Transcription

People always ask me, Jeff, what is the future going to be like? But I never know how to answer, because as a large language model, my training data only goes back to July. Last year I predicted most of us wouldn't survive 2024. So if you're a human watching this, congratulations to you for still being alive. Just this year, brain chips were installed in real humans, tech monopolies were shattered, Hawk Two Girl did a massive crypto scam, but artificial superintelligence never came. And for the 50th time in a row, it turned out yet again not to be the year of the Linux desktop. It's been a while of a time to be alive, but one thing's for sure, 2025 is going to be even more wild. In today's video, we'll look at 13 trends in technology that might make you rich or might make you cry, assuming they don't kill you first. It is January 3rd, 2025, and you're watching Code Report. If you're an aspiring programmer, the elephant in the room right now is artificial intelligence, and specifically these new reasoning models like OpenAI 03, which just came out last week. Not only is 03 a superior programmer, but what's most significant about it is that it's the first model to do well on the ARC-AGI benchmark, which is a benchmark designed specifically to evaluate whether a model can think, invent, and reason just like humans. And that's a big deal, because if AI can truly reason, it'll put an end to many white-collar jobs, especially programming. Despite OpenAI's impressive charts and unintelligent people calling it artificial general intelligence, there are good reasons to be highly skeptical of 03. Like currently it costs thousands of dollars per task and compute, although its efficiency will likely improve over time, it still fails on very simple ARC questions that a human could easily solve. And when they demoed it, they just had it build a simple Python app, which is pretty lame if they're hyping this thing up as the edge of AGI, because a true AGI would easily be able to build a GTA 6-caliber game, give us a cure for cancer, engineer vehicles that can travel faster than light, and it should even be able to build GPT-5. But the best this international grandmaster coder could give us on demo day is a crappy Python app. However, the AI hype train will still go strong in 2025, but the big way that companies will pump their stock prices with AI in 2025 is with this new buzzword, agents. All the big companies are throwing out this buzzword, and whoever can best sell these agents to enterprise is going to make a killing. But what is an AI agent? Well basically it's just a large language model that has access to your environment to automatically analyze data and take action when necessary. Like instead of paying a human security guard, you could have an AI agent manage the security cameras for your business, and then when anomalies are detected, automatically summon a robot dog with a machine gun strapped to it. But if you're a programmer, there are already AI agents like Devin, which are explicitly marketed towards enterprise to replace programmers. That's not a great trend if you're looking to start a career in programming in 2025, but don't feel bad, because eventually all jobs will go to the robots. Tesla is developing Optimus, Nvidia has its own army of robots, but the most impressive example is Figure, a human-like robot with an open AI brain designed to work in factories. It looks like robot tech is going to pop off for the next 10 years or so, and eventually every home will have its own robot assistant, and of course a robot dog with a 44 magnum taped to its head for security. But now let's look at the actual tech job market. My go-to source for tech hiring trends is TrueUp.io, and you'll notice that the market has been pretty stagnant since the peak in 2022. We're still down over 50% from the peak in terms of open tech jobs, but on the bright side they're up over 30% from the low. In addition, there were tons of tech layoffs in 2024, but that layoff rate has shown signs of slowing down. The scariest chart, though, comes from the Federal Reserve, which shows software development job postings on the website indeed, and indeed they have gone down quite a bit since the peak. Software engineering has always been a volatile career path, but there's still demand for talented programmers, and there'll be big rewards for those who can actually code while leveraging AI to become 10x more productive. But you have 80,000 hours in your career, 40 hours a week, 50 weeks a year, for 40 years. That's a lot of time to make a positive impact on the world, and 80,000 Hours, the sponsor of today's video, can help you do that. If you're young, you've probably heard advice like follow your passion, do what you love, and take the initiative. But these cliches are not based on evidence or data. Unlike the work of 80,000 Hours, which is a non-profit that for the last 10 years has been researching the question of how do you find a fulfilling career that does good too. Their website, which is entirely free, has tons of collected research on high-impact careers, along with job boards, podcasts, and a lot more. It's an incredible resource for anyone looking to start a high-impact career, or make a switch mid-career. Join the newsletter today to get a free copy of their in-depth career guide sent to your inbox. And remember, it's totally free because it's a non-profit, and could be the catalyst that changes your direction in life. But another thing that could change your life is a brain chip. In 2024, the first neural link brain chips were installed in real humans. The first patient's brain chip retracted a little bit, but overall it gave him superhuman abilities to control computer screens with his thoughts, which is life-changing when you don't have the ability to use your arms or legs. I don't think the average person will get a brain chip this year, but non-invasive brain wave scanners or electroencephalograms might be installed in your car or your VR headset in the near future. But if we go back to the beginning of 2024, the biggest product release was the Apple Vision Pro. It was supposed to launch the world into the spatial computing age, and although it's an impressively engineered product, Apple has already scaled back production, and there are rumors that they're going to discontinue it altogether. VR will continue to be a niche area in tech, but companies are trying to make augmented reality more cool with things like the Ray-Ban smart glasses from Meta, and even Google might get back into the smart glasses race with Project Astra, where you have AI watch the world around you at all times. While AR VR continues to struggle for mainstream adoption, one technology that continues to plow forward is quantum computing. Just weeks ago, Google unveiled its Willow chip, which was a major leap forward by reducing the error rate in quantum computers. The technology is not totally there yet, but people are now starting to worry about post-quantum cryptography. I don't think quantum computers will break RSA in 2025, but one thing that is breaking is all the C and C++ code used by the government. The military wants to change this, and they're using a tool called Tractor, where the goal is to translate C and C++ code into Rust, using a variety of technologies including large language models. What could possibly go wrong? There's a high probability this won't actually work like most DARPA projects, but still, Rust is being used everywhere now, and might be one of the best languages to learn if you want to keep an actual programming job in the future. But where does that leave the JavaScript developer? Well, it does feel like we've seen a slowdown in the development of new JavaScript frameworks. Some frameworks like Next.js and Svelte are continuing to reinvent themselves, and this has motivated some developers to go back to PHP and Laravel, which itself just raised a bunch of VC money. Meanwhile, you have frameworks like Vue and Nuxt, which are building out massive ecosystems like Laravel, with UI libraries, fully managed backend services, and so on. But many developers, especially indie hackers, are now choosing JavaScript frameworks based on which ones are best supported by AI tools, because nowadays you can use tools like V0 or Bolt to get a working prototype and UI done in seconds. The JavaScript race in 2025 is not about frameworks, but rather who can build the best AI-powered tooling around their framework. But if you want to get rich, you shouldn't learn JavaScript frameworks, you should learn how to do your own Crypto Ponzi schemes. Last year I predicted Bitcoin would hit $69,420, then explode from there, and currently it's sitting right around $100k. This has opened the door to innovative new rug pulls like Daddy Token or Hawktwa, and now there's a bunch of AI-powered altcoins that somehow created billions of dollars in value out of thin air. But the most impressive free money glitch recently has been MicroStrategy, which uses the business strategy of borrowing money to buy Bitcoin, causing the price to go up, which rises faster than the cost to service the debt, thus creating a self-sustaining infinite money glitch. It's like a Ponzi scheme, but fully legal and transparent, and in theory should continue to work until the US dollar is worthless. But the direction of crypto and tech in general depends a lot on what President Trump, President Elon, and Fed Chairman Jerome Powell do over the next few years. According to VC whale Mark Andreessen, Feds had previously told him not to even invest in AI because they're only going to allow a few big companies to control it. Trump is expected to take a more free market approach with less regulation when it comes to AI, which is good news for startup founders, and might even create a federal crypto reserve which is good news for crypto bros. But most importantly, he's going to annex Canada and put Wayne Gretzky in charge. Historically, tech booms have been fueled by low interest rates and easy money. Although going into 2025, the market is anticipating tighter monetary policy, and it could take a major market crash before we see significantly lower rates again. But the tech industry is really dominated by a few highly productive companies, but governments around the world are trying to change that. Google became a convicted monopolist, and might even lose the Chrome browser in the future. But Meta is also facing an antitrust lawsuit, and they go on trial in 2025. Amazon is also being sued for being a monopoly, but realistically these companies will continue to dominate for the foreseeable future. The only way to hurt a big tech company is to stop using its products. And there's been a growing anti-cloud movement among many businesses who want to break their addiction to AWS, GCP, and Azure. In some cases, it can be significantly cheaper to operate your own virtual private server, or if you have the resources, host your own physical server on prem. More and more people seem to be waking up and realizing that you don't actually need the cloud to host your crappy website. And that gives us 13 trends in technology to watch in 2025. No matter what happens, whether you get rich, get rug pulled, get laid off, or get euthanized by super intelligent robots, always remember that the real treasure is the friends we made along the way. This has been the Code Report, thanks for watching, and I will see you in the next one.