Pioneer.app is a fully remote accelerator. They fund exceptional individuals — usually outsiders — and provide them with a world class network, resources (cash/crypto/credits), a private inner circle, and help with the media.

When you win Pioneer you receive:

• $1,000 in cash + $6,000 in Stellar Lumens
• A free round-trip plane ticket (and help with associated logistics) to Silicon Valley.
• $100,000 in Google Cloud credits.
• $100,000 in AWS credits.
• Access to the private online Pioneer community.
• And a bunch of other stuffs: Find the complete offer here

→ More info on Pioneer here: New York Times article on Pioneer

You get funded by Pioneer — or “win” Pioneer — by submitting a project and competing in a global tournament against other aspiring Pioneers, submitting weekly updates that reflect how big a hustler you are.

You climb the leaderboard by receiving points from other competitors based on your weekly updates. The entire tournament runs using a peer to peer reviewing process.

It happens that while being in San Francisco this summer I spend a fair amount of time in hipster coffee shops writing my master plan. Part of it involved “winning” Pioneer, which means climbing to the top of the leaderboard and receiving a Pioneer offer from one of the experts reviewing competitor’s projects.

I am excited to announce that I have been funded by Pioneer recently and that the new Pioneers have just been announced.

I was waiting for the news to be made public to publish this blog post!

The project and the team

My Pioneer story is a bit different from the usual one.

I entered Pioneer with my differentiable physics engine (I still have a landing page here: https://magic.tf/). It is a good idea and I’ll probably keep working on it but another project of mine took off while I was competing.

I had been working with two friends of mine — Jay Yeung and Florian Muller — on an automated customer support software powered by Machine Learning (https://clearcall.app/).

Our one line pitch is that we turn voicemails into reports. Support teams of e-commerce companies can save time by not having to listen to hours of audio every morning. Instead, agents come back to work with reports and recommended actions (refund customer A, send tracking link to customer B) for every call that was received during the night.

We suddenly gained more traction and our product became good enough for us to show it to the world.

I thus did a “Pioneer Pivot” and added Jay and Flo as my two team members.

This led to a stratospheric rise in the leaderboard because we were building a project with real revenue as opposed to most competitors. We received our Pioneer offer three weeks after the pivot.

How to win Pioneer, an opinionated guide

I approached competing into the Pioneer tournament very scientifically.

The first thing aspiring Pioneers need to understand is that you are judged entirely on your weekly updates. Peer reviewers don’t see your project description, your past updates, or your team. Only experts do review this part of your application when you reach the top 50 in the leaderboard.

This means that writing good weekly updates must be your first priority if you want to win.

Good weekly updates will bring you to the top 100 but to reach the first spot (like we did with Clearcall) you will have to stand out.

From Bad to Good

Here is my first weekly update:

What are you building, in a sentence?

I am building a new kind of simulator to train AI for robots 1000x faster than the current methods. How? The simulator is differentiable.

What did you accomplish this week?

I got the fundamental equations of motions working in Julia (the programming language I am using for this project).I also experimented with propagating the gradient (the “signal” pointing to the thing you optimize for) through the simulator but it’s not quite working yet.

What do you want to have done by the end of next week?

I want to have the gradient propagation working (this will involve some serious engineering and hacks as the technologies I am using are not that mature) and use it to optimize some very simple control problems like balancing a stick on a cart (CartPole for the connoisseurs)

It’s bad. The sentences are too long, the vocabulary is too technical — remember that anybody can compete — and it just looks too dense.

When reviewing updates, competitors read up to 40 of them. You don’t want them to stop reading immediately because your huge paragraph is hard to digest. Make it easy for them.

Here is a better update:

What are you building, in a sentence?

I am building Magic, the first fully differentiable physics engine. Why is it a big deal? Check out https://magic.tf/ for more explanations!

What did you accomplish this week?

🎩 Built a website to explain the concept https://magic.tf/
🎩 Collaborated with the creator of a key component of the simulator on solving some problems with the gradient
🎩 Applied to YC and got 3 recommendations from alumnis
🎩 More progress on a first working version in Julia

What do you want to have done by the end of next week?

📌 Get a working version! Building Magic is complicated so I don’t know if I’ll be done in 7 day
📌 Get a video of the simulator simulating the dynamics of a simple system
📌 Write a blog post explaining how Magic works in depth
📌 Get more numbers on how fast it is

Less technical words — although “gradient” is still jargon — and fancy lists pleased my reviewers. My upvote/downvote ratio was 3 times better than with the first update.

Use lists extensively. Dumb down your message. Remove all the unnecessary words.

From Good to Great

Writing clear updates without jargon is a good start.

Most competitors already do that so it becomes hard for reviewers to decide which update is actually better. They often end up voting for the project that they think is better.

You don’t want that to happen. Reviewers do not understand your business or your project. They have not spent hours thinking about it like you did and that’s why a lot of really crazy projects without any hope of success and actual tangible work from the founders end up on top of the leaderboard. Reviewers vote for the projects that grab their attention!

A way to fix this problem is adding content to your weekly update.

When I started writing “extra information” on a Notion workspace and then sharing it in the update I saw a massive increase in our upvote/downvote ratio. The weekly update text box in Pioneer only accepts a really small amount of text so you can’t go through all the things you accomplished that week.

You thus want to summarise the high level achievements in a list inside the actual update.

Then, you add a link to some “extra materials” with screenshots, low level descriptions of the different goals you crushed during the week, and graphs if you have some.

Here is an example of one of our update with extra info:

What are you building, in a sentence?

Automated customer support for e-commerce companies w/ ML and NLP. $1k MRR. Growing Fast (1+ deal/week). Extra Info: https://bit.ly/2VGT9hf

What did you accomplish this week?

This is a Pioneer pivot! I am focusing 100% of my time to my other project which is blowing up

🎩 Got a grant from Google and The University of Edinburgh. 4000£.
🎩 Releasing the V2 of our UI. Screenshot: https://bit.ly/2IPPJ6G
🎩 Now taking 100% of the call of our biggest client

What do you want to have done by the end of next week?

📌 Release the new dashboard
📌 Do new user interviews with the leads we collected last week
📌 Redo the landing page
📌 Make a special page for Pioneer to explain the concept! (Much like I did for my previous project)
📌 Sign more deals!

Notice the link to Imgur with a screenshot of our new dashboard and the link to our Notion workspace with a pitch deck for the product.

In our case we knew that the product we were building did not look like such a great idea if you were not familiar with the problems our users were facing before being introduced to our solution. Adding our pitch deck with the classic problem/solution/market sections helped our reviewers understand the context of our projects.

From Great to 1st

All those tricks worked wonderfully and we reached the top 10 of the global leaderboard in a few weeks.

Still this is not enough to actually beat the rest of the competitors.

You need to go the extra mile, do something other people would not do.

That’s why we started including unedited videos of us going through the goals we accomplished during the week and demoing the new features we implemented live.

This works crazy well. The reason why is because you are being more authentic than everybody else.

Quoting Naval:

Escape competition through authenticity.

— Naval (@naval) March 20, 2018

Here are some of the videos we shot:

(We used Wistia and Loom extensively to host our video content)

Conclusion

Pioneer is great and you should join the tournament if you are working on any project, whether it’s art, research, or a business.

The weekly updates will incentivise you to be productive and the interactions with the other competitors will often lead to new friendships.

Sign up for the tournament here (with my referral link): https://pioneer.app/

-Justin