Tales from the jar side: Managing your manager and chess updates
It turns out Alireza Firouzja is good at chess, but he's no Leela Chess Zero.
Welcome to Tales from the jar side, the Kousen IT newsletter, for the week of Apr 12 - 19, 2020. This week I taught my revised Managing Your Manager course and my Basic Android course online on the O’Reilly Learning Platform, among other activities.
Managing Your Manager
This week I used my updated slides for my online course, which spend a lot more time discussing the basic problem of why conflict with your manager is inevitable. The idea is that since you’re a working professional, rather than a line manager, you care more about the technical details of your job, rather than making another dollar for the company. Managers — line managers, who make decisions about raises, promotions, and assignments, as opposed to technical leads on a project — care more about the money. There’s an overlap, of course, or why else would we be working together?
Because what you care about is different from what your manager cares about, sooner or later a conflict will arise. That’s natural and understandable, and you need to know how to deal with it. Of course, what the company would prefer is for you to do in that situation is what your manager tells you to do. Unfortunately, that will not get you what you want in the long run. Nor will the opposite — not doing what the manager tells you to do puts your job at risk.
Part of the message of the book I’m writing is to give you strategies to help navigate the conflicts in a way that builds the constructive loyalty relationship rather than damages it. Ultimately you want a relationship with the manager such that when they make a decision that affects you, they take your priorities into account, so that you get what you need when you need it. It’s not a trivial process, and the techniques I discuss won’t always work, but they at least give you a framework to try. Otherwise, when a conflict arises, most of us feel we only have two options: we can either just do what they say, or we can leave. Neither of those options gets you what you want.
I also changed the emphasis of one particular strategy — that of responsiveness. I run into this all the time. People who respond to requests in a timely way, even if they don’t know the answer, are viewed much more favorably than people who take too long to answer, even if they’re right.
Here’s the phrase to keep in mind:
A good answer today is much, much better than a great answer next week.
If your manager asks you a question and you don’t know the answer, don’t spend lots of time figuring it out. Instead, respond as soon as possible with “I don’t know,” followed by:
Here is what I do know.
Here is what I think the answer is.
Here are the steps I would take to find out the right answer.
The magic question: do you want me to spend the time necessary to figure it out?
Nine times out of ten or more, the response will be, “no, that’s fine. That’s all I needed.” I spend time in the book motivating why that works, and why we’re so reluctant to use it, but that’s the best approach I’ve ever found. It’s made a huge difference in my career.
Now that so many of us are working remotely, responsiveness becomes even more valuable. If your manager knows that you’ll respond quickly to any request, even if that response consists of “I don’t know” followed by the steps above, they’ll be happy with you. Responsiveness builds the loyalty relationship. Lack of responsiveness is very damaging.
Anyway, more when I write it all down.
Banter Blitz Update
Last week I talked about the upcoming Banter Blitz final chess match between the World Champion Magnus Carlsen and the 16-year-old prodigy, Alireza Firouzja. To give you an idea of how that played out, here are the round-by-round results in their best of 16 games match, from this excellent summary article:
Each column represents a single game, with 1 for a win, 0 for a loss, and 1/2 for a draw. So Firouzja won the first game, the second was a draw, Magnus won the third, and so on. Each time one of the players won a game, the other fought back right away. With Magnus down by 1, he won the 10th and 11th games consecutively and looked like he was going to win the match. But then Firouzja won both the 12th and 13th games to retake the lead. Then Magnus won again to tie, making the score 7 1/2 - 7 1/2. That set the stage for the final game, which Firouzja played brilliantly to win.
Magnus was unhappy with his play, but complimented the kid afterwards:
Ok, good game, Alireza! That was really horrible, I just was way out of shape, but he deserves full credit. I’ve just got to be better, but yeah, he’s amazingly strong! Full credit.
The match lived up to all the hype. The “Banter” part of Banter Blitz meant that both sides were providing commentary as they played, so you could watch one side or the other as the games went along. Here is the YouTube recording from Magnus’s point of view:
This was my favorite game, where the World Champion managed to deliver mate with 0.2 seconds left on his clock:
The reactions were great. Keep in mind, though, that this match didn’t actually mean anything other than the prize fund. It didn’t affect overall standings, or any title, or even any official tournament. It was for fun, but it sure was a lot of fun. :)
The amazing part is that the results even made it into the popular press:
Yes, that’s referring to this article from yes, that’s freakin’ CNN.
Since timing is everything, GQ (yes, that Gentleman’s Quarterly) also just published an interview with Magnus Carlsen.
I have to admit, the results made me a little sad. I really enjoy rooting for Magnus, both as a player and as a personality. I have nothing against Alireza, but he has his whole career ahead of him. It will be fun watching how he progresses over time. He’s obviously brilliant, but there is a big difference between blitz and standard. We’ll see.
Magnus Carlsen Invitational
The first chance for Magnus to get revenge will be tomorrow (Monday). The Magnus Carlsen Invitational started on Saturday, in a very different format. Each of eight players has a round-robin match against the others, consisting of four “rapid” games (15 minutes + 10 seconds per move), followed by tie-breaks. After that, the top four players will meet for the semifinals and finals.
The tournament is entirely online, so for Hikaru Nakamura, playing in the US, the games start at 10 am, while for Ding Liren, playing in China, the games start at 10 pm, and everyone else is in the middle.
Monday, Carlsen and Firouzja play their four games. I have no idea what will happen, but I won’t be surprised if the kid wins again. We’ll see.
Aren’t Computers Good At Chess?
If you’re into such things, you may vaguely recall that back in 1996 and 1997, the reigning World Champion Garry Kasparov played two matches against an IBM computer known as Deep Blue. Kasparov won the first match, but then famously lost the second, marking the first time a computer beat a World Champion under tournament conditions.
Deep Blue used a pretty standard alpha-beta algorithm, which basically searches millions of positions and evaluates them one by one. Since its opening “book” consists of all the current thinking on every opening every created, and the computer holds a “tablebase” of solutions for any endgame position involving six total pieces or less, the opportunities to beat it are pretty limited. To defeat a computer at that level, you have to force it into non-standard positions where human evaluation skills are better than a brute force search, and then never make a tactical mistake of any kind. That’s a tough challenge, and has only gotten worse in the twenty years since then.
At the time, some people thought the computer victory would be the death of chess. After all, tic-tac-toe has a sufficiently small number of possible positions that you can work out the best strategy for all of them, and there’s no point in playing any more if you know that.
Once during one of my summer jobs, I was sufficiently bored that I actually did that — I wrote out every possible tic-tac-toe position and move, to see whether it won or lost. But here’s the thing: I didn’t learn anything from it. There was no actual insight, other than knowing a few positions to avoid at the beginning. Sure, brute force searches work on small enough sample spaces, but people don’t think that way. Chess Grandmasters don’t search every single possible position. Their talent, experience, and intuition helps them quickly select which lines are most favorable, and then they dig deeply into those. It’s about understanding, rather than just memorization and trying every possible move.
Also, just because a computer can beat a person doesn’t mean people stop doing activities. A horse can run faster than a man, but men still run races. A car can go faster and farther than both, but just because cars exist didn’t put an end to track and field competitions.
A brute force search is, if you’ll pardon the expression, mechanical. I can use a machine to solve a problem, but it doesn’t replace the way we think.
At least that was all true until around 2018, when the artificial intelligence company DeepMind produced the deep neural network programs AlphaZero for chess and AlphaGo for the game of go. Each consisted of a generalized neural network that trained itself solely by self play, and within 24 hours they were stronger than any human alive.
The Go version is the subject of a documentary that you can watch on Netflix if you’re interested. The impact of both programs on the chess and go communities was roughly comparable to the impact of that asteroid that killed all the dinosaurs.
While AlphaZero was proprietary and no longer active (I have no idea why not), the open source community responded by creating Leela Chess Zero (abbreviated LCZero or Lc0) based on similar principles. You can download the program here if you’re so inclined. The reason I mention it is because also going on this week is the finals of the Top Computer Engine Championship, being fought between the current versions of LCZero and Stockfish, the best brute force program in existence. At the time of this writing, after 87 (of 100 planned) games Lc0 is leading Stockfish by a score of 44 to 43, where Lc0 has won 11 games, Stockfish has won 10, and the rest are draws. This “superfinal” consists of 90 minute games with a five second increment. The fun games are where the two engines disagree, and then we see which one was right. It’s also amusing when Stockfish announces something ridiculous, like mate in 43 moves. Yeah, okay, I saw that immediately too, as far as you know.
Ultimately it’s all about insight. With that in mind, last year Matthew Sadler and Natasha Regan produced the book Game Changer: AlphaZero's Groundbreaking Chess Strategies and the Promise of AI, which sold enormously well (for a chess book) and won every relevant award. Heck, they even got Garry Kasparov to write the introduction.
The closest I personally came to dealing with AI (other than back in the 90s, before AI was cool) is that my Tesla has a self-driving mode that I don’t entirely trust. It’s good at following lanes, and adjusting speeds based on the surrounding cars, but I don’t let it make lane changes for me. To be honest, it doesn’t drive like a person, and that’s a problem when you’re surrounded by cars that are driven by people. Someday maybe I’ll be able to get in my car at the airport, close my eyes and say, “take me home,” but we’re a long way from that point.
Maybe I’ll discuss some of the related issues in future newsletters. I do promise that this newsletter isn’t going to become all chess, all the time.
Friday Musicals
On yet another unrelated note, the past two weeks Andrew Lloyd Webber has been posting awesome video recordings of his stage shows through is channel The Shows Must Go On. A week ago Friday was Jesus Christ Superstar, and the other day was Phantom of the Opera. The catch is that each video is available for only 48 hours before it’s removed. Both the ones I mentioned are now gone. There will be a new one this coming Friday, however. I don’t yet know what he has planned or how many there will be in all, but they’re really good so far.
On a completely unrelated note, I should mention this program, for no reason.
This week:
Managing Your Manager, online on the O’Reilly Learning Platform.
Basic Android, same place.
Writing my Managing Your Manager book.
Next week:
Deep dive into Spring and Spring Boot, NFJS Virtual Workshop.
Groovy Podcast on Wednesday afternoon, with co-host Baruch Sadogursky and special guest Szymon Stepniak.
Finish book seven of The Expanse and start book eight. I really need to finish the series so I can find out what happens and get back to real work. :)
