00:00:00 → 00:01:02
Right now, I have AI agents running on a computer in my office while I'm out of the house, at the beach, at dinner, wherever, they're working, coding applications, running market research, and sending me reports automatically. Not long ago, I experienced a watershed moment with AI. I was right here on the beach texting my AI employee from my phone, iterating on a project it was coding all day and getting responses in real time, all from an autonomous system in my office while I was standing here on the sand. That is the moment why I'm making this video. My name is Mars Deutsche. I'm the founder of AI Edge and I've been deep in this technology since Chat GBT first launched. Last year, I founded an AI tech and media company because I saw where this was heading. In October, I flew back to Melbourne to visit friends and family. Went to the pub, went to dinners, caught up with people I hadn't seen in months, and I realized something that shook me. They're not using AI. Well, not really. Maybe chat GBT for a question here or there, maybe a spreadsheet formula, but that's it. No workflows, no automations, no experimentation. Then I visited my dad. He runs an art company. Great business, but hands-on sales, leads,
现在，我在办公室的电脑上运行AI代理，而我可以在家外、海滩、餐厅等任何地方，它们在工作、编写应用程序、运行市场研究，并自动向我发送报告。不久前，我经历了与AI的一个分水岭时刻。我就在这里的海滩上用手机给我的AI员工发短信，迭代它整天编写的项目，并实时获得响应，所有这些都来自我办公室里的自主系统，而我站在沙滩上。这就是我制作这个视频的原因。我叫Mars Deutsche。我是AI Edge的创始人，自从ChatGPT首次推出以来，我就一直深入研究这项技术。去年，我创办了一家AI技术和媒体公司，因为我看到了这项技术的发展方向。十月份，我飞回墨尔本看望朋友和家人。去了酒吧，参加了晚宴，与几个月没见的人聚在一起，我意识到一些让我震惊的事情。他们没有使用AI。好吧，不是真的在用。也许在这里或那里用ChatGPT回答一个问题，也许用一个电子表格公式，但就这样。没有工作流，没有自动化，没有实验。然后我拜访了我的父亲。他经营一家艺术公司。很好的业务，但需要亲自处理销售、线索。
00:01:02 → 00:02:03
clients, databases, cataloges, print design, email responses. But when I asked him about AI, he said he was still figuring out how to prompt chat GBT. None of the systems were automated. Not one. Then I flew home to Dubai and sat down with my housemates and asked, "You guys are using AI, right?" But they weren't. They didn't even really know where to start. They weren't hostile towards it, just unaware. Think about this. I have a 30 employee AI and media business. And if that's happening in my own house, then the gap is way worse than I first thought. I think people let what they want to believe in get in the way of what's actually happening. They don't want to believe that this is the future, so they tell themselves it's not, but it is. And pretending otherwise doesn't slow it down. A year ago, I was using AI the way most people do, as a chatbot. I would ask it a question, get an answer, and that's about it. Then I started pushing it. I got it to do my accounting, chuck my bank statements into a folder, and within minutes, everything was sorted, categorized, and reconciled. Something that used to take me hours. Then I discovered I could build applications, full applications, without a developer or a budget, just me
客户、数据库、目录、印刷设计、电子邮件回复。但当我问他关于AI的事情时，他说他仍在弄清楚如何提示ChatGPT。没有一个系统是自动化的。一个都没有。然后我飞回迪拜，和我的室友坐在一起问："你们都在使用AI，对吧？"但他们没有。他们甚至不知道从哪里开始。他们并不反对它，只是不了解。想想这个。我有一个30人的AI和媒体业务。如果这发生在我自己的房子里，那么差距比我最初想象的要大得多。我认为人们让他们想相信的东西妨碍了实际发生的事情。他们不想相信这是未来，所以他们告诉自己它不是，但它确实是。假装不是这样并不会减缓它。一年前，我使用AI的方式就像大多数人一样，作为聊天机器人。我会问它一个问题，得到一个答案，仅此而已。然后我开始推动它。我让它做我的会计，把我的银行对账单放进一个文件夹，几分钟内，一切都被分类、整理和核对了。这曾经花费我几个小时的事情。然后我发现我可以构建应用程序，完整的应用程序，没有开发人员或预算，只是我。
00:02:03 → 00:03:04
describing what I wanted in plain English. For example, I do a lot of trading and investing, and I couldn't find an analytics platform that gave me the specific correlation data that I needed. So, I described my painoint to Claude. And a few hours later, I had a full working custom market scanner to back test correlations for any data point or indicator. Exactly what I wanted without AI. That's probably a $20,000 project. And I did it in a few hours with a $20 subscription. And after that, I started building everything. Habit tracking apps, custom dashboards, second brain. Each one took 30 minutes or less. And that's when it finally clicked. I can build anything that's in my head, any workflow, any tool, and anyone can. Then I went even further. I set up autonomous AI agents on a Mac Studio, a coder, a researcher, a strategist, a content writer, a market analyst, all running on my computer. And I can leave the house while they work. I can text them from my phone and iterate on the go. I can call them, they send me updates on Telegram, even without me asking. 12 months ago, I was just a
用简单的英语描述我想要什么。例如，我做了很多交易和投资，找不到能给我所需特定相关数据的分析平台。所以，我向Claude描述了我的痛点。几个小时后，我有了一个完整的工作自定义市场扫描仪来回测任何数据点或指标的相关性。正是我想要的，没有AI。这可能是一个20,000美元的项目。而我在几个小时内用20美元的订阅完成了。之后，我开始构建一切。习惯跟踪应用程序、自定义仪表板、第二大脑。每一个都花了30分钟或更少。正是在那一刻，一切终于点击了。我可以构建任何在我脑子里的东西，任何工作流、任何工具，任何人都可以。然后我走得更远。我在Mac Studio上设置了自主AI代理、编码员、研究员、策略家、内容作者、市场分析师，所有这些都在我的电脑上运行。我可以在出门时让他们工作。我可以从我的手机给他们发短信，在路上进行迭代。我可以呼叫他们，他们甚至在我没有要求的情况下在Telegram上向我发送更新。12个月前，我只是一个。
00:03:04 → 00:04:04
chatbot user. Today, I run an AI powered operation. So, here's the question that keeps me up at night. If that happened in just 1 year at an exponential pace, what happens in five with hundreds of billions in capital investment that hasn't even come online yet? Data centers only just getting green lit, energy infrastructure still just being built. I see the writing on the wall. I need to tell you something I'm not proud of. Recently, I let go of five employees and I replaced them with AI agents running on my computer. a crypto researcher replaced with an AI that scans the market around the clock. An operations person replaced by a strategic AI system that handles planning, a content writer. I trained a model on my own voice and writing style. And three developers, a website dev, a software dev, and a general coder replaced with a singular AI model. That's not a flex. That's the reality that I had to face. And if it makes you uncomfortable hearing it, good. Because that discomfort is exactly what I'm trying to give you before your boss
聊天机器人用户。今天，我经营一个由AI驱动的业务。那么，让我夜不能寐的问题是什么？如果在仅仅1年的指数级步伐内发生了这种情况，那么在5年内会发生什么，还有数百亿美元的资本投资尚未上线？数据中心才刚刚获批，能源基础设施仍在建设中。我看到了墙上的字。我需要告诉你一些我不为之自豪的事情。最近，我解雇了五名员工，我用在我电脑上运行的AI代理替换了他们。一个加密研究员被一个全天扫描市场的AI替换。一个运营人员被一个处理规划的战略AI系统替换，一个内容作者。我在自己的声音和写作风格上训练了一个模型。还有三个开发人员，一个网站开发人员、一个软件开发人员和一个通用编码员，由一个单一的AI模型替换。这不是炫耀。这是我必须面对的现实。如果它让你听着不舒服，很好。因为这种不适正是我试图在你的老板之前给你的。
00:04:04 → 00:05:04
makes the same decision about you. Because if I'm doing this as a small business owner with 30 people, imagine what's happening at companies with 5,000 or 50,000. Now, I'm still hiring some humans, but every person that I hire from now on will leverage my AI systems. That's the baseline. If you can't work alongside AI, you can't work for me. And I'm far from the only person thinking this way. I have friends who are accountants. I'm worried about them. Friends in finance, worried about them. any role where the core work happens on a screen, reading, writing, analyzing, deciding, decoding, AI is coming for significant parts of it. Not in 10 years, but now >> AI um will is is really still digital. Ultimately, um AI can improve the productivity of of humans who who do things with their hands like plum, you know, literally welding, electrical work, plumbing, anything that's that's physically moving atoms, like anything that's that's physical, those jobs will exist for a much longer time. But anything that is digital, uh, which is like just someone at a computer doing
对你做同样的决定。因为如果我作为一个30人的小企业主在这样做，想象一下5000人或50000人的公司正在发生什么。现在，我仍然在雇用一些人类，但从现在开始，我雇用的每个人都将利用我的AI系统。那是基准。如果你不能与AI一起工作，你就不能为我工作。我绝不是唯一这样想的人。我有做会计的朋友。我为他们担心。朋友在金融领域，为他们担心。任何核心工作发生在屏幕上的角色，阅读、写作、分析、决策、解码，AI正在针对其中的重要部分。不是在10年后，而是现在。AI嗯确实是真的仍然是数字。最终，嗯AI可以提高那些用手做事的人的生产力，比如水管工，你知道，字面上的焊接、电工、水管，任何物理移动原子的东西，比如任何物理的东西，这些工作将存在更长时间。但任何是数字的东西，呃，这就像某个人在电脑上做某事。
00:05:04 → 00:06:04
something, AI is going to take over those jobs like lightning. >> My videographer, the guy filming this right now, said to me this morning while I was preparing the script, "The world is fucked." Now, he's in a creative field that's also under threat by new video models, which will soon be able to replicate full-scale Hollywood movies in just a matter of minutes. The quicker that you accept that no one is safe, the quicker that you can get ahead, because the window to be early is still open. But admittedly, it's closing. So, what can you actually do? Well, here's what I'm telling my family and friends. Firstly, start using AI seriously. Sign up for Claude, which is my favorite model, or even chat GBT and use the best model available, not the default. Free version of these tools are often months behind what paying users have access to. If you tried the free version and thought, well, this isn't that impressive, you haven't seen what it can actually do. It's mind-blowing the power that these models now have. Secondly, write down everything that you do on a daily or weekly basis. Then pick one workflow to automate. Just start with one workflow at a time. Focus on one thing every week and keep stacking
什么，AI将快速接管那些工作。我的摄像师，现在正在拍摄这个的人，今天早上当我准备剧本时对我说，"世界完
00:06:04 → 00:07:04
workflows. Start with the thing that you spend the most time on, the repetitive thing that eats your afternoon. Just feed it to the AI and see what happens. I'm willing to bet that you'll be able to automate it. If you're an accountant, give it a full client return. If you're in sales, give it your CRM data. If you're a manager, paste in your quarterly numbers and ask it to find the story. Don't assume it can't. It probably can. Just try it. And if it kind of works today, in 6 months, it'll probably be able to do it near perfectly. Thirdly, spend 1 hour a day not just reading about AI, but actually using it. Every day, try something new. Then try something slightly harder, something you're not even sure you can handle. Keep a log on Excel or your AI dashboard of your experimentations over time. You'll be mind-blown what you can achieve in a few months time. And if you do this for, let's say, 6 months, you'll understand what's coming better than 99% of people around you. The bar is low because most people are still far too complacent. I could get into more specifics now, but this is exactly what I break down on this channel. The workflows, the automations, the systems.
00:07:04 → 00:08:04
So, subscribe if you want the exact systems that I use to scale and run my business. Fourthly, get your finances in order. I'm not a financial adviser, but if you believe even partially that the next few years will be disruptive, having savings and flexibility matters much more than it did just a year ago. The best part is AI can help you with this. whether it's automated accounting or financial planning. And lastly, rethink the old playbook. Many of the old tropes you were taught by your parents or you learned at school are sadly no longer true. I actually told my younger sister recently to seriously consider whether university is the right move right now. Knowledge is free. It's in your pocket, available at any time. There's no premium on what you know anymore. The premium is now on agency and emotional intelligence, on taste, on curation, on knowing how to direct these tools at problems that actually matter. It's never been a better time to be an ideas person. The winners will be the people with enough self-awareness to leverage what's available. And while you're doing all this, make sure to prioritize your health. That's the one thing that AI can't automate for you.
00:08:04 → 00:09:04
And ironically, it might be AI that actually unlocks major longevity breakthroughs. So, make sure you're around to benefit. I've spent a lot of time in this video on the threats that AI represents. But let me also tell you the other side. Being able to take any idea stuck in my head and actually build it for 20 bucks that used to cost a million dollars in development is absolutely wild. I'm more excited than I've ever been. I'm building an AI media company, building applications, investing in the underlying technology, and the past few weeks alone have made it crystal clear that this is the biggest opportunity of my life. But I'm also scared. I'd be lying if I said I wasn't. This may be the biggest change that humanity has ever undergone in such a short time span. It's moving faster and more aggressively than the industrial revolution. Here's what I keep coming back to, though. Thinking about how scared you are doesn't change the reality of what's happening. You can either bury your head in the sand or embrace it. And if you bury your head, you might not like what you find when you finally look back up again. I read an article by Matt Schumer that put this feeling into words better than I could. Links in the description below. It's
00:09:04 → 00:09:30
worth your time and it inspired this video. So, dear family, dear friends, dear anybody watching this, I don't have all the answers. No one does. The window to be early is still [music] open, but not for long. I'd rather be the person who warned you too early than be the one who said nothing and watch you get blindsided. Now, go download Claude or ChatgPT and start experimenting. I'm [music] serious. And if you want help, that's exactly what this channel's for. Subscribe and I'll see you in the next video.