Hey I know something about this! I just mailed my organization's RRF-1 a couple of days ago. The author of this post seems to be confused. My organization is on the same fiscal year as OpenAI, and our RRF-1 had to be mailed by November 15th. That explains the supposed "six month" delay. Second, if it's mailed on November 15th, it might not have even been received yet, let alone processed. This post feels like grasping at straws on the basic facts, setting aside the fact that it just doesn't make any sense to imagine a board member filling out the RRF-1 and going "oh wait, was there financial fraud?" the morning of November 15th. (That's ... not how the world works? Under CA law, any nonprofit with 2M of more in revenue has to undergo an audit, which is typically completed before filling out the 990, and the 990 is a pre-req for submitting the RRF-1. That's where you'd expect to catch this stuff, and the board's audit committee would certainly be involved in reviewing the results well in advance.)
The six-month delay is probably due to an automatic extension if you get an extension from the IRS, and also, you can file the form electronically, in which case mail delays are not a problem. But neither of those issues is the point. The point is that the form needed to be filed at all, and representations needed to be made accordingly.
OpenAI handled their audit years ago and hasn't had another one since according to their filings. So that does not seem like it would have been an issue this year.
Take a look at the top of the RRF-1 for the instructions on when it's due. Also, the CA AG's website says that OpenAI's was due on May 15th. They just have been filing six months later each year.
I'm the co-founder and CTO at U.S. Digital Response (https://usdigitalresponse.org/) - we're a non-partisan non-profit that started 2 years ago to help state and local governments and non-profits around the US address the challenges from COVID.
We've expanded our work beyond COVID to support these organizations addressing the critical needs of the public and bring the best technology has to offer to these problems. We've partnered with more than 230 government and nonprofit partners on 300+ projects, impacting more than 42 million people across 36 states and territories. As one example, we helped the city of Memphis and Shelby County distribute almost $14M in rental assistance (https://www.usdigitalresponse.org/case-studies/keeping-famil...) in a matter of weeks at very low cost.
One of the things I like about this work is that we are tied directly in with local experts who understand the problems very well, and we can bring the lens of understanding how modern tech can be helpful, rather than trying to lead with solutions.
In doing this work, we've discovered a number of common problems that have scalable solutions that we're working on in both access to benefits and election administration (e.g. our poll worker management tool - https://www.usdigitalresponse.org/projects/poll-worker-manag...) and are starting to scale our team to tackle these challenges.
I think my favorite insight from this post is that the difference between founders and others is the bias to action to go after an idea that seems like a good one, and take it where it may go.
(original author here) I got inspired to write this post after helping contribute a few chapters to The Holloway Guide to Technical Interviewing and Recruiting, which is in early access and has a more material that might be interesting+relevant to you: https://www.holloway.com/g/technical-recruiting-hiring/about (It goes beyond just interviewing, as well, to cover the whole hiring process end-to-end)
I wouldn't say that it felt like I was thinking really hard in times I wasn't writing. It's a bit like the technique of putting down a crossword for a little while when you get stuck - you might not be thinking much about the puzzle, but somehow you're turning it over in your head.
FWIW, I do agree with the original comment at the top of this thread that this technique couldn't yield 48 books in 2 years - there's only so much headspace available, and this technique is one that taps into that headspace effectively for long-term projects, but it's not magic.
I've been using a technique similar to what I described in the original post to teach myself Haskell, and your comment 100% resonates with me. In the case of learning a new language, just getting a piece of code to run and testing it out has been a great way for me to get into the groove. I suspect that writing unit tests might be another good entry point.
It's worth noting that you are better off tax-wise saving one dollar than earning the same dollar. I didn't expect it to impact your argument that much, but given that the 2010 income tax rate in Israel is 44% for anyone making over ~90K/yr, 2K in savings is roughly equivalent to 3.5K in marginal income. From what I gathered last time I was in Israel, that was a pretty significant amount of income.
On the topic of food...it's interesting that Tabasco is specifically mentioned. I'm originally from Louisiana, and my family and I prefer Crystal hot sauce; we think it has more flavor.
Did you mean to compute the cost/user? If so, that comes out to a little over 13c/month. Still, that comes to $1.60/year, meaning that a user would start to cost money right around the end of the 6th year of usage (assuming hosting costs don't go down).
But if they're not paying for archival, are they really the folks who are causing the high hosting costs? As other have speculated, it would seem that the archival service might be the major reason that hosting is so expensive.
