Florian Dedov

I was able to meet Jensen Huang the CEO of @nvidia at #GTCParis. This guy is a legend!

Teaching and creating content in tech is one of the best decisions I ever made. After recording over 1000 video tutorials and doing content production in tech for more than seven years, I want to summarize the biggest benefits and maybe inspire more engineers to share their knowledge and insights. So here we go: 𝟭.🧠 𝗧𝗲𝗮𝗰𝗵𝗶𝗻𝗴 𝗙𝗼𝗿𝗰𝗲𝘀 𝗨𝗻𝗱𝗲𝗿𝘀𝘁𝗮𝗻𝗱𝗶𝗻𝗴 Explaining something to an audience requires a higher level of clarity and understanding than just applying the concept. In order to use something, it's enough to roughly get the idea. In order to teach it, you need to break it down to its essence and understand it fully. It's quite similar to the Feynman technique. You need to be able to explain something to a child (or unskilled person) and if you can't do that, you probably don't really understand it deeply enough yourself. Bonus: You can oftentimes combine learning new technologies or concepts (that you maybe need for other ventures and side projects) with teaching. This way, you continuously expand your skill set, while also building an educational brand and portfolio. Which brings us to the next point. 𝟮.⌚️ 𝗜𝘁 𝗦𝗵𝗼𝘄𝘀 𝗣𝗿𝗼𝗮𝗰𝘁𝗶𝘃𝗶𝘁𝘆 Having any sort of portfolio - be it a collection of YouTube tutorials, numerous blog posts or impressive GitHub repos - signals proactivity to potential employers, clients and business partners. This is especially true if you do it consistently for years. There is no better proof for discipline and industriousness than consistently delivering on your commitment over and over again. And in times where people are thinking twice about hiring "real people", it is more important than ever to stick out. All this work can speak for itself and it can sometimes help you to skip job interviews and CV requests. 𝟯.🚪 𝗥𝗲𝗮𝗰𝗵 𝗢𝗽𝗲𝗻𝘀 𝗗𝗼𝗼𝗿𝘀 This one only applies once you have a certain level of success. It doesn't matter whether you have a blog, a YouTube channel or a large following on LinkedIn. The moment you reach a certain size, you are a "somebody" in the eyes of many people. I personally find this way of thinking stupid but that is how a lot of people will perceive it. And because of this, more opportunities will present themselves to you. More offers, invitations and collaboration opportunities will come your way, just because people associate a large following with a bunch of other unrelated characteristics. I don't think that this should be the primary driver behind producing content or teaching, but it is one of the most significant benefits that you will experience, if you do this for long enough. And it can translate into success in many other areas of business. So if you have skills in IT, share them. If you don't like cameras, just write about them. But don't stay invisible.

@canmachinethink It's crazy. A month ago it was able to implement complex feature almost flawlessly. Now it makes ridiculous mistakes all the time.

@NeuralNine Lobotomized version to save compute? I feel like model degradation is very noticeable in gemini (through chat, not api)

Why is Claude so stupid all of a sudden?

Every week, major new models, papers, and frameworks are released in the field of AI/ML. Everything is evolving and changing rapidly. Don't forget that Claude Code is only about one year old at this point. So here is how I stay up-to-date on the most important things, by spending about 15 minutes per day. 𝟭. 📰 𝗥𝗲𝗮𝗱𝗶𝗻𝗴 𝗛𝗮𝗰𝗸𝗲𝗿 𝗡𝗲𝘄𝘀 I would argue that Hacker News is by far the best resource for tech people to stay updated. I've never seen an interesting model, paper or framework being released without a corresponding Hacker News post on the same day. And the best thing: For every post or link there is a whole discussion going on. Competent people giving opinions, asking questions and debating. Also, compared to platforms like Reddit, the community isn't toxic and discussions are not devolving into personal attacks or politics. The only challenge with Hacker News can be to filter out all the noise. Oftentimes, people will post fun side projects or interesting scientific facts, which may not be related to your specific field of interest. But there are also various websites offering filtered feeds of Hacker News. Link: news.ycombinator.com 𝟮. 📹️ 𝗪𝗮𝘁𝗰𝗵𝗶𝗻𝗴 𝗧𝗵𝗲 𝗥𝗶𝗴𝗵𝘁 𝗬𝗼𝘂𝗧𝘂𝗯𝗲 𝗖𝗵𝗮𝗻𝗻𝗲𝗹𝘀 While I would like to plug in my own channel at this point, it wouldn't fit, since I am mainly doing tutorials and courses. However, there are many great YouTube channels out there focusing exclusively on tech news and recent developments in AI/ML. For quick updates, I personally enjoy Fireship (@fireship_dev) and for in-depth analyses I like AI Explained (@AIExplainedYT). Link: @Fireship" target="_blank" rel="nofollow noopener">youtube.com/@Fireship Link: @aiexplained-official" target="_blank" rel="nofollow noopener">youtube.com/@aiexplained-o… 𝟯. 👥 𝗧𝗮𝗹𝗸𝗶𝗻𝗴 𝗧𝗼 𝗣𝗲𝗼𝗽𝗹𝗲 / 𝗝𝗼𝗶𝗻𝗶𝗻𝗴 𝗖𝗼𝗺𝗺𝘂𝗻𝗶𝘁𝗶𝗲𝘀 This one is a game changer. If you connect with like-minded intelligent people in the AI/ML space, you can keep each other updated, whenever something interesting happens. This works like a natural filter that only sends you updates that you will really be interested in. I like to do this both with individual people and in communities or groups. You then no longer have to actively seek out news. They come to you. All in all, this should easily be doable with 15 minutes per day on average. Usually, you won't have 2-3 groundbreaking updates per day. There will be 1-2 relevant Hacker News posts per day, 1-2 relevant messages every 2-3 days and 1-2 interesting deep-dive videos every 1-2 weeks. And that's all you need to not lose track.

The best thing about coding agents is that I never ever have to center a div again...

More than ever, it pays off to be a generalist in tech.

lol

We need a slur for people who use Microsoft products

Thanks to @NVIDIAAI & @Dell for sending this device to me and hopefully I will discover new interesting use cases for it over time! (🧵7/7) Link: dell.com/en-us/shop/des…

5 - 🖥️ Portable Data Center Finally, this device works perfectly as a portable data center. If you're moving to a different place for a couple of months, this device alongside a good laptop is the perfect combination. It easily fits into a backpack and doesn't need anything special to run. (🧵6/7)

This is what one PetaFLOP of compute looks like on your desk. I received a powerful device from @Dell and @NVIDIAAI - The Dell Pro Max with GB10. An AI Supercomputer that easily fits on your desk and allows you to run huge models at home. Here are my top 5 use cases. (🧵1/7)

I don't think SaaS is dead at all. People underestimate how busy everybody is. Sure, soon everyone will be able to vibe-code their own software. But will they do it? People with money usually still want to outsource it.

Guess who found a new hobby...

@elianxai I'd recommend learning the basics of Python, traditional ML, deep learning etc. Once you're comfortable with that I would go deeper into the SOTA architectures and papers. I would mainly learn by building stuff at this point. After this, use agents to speed-up development.

@NeuralNine Hi i love your videos , and i hope you answer to my question ; I want to be an ai engineer like you , but there are too many tools to learn and i am confused where to start what things should i learn for that ? Thank you !

Currently preparing a JS crash course. This language is a pure fever dream...