Recent @mathgladiator Activity
As a fan of earlier Final Fantasy games, this is giving me chills. #soexcited.
The gap emerges because doing everything under is vastly different than doing strategic work across multiple organizations. It's the difference between captaining a small boat versus a super carrier.
I am, and part of the challenge is
(1) how do I leverage my position to do the next thing at massive scale
(2) do I give up on being charitable with resources at scale. For instance, I have my name on a donor wall at a zoo. I believe strongly in conservation and have donated a great deal of capital to various organizations throughout the country.
Ultimately, I think I have one final idea before I get out of big tech.
My golden handcuffs are sooo heavy. I'll wipe my tears with a hundred and then burn it to feel something.
This is the kind of stuff that inspires me and has me yearning to escape big tech to make again for sheer joy.
I've been thinking about two interesting problems.
First, differentiating code to build a client side predictor with privacy as a consideration. I have code that describes how to translate domain messages into state changes, and I'm trying to figure out how to predict the effects of sending a message on a client even though the client has imperfect knowledge.
Second, AI for games with imperfect information. Specifically, how to build an AI for Battlestar Galactica.
These are in context of http://www.adama-lang.org/
> Does it mainly happen at global corps that run essential internet infrastructure?
Being on-call happens at scale because machines are not to be trusted, and the limits of our skulls meet their match when machines vastly outnumber people. When I worked at Amazon S3, we had rotations where people had physical pagers. On-call was a duty, and everyone shared a rotation.
Honestly, I thought it was an awesome responsibility that there I was at 2 AM and some stupid shit happened in dublin and I got to fix it. (Or at least try to fix it as my first on-call shift was rough)
I think Amazon is one of the few companies that get this right with their cultures because you can own efforts that people use. It's not going to be something discarded after a year unless a massive rewrite is planed. The massive rewrites can be super fun (but hard) as well because you have to do it in-flight.
The reason Amazon gets this right is that you can get a bit of the stake on the outcome, and this is awesome. For instance, I designed how Amazon S3 does URL rewrites on the website endpoint. It's a small feature, but I lead the effort from start to finish. For all practical purposes, that feature is most likely going to last until the end of time. However, its that relationship between myself as an engineer with the customer that makes it worth it (at least emotionally).
Here is the poem:
I will not do this PSC. There is a future that I foresee. No longer shall I bend the knee to this process called PSC.
So, recently I achieved happiness because I just don't care about the job anymore. I didn't even do my performance self review, and I wrote a poem instead. I got a Meets-Most, and I've never been happier.
I'm writing my final proposal as one final swing at the bat and the title is "XYZ or Bust" where either I get buy-in or I'm leaving. I've never felt better in my entire career.
This is made possible because I'm old enough with enough assets that I can quit and retire, so now it is time to leverage the positives.
From reading this, it is clear the team he is on is toxic. The key to not being on-call all the time is not answering the call unless you in rotation. It's hard to do for sure.
Given the Principals only ask, this is interesting. There is something humble about asking for mentor, and this is something that I wish I had when I was younger.
I wonder how far I could have gotten if I embraced asking for help sooner instead of just tinkering away and relearning old lessons.
There is a level of mastery, but it is borderline useless. I've been called an architecture astronaut (ref: https://www.joelonsoftware.com/2001/04/21/dont-let-architect... )
And the thing here, software is simple. You take bits in, do stuff, and then its bits out. State machines there. Disks over there. Algorithms in hand. There is no real end to understanding all the details, but the game is well understood in a very reductive manner.
There is the game that we study which is making these infernal machines do useful things. Then, there is the meta game of getting humans to do things with these infernal games. These games are infinite and mastery demands you to see them for the games that they are.
In this, there is no plateau. You can operate with peak resources to explore, but the game is the game to play.
When I switched from ANTLR to hand written for Adama ( http://www.adama-lang.org/ ), I felt way better about things. I was able to get sane error messages, and I could better annotate my syntax tree with comments and line/char numbers.
A killer feature for a parser generator would be the ability to auto-generate a pretty printer which requires stuffing comments into the tree as a "meta token".
I've had a position for a while that we need to rethink our databases such that privacy is a first class citizen to minimize the potential space to "hook into" sensitive information.
Databases have access control mechanisms that scale to an organization, but web scale demands access control for the world which means that every developer working with a DB must invent their own access control to the data on behalf of their users. This basically means the database is open to all developers in a company.
Given the emerging regulatory environment, I suspect there is room to launch new database technology such that access control is an atomic feature built in.
I've experimented with this using my programming language for board games, and it is surprisingly amazing. I can model data schema and privacy at the same time, and I love it. http://www.adama-lang.org/
I was there, and then I started to do weed.
Every day was the same without any enjoyment, and I considered just throwing myself in front of a bus. I tried some drugs from the doctor, and I lost control of my mind. It wasn't pleasant. The process for diagnosis was laughable. So, I tried weed for the first time.
Now, weed wasn't a magic fix, but I did a few bong hits and I didn't feel anything at first. Then, I went to drink some water and it was as if I saw the radiant beauty of glass and light interact for the first time. There I was standing in my kitchen staring at the bottle of a glass. I was in awe.
Now, I don't really need weed anymore, but I did develop a sense of awe of the wonder of reality. It's kind of amazing to think about things like why do we exist.
Purpose is how we distract ourselves or amuse ourselves, but curiosity is how we can find gratitude for the positive things in life. Here we are. We are a part of the universe that can observe the universe. What the fuck are we? What is consciousness? What does it mean to be alive?!?
I give a fuck, but I also don't know how to help. Most people barely know how to take care of themselves because we live in a chaotic world. We are frail beings living fragile lives in a psychotic world.
I've been suffering for a long time for a variety of reasons.
The reason you suffer may be to due to expectations or desire.
I feel like I'm not longer suffering since I let go and started to embrace that life is a journey. What matters is the present, but it is hard to recognize that when we can look forward.
Underneath this is the hard truth that streaming is hard and costly. I've run into many people with optimistic dreams around how to take an SQL or other query languages and then make it update based on data changes rather than periodic polling.
These systems are always expensive and yield marginal improvements when compared to a dumb publish subscribe system. I've got notes on publish subscribe systems, and I am publishing a paper next month about the end result system that I built over five years. For a clue about the midpoint, my SRECON talk addresses this: https://www.usenix.org/conference/srecon17americas/program/p...
A key note is that publish/subscribe over-commits the problem space (i.e. the problem space can potentially be quadratic depending on use), but if you give up on durability then your costs can go way-way-way down.
The lack of durability is a problem, BUT it can be leveraged to discover non-marginal use-cases such that further investments make sense. Unfortunately, once you invest in the imperfect system such that it operates at five+ nines during normal operation, then very few query based systems are competitive.
There is one more level where you hate everything, but then embrace loving the people that make shit.
I bought a new truck, and after putting 100 miles on it the transmission started acting strange. I take it in, and they don't know what is wrong with it. So, they are going to replace the entire transmission next week.
Now, I could say "oh, why did I buy that brand" or be like "lolz, these people are n00bs". Instead, I realize that they are making something at a massive scale and bad things will happen, so I'll accept the inconvenience and move on with my life. Shit happens, and that's life.
I've hated software for a long time because I value precision and reliability, but these are exceptionally hard to scale organizationally. Suppose you fix reliability today, well tomorrow's feature could just fuck it up.
Now, I still hate software, but I come at with a mindset that software is an artifact of a culture which requires nurturing and love.
I went into big tech to build up a warchest, and now I have no desire to build a company since I don't want to hire anyone. I'm going to "retire" and do some solo products around board games. I've got a secret weapon: http://www.adama-lang.org/
I would use it, but given my eclectic nature would need an backup button.