Join 7,000+ data professionals who are advancing their data and tech careers by learning from real-world projects, gaining first-hand experience, and accessing top resources delivered to your inbox every week.
Share
I met some folks from Google DeepMind. Here's what I learned.
Published about 10 hours ago • 4 min read
Hi Reader,
Last week, I was in New York City for the first-ever YouTube Developer Summit, organized by Google.
It was a small event, the first of its kind, with just 16 tech YouTubers from around the world who teach programming and other tech topics.
(Wow, I'm officially a "tech YouTuber" now 😅)
I learned a ton, met some amazing YouTubers whose work I've admired for years. And I got to chat with some very smart folks from Google DeepMind who are building Gemini models and other AI tools.
I even got to meet Corey Schafer, who's been making Python tutorials for years. He's one of the kindest and most humble human beings I've ever met!
If you ever searched for Python tutorials, you've probably watched his videos.
Of course, we were all half-joking about how doomed we are because AI is going to replace all data scientists and engineers. And soon nobody will learn programming or technical skills anymore. 😏
(quite reassuring when your entire career is built around teaching those things...)
But I actually think the fact that Google cares enough about tech creators to organize this first-ever summit says something...
Human scientists, engineers, and educators are still key. And the need for technical education is only expanding, even if what and how we teach is changing.
Maybe I just need to get smarter about making content using AI, lol.
Anyway, jokes aside, I came home with a few interesting insights that I wanted to share with you:
1. How to ACTUALLY keep up with AI tools
Like it or not, the AI landscape is changing every day.
Something that didn't work for you 3 months ago might already work surprisingly well today. 6 months in AI is practically a geological era.
The trouble is, it's impossible to keep up with everything everyday without having a mild existential crisis.
My approach is to revisit AI tools every month or two, try them on my own use cases, and stay aware of what's possible.
For example, I was recently asked by a recruiter to send an updated CV for a position I was considering.
My first instinct was to spend 2 hours putting together a (possibly crappy) PDF resume. Because the last time I tried using AI for this, the results were pretty 'mehhh'.
Then I realized... wait a minute. It's 2026. You can practically turn anything into an interactive web app these days.
So why not ask Gemini, Claude, or ChatGPT to build me a fancy interactive resume based on my LinkedIn profile and GitHub repos? It would be cool and fun to make.
My point is, don't let your past experience with AI become your permanent opinion of what it can do. If something didn't work a few months ago, give it another try. You might be surprised.
And no, I'm not suggesting to chase every shiny new tool or follow every AI guru on LinkedIn... I'm just saying, don't become so cynical that you stop experimenting altogether.
Stay curious and open-minded. We can't affort to be cynical 🥲.
2. You don't need AI for everything, but two ideas are worth trying
During the summit, a Google DeepMind fellow was demoing some pretty cool things you could build with Google AI Studio.
One of my favorite questions from the audience was: "What are the must-have AI agents that everyone should have?"
His answer was: (1) a research agent and (2) an information extraction agent.
Why?
Research agent: As humans, we can only read at a certain speed, and proper research can take days or even weeks. Having AI do that can really speed up that initial process. Of course, you still need to verify the information and draw your own conclusions.
Information extraction agent: If part of your job involves extracting specific data from PDFs, websites, invoices, etc. and pasting it into a spreadsheet, you should seriously consider automating that.
And I appreciate the honesty in his answer because, frankly, I'm getting tired of the whole "use AI for everything" mantra.
I think having clear principles about when to use it (and when not to) is important. Just decide what feels comfortable for you.
For example, for me:
Repetitive, tedious tasks: YES to AI.
Writing my original thoughts, ideas, or experiences (i.e. the source): NO to AI.
If you're like me and sometimes feel conflicted about how to use AI in your work or business, I'd highly recommend the book Return to Real by Ryan Levesque.
One of my favourite principles from the book sums it up perfectly: "Automate the distribution, NOT the source."
3. Get yourself out of your comfort zone
This last one has nothing to do with tech, but it's probably what I'm most proud of.
I'm terrified of public speaking! But then, 2 days before the event, I changed my mind and decided to volunteer for a 5-minute lightning talk. (I like living on the edge, lol.)
My talk was about how becoming a parent made me rethink what I create as a YouTuber.
I prepared quite a bit, rehearsing in front of the mirror and all. But once I got on stage, I was so nervous that I forgot most of it, including all my cool punchlines and jokes.
To make things worse, it turned out only a few people (3 to be exact!) in the room actually had kids. The rest were probably wondering why I'd decided to give a parenting talk at a developer summit 🫣...
I managed to finish, but walked off stage feeling embarassed. I felt like I'd just delivered the worst talk of my life. (I still refuse to watch the recording, lol.)
To my surprise, quite a few people came up to me afterwards and told me they really enjoyed it. My reaction was basically, "Thanks... but are you serious??"
It reminded me that we're often our own worst critics. We notice every little mistake, while other people might have a completely different experience.
I'm still terrified of public speaking, but I'm really glad I pushed myself to do it.
Sometimes you just need to get over yourself and do it.
And more often than not, it's not nearly as bad as you imagined. :)
Anyway, that's my little recap from NYC.
Awesome event and I came home exhausted, with lots of new ideas and connections.
And a recording of myself speaking in public that I may watch in ...10 years.
Master data science & AI skills, build awesome projects.
Join 7,000+ data professionals who are advancing their data and tech careers by learning from real-world projects, gaining first-hand experience, and accessing top resources delivered to your inbox every week.
Hi Reader, I'm in Vietnam for a few weeks visiting family. I'm enjoying it a lot. What I did not fully appreciate beforehand is how tiring it is to travel with a baby under 1 🫣. Last week, I climbed a mountain here with a bamboo walking stick. Here's me catching my breath at a viewpoint midway up. The place is called Yen Tu, one of those hidden gems most foreign tourists never hear about (you're welcome, and also please don't all go at once 🤣). I'd never climbed with a stick before. And wow,...
Hey Reader, If there's anything I've learned over the past years, it's that you have to advocate for your own work. No one else will do it for you... And yet, I think most of you aren't advocating for your work hard enough (I have evidence for this!) Let's say you've put in many hours learning something new and built a data or AI project, like a data analysis on a real dataset, an AI automation app, a dashboard, a machine learning model, or whatever. You get it working. You're proud of it,...
Hey Reader, Last year I worked with about 500 people in my program on building projects with AI and Python. To my surprise, the biggest challenge of all wasn't technical. It was: lack of TIME. ⏳ As much as I don't want to become a productivity self-help guru... (because I'm not qualified to give anyone advice on this 😅) ...I feel like I need to write about it anyway. Because "I don't have time to learn this" might be the single biggest thing standing between you and whatever you're actually...
