of AI,
15:49and to seize this opportunity,
15:50AI agents in 2026,
15:52You have all the necessary training courses listed in the description.
15:55I'll see you soon, until next time.
15:57Subscribe if you haven't already.
15:58See you later.
Commentaires