I've applied to YC 11 times over 5 years, which demonstrates both my persistence and how deeply I value what YC offers. Building startups is incredibly challenging - I see YC as a force multiplier that can significantly accelerate our growth through its proven methodology, network, and advice.

We charge a $20 subscription fee from individual engineers and a custom fee from enterprises starting at $20 per engineer per month. We also intend to have a free tier. At 1,000,000 engineers, we could make $240,000,000 ARR

Most open source robotics projects today are either linked to Github ( for software and file management ) + onshape ( for hardware designs ) or linked to a software documentation template (See K-Scale Lab's (YC W24) Documentation ). Closed source teams say they use tools like Notion and Confluence for their documentation. While these tools are great at what they do, they represent "hacky" ways to get the job done for teams building robots. Hence, they take significant time and effort.Engineers want to focus on what they do best ( building cool things ) and the tools that exist today force them to spend significant time describing things they have built already. We believe that there is a need for a new kind of knowledge management tool for robotics teams one that can be integrated deeply with tools they already use to do their work and automate the long and tedious documentation process.

Competitor 2. Software documentation templates.

See

We chose to work on improving robot documentation because we've personally faced the challenges it presents. 7 months ago, Olu and I attempted to build an open-source robot and found that existing documentation tools—primarily designed for software—were ill-suited for robotics projects. This made it difficult to understand the robot's design and functionality, leading to numerous questions and delays.Our co-founder, Naveen, is a robotics product designer, who has helped multiple robotics companies develop advanced industrial robotic arms and humanoid robots. In his experience, he has consistently found documentation to be one of the most daunting yet essential tasks.To validate the need for a better solution, we spoke with over 70 engineers who echoed our frustrations. They reported spending about 25% of their project time on documentation and an additional 10% answering questions from others who are less familiar with their work.Through our personal experiences and these conversations, we realized there's a significant demand for a specialized tool to streamline robot documentation. By addressing this pain point, we aim to save engineers time and improve collaboration within the robotics community.

Nvidia Inception Program for Startups.

We have 5 robotics companies in our pipeline ( 4 open source, 1 closed source ) giving us access to 2,000 engineers ahead of our launch next year. We expect to be at $250k ARR by June next year.

We have 5 robotics companies in our pipeline ( 4 open source, 1 closed source ) giving us access to 2,000 engineers ahead of our launch next year. We expect to be at $250k ARR by June next year.

Revenue came from the 25 users that paid 20 dollars to use our web-based robot explorer to learn about the Parol 6 robot arm and the RX1 humanoid robot.

Revenue came from the 25 users that paid 20 dollars to use our web-based robot explorer to learn about the Parol 6 robot arm and the RX1 humanoid robot.

These 25 users paid us $20 to get access to the explorer and learn about specific projects on our platform as they prepare to build it. They are majorly software engineers and product designers who have always had a knack for robotics but never had a clear and easy way to get started.

We have over 100 active users in our Discord community, 25 of which are now paying customers and actively use our most recently launched explorer ( They have used it more than 5 times in the last month ).

We've launched twice and are currently on a third iteration of the product. We generated $500 in revenue from our most recent launch, received feedback from engineers on what we need to focus on to make the product more valuable and now, we're working on a new iteration to serve engineering teams.

We understand the need to be in San Francisco to build our product

Olu and I have known each other the for more than 5 years. We've known Naveen for about 3 months. Naveen was one of our first and most active users.

express functions

argument

Node, and

It is worth pointing out again that there are other ways to get this information, e.g. instrumentation, core dumps debugging, etc. This is for those who can’t modify the application, but also can’t hinder performance. It’s also useful for maintainers who want to do this without deploying code changes.

P.S. I initially planned to walk through the AST at the start of this post and interpret it with the code. But, I think it’d be an interesting exercise for you to see if you understand how the Visitor pattern works. Fancy trying it?

I’ve built a small client for the evaluator that you can play with online, and you can find the complete explanck project on GitHub.

there might be slight variations on the indifferent books/articles

In the accept method for Literal we have visitor.visitLiteralExpr(this);. We know from Literal(9).accept(Interpeter()) that the current visitor is an interpeter object, therefore visitor.visitLiteralExpr(this); can be represented as Interpreter().visitLiteralExpr(this) for clarity. Also, as before this refers to the current object which is a literal object: Literal(9). So Interpreter().visitLiteralExpr(this) becomes Interpreter().visitLiteralExpr(Literal(9)).

In the Interpreter class, create a method to evaluate each type of expression and then use conditionals to call the correct method

implementation path

path

Binary(Binary(Literal(120), PLUS, Binary(Literal(12.3), STAR, Literal(53))), MINUS, Literal(9))

tiny evaluation language

This point is critical because it highlights how we enforce precedence. Remember that we were trying to resolve the term’s right operand from 7, but because we found ‘*’ (i.e. a factor expression) in the path at 10, we resolved it first!

Here’s a diagram detailing the parsing steps:

parse() gets called and it calls expression()expression() expands to term()At term(), the first thing we need to do is to evaluate the left operand. To do this: a. term() calls factor() b. factor() calls power() c. factor() calls primary().At primary, we see a NUMBER token at index 0, so we create a Literal expression and increment the index to 1. After that, we return Literal(53) up the stack to power()At power(), we check to see if we have a POWER operator at index 1, but there’s none, so we return Literal(53) back up the stack to factor().At factor(), we check to see if we have a STAR or SLASH operator at index 1, but there’s none, so we return Literal(53) back up the stack to term().We’re back at term() and find a MINUS operator at index 1, so we enter term’s while loop and increment the index to 2. To resolve the right operand, a. term() makes another call to factor() b. factor() calls power() c. power() calls primary()At primary(), we see a NUMBER token at index 2, so we create a Literal expression and increment the index to 3 ( which indicates the expression has been completely evaluated). After that, we return Literal(9) up the stack to power()At power(), we don’t check for any operator since the expression has been completely evaluated. So, we return Literal(9) back up the stack to factor().At factor(), we don’t check for any operator since the expression has been completely evaluated. So, we return Literal(9) back up the stack to term().Now, we are back at term. We have Literal(53) as the left operand, Literal(9) as the right operand and MINUS and the operator between them, so we create a Binary object like Binary(Literal(53), MINUS, Literal(9)) and return that to expression().

Here’s a diagram detailing the parsing steps:

%%> In this sample expression, an operator doesn’t exist, so there is no need to evaluate a “right operand”. Anyway, let’s focus.%%

// Checks if the next token is a particular token type // If it is, it moves the index by one step private boolean nextToken(TokenType type){} // Returns the most recently consumed token private Token previous(){}

| represents an if statement

As

Expr expression

\

expression and

so I won’t explain again

“books.”

A rule name is defined as a non-terminal because it’s not a final state. As in, it is not terminating. As in, it produces other things.A terminal is defines a final “state”. As in, it terminates. As in, it does not produce other things.

post-order traversal

Abstract Syntax Tree 🔗

it

search it on explorer

as in our final consideration for this essay, not the final problem Slack faced in this part of their timeline

We spoke briefly about the MySQL server being sharded by workspace. Say this mechanism created too many hotspots; how do we adjust it?

Answered Questions From The Space 🔗

This Flannel instance in that region that stores the snapshot is the instance that will respond to any future requests in that region from users in that workspace - in this way, we say Flannel has workspace affinity.

Issues

There’s no commitment, though