Hey friend!

My machine learning journey to where I am right now took around 10 years.

If I had a clue about what what I was doing I could have done maybe done it in 3 years.

But what if, absent unfair advantages, getting to good things in life DOES take 10 years?

Here is the problem.

When going into something new you cannot expect you will be able to find let alone identify expert advice that is worth following.

So it might be that indeed the median of getting to a happy place in a domain is indeed 10 years.

What consequences does this have for your life?

The answer is simple but many of us miss this.

If you can get to a good outcome in some amount of time, regardless of the field... all that makes a difference in the long run is the field that you chose!

I realize this sounds outrageous, but bear with me.

If I would have made a different decision 10 years ago, today I could be:

• an expert corporate manager
• an expert web developer
• an expert tile layer

Many of us see machine learning as this absolutely hardest field to get into.

Some of us often point to math and say "look at all the math I know! this is why I can do the things that you will never be able to! MUAHAHAHAHAHA".

But if we looked closer at web development, an unbelievable complexity would be revealed to us.

Distributed systems at every corner. Multiple languages (HTML, css, javascript, SQL) and complex abstractions (DOM, request lifecycle, css content box) you need to learn JUST to get started.

In data science the transition from R to Python is often a mountain impossible to scale.

And yet a web developer might use javascript, coffee and typescript (various flavors of js) on any single day and that is just one of the possibly 5+ languages they might be using! (Ruby for backend, SQL for optimizing db calls, HTML to display stuff, bash to deploy and automate stuff, CSS to style stuff)

So if machine learning is just as hard to learn as web development... which possibly might be as hard as learning to play music at a high enough level... or becoming an expert carpenter... where does this leave us with?

That the choice of the game you decide to play is really important!

Possibly even more so than how well you play a game.

In my home city, Lodz, machine learning engineers can live like kings. Web developers are aristocracy.

But folks who, like my wife, studied physiotherapy for 5 years, make less than a cashier!

This might sound trivial. And there is way more to being in machine learning than just good money.

The challenge.

The exciting pace at which new techniques are being developed.

The fact that we get to play with ideas on how intelligent behavior emerges.

But for many practical reasons, if you strive to be an employee, Machine Learning is one of the best fields you could have chosen.

None of the stringent regulation as in medicine.

Very little repetitive work.

Attractive job prospects.

All in all, this feels like a game worth playing.

My favorite things ❤️

🤖 Use text to edit images - My friend, Sanyam Bhutani, is working on a super cool project. It is an app that runs on a single GPU that allows you to edit images using text 😱. This is a fascinating project from a technical standpoint. I also love the hacker ethos of taking cutting edge tech and making it available on off-the-shelf hardware. Please find a demo below. A full walkthrough is coming to Sanyam's YouTube Channel soon!

🏆 How do people win ML competitions? - This is an important question, because very often the solutions that fare well in a competitive context also do well in real life. The State of Competitive Machine Learning is a report on the competitive landscape of 2022 -- what tools did people use? what techniques? who are the winners? It is a fascinating read. One thing that does stand out to me is how many first time and solo winners there are. Yes, winning a competition does take skill. But mostly it is hard work. Also, if you approach a competition properly, you can develop many of the skills needed on the fly. It does take time but for many, including myself, it has been 100% worth it. The report is a recommended read!

My new video 📺

Often you want to share some information, but not necessarily with the entire world.

Maybe you are building a knowledge base for your team. Or a blog you would like to make accessible to a certain community. Or you might be selling a course and creating a course site.

How do you limit access to a static site in the easiest way possible totally for free?

Turns out Hamel Husain has the answer and shares it with us in the video below.

Hope you enjoy the watch! 😊

​

Quote(s) of the Week ✍️

I overheard a statement to this effect from one of the most inspirational CEOs in tech and it resonated very deeply with me.

The productivity gains using tools like Copilot or ChatGPT are no joke.

Ethan Mollick @emollick 👀Two early papers find the effects of generative AI on knowledge work are completely unprecedented in modern history Separate studies of both writers and programmers find 50% increases in productivity with AI, and higher performance and satisfaction. And this is just the start. pic.twitter.com/r8TeOv0Uui March 3rd 2023 797 Retweets 3,480 Likes

Sure, the programming scenario in that paper is contrived (write a web server from scratch).

But the writing tasks aren't. And on top of that, those were people who were first exposed to these tools and it does take time and effort to get the hang of how to use them.

I am not an early adopter of any tech by any stretch of the imagination, but I do hope all this is evidence convincing enough for me to dive deeper into these applications!

After all, using these tools, not merely understanding how they work, might be pass to gainful employment within a year or two...