This is a poor analogy. Magnus Carlsen stays because chess consumers decide to pay for humans even though they are inferior to Stockfish. BigCorp will always pick machine over you if they can.
I agree, that was a weak analogy. Magnus stays employed because chess fans value watching humans compete, not because engines didn't replace his capabilities.
I've updated the post title to "Train with coding assistants like Magnus Carlsen trains with chess engines" to focus on the main point: the methodology. Magnus uses chess engines as sparring-partners to improve his game after matches. Same can be done by developers who will use coding assistants to level up their skills.
Companies automate the parts that are commodity. On messy product work (drifting specs, integration, liability), human + AI + good process > AI alone. The machine proposes; the human sets goals, constrains risk, writes/reads tests. That combo ships faster and with fewer costly mistakes than letting an ai free-run.
Productivity replaces people, if you get more done from a team of 5 than your old team of 10 you generally fire 5 people.
Programmers have significantly higher unemployment than the general workforce today, but it’s hitting a wide swath of white collar jobs and that’s not going away. The industrial revolution replaced manual labor, so people moved to more mentally challenging jobs but AI can eventually replace anybody from CEO’s on down.
Demand isn’t doubling economy wide any time soon. So you might keep a team of 10 if you’re outcompeting a different team of 10 and getting all of them fired.
Even if you’re keeping your job expect huge downward pressure on wages.
Computer ratings are kind of random because there's no meaningful sample of them playing top players. They're obviously much stronger, but specific numbers are kind of meaningless beyond comparing them to other software which they have competed against.
But Stockfish is not going to beat Magnus with queen odds. Nakamura has beaten Lc0 with knight odds in blitz (though he got crushed overall), and queen odds in bullet, all while chatting on stream. And fast time controls favor computers simply because they're practically impossible to flag and will almost never miss a tactical idea.
So you would enjoy playing Magnus more than playing Stockfish. Because it would actually give you a chance. And that's the whole point of playing games: to enjoy them, and to both have a chance at winning, even if that chances is remote. Stockfish takes that joy away. For me, that makes Magnus far superior to Stockfish. If I want to lose a game playing with myself I'll play solitaire. Or better yet, I'll go read a book. But the chance to interact with a human at the peak of their skill would be unbeatable for me in terms of enjoyment, far to be preferred over playing a game with a computer program. Though I can see the satisfaction in programming a computer to be that good, I would not enjoy the game because the other side would not enjoy it either.
>So you would enjoy playing Magnus more than playing Stockfish.
Oh, no, I wouldn't enjoy playing him more. I'd enjoy playing Stockfish infinitely more because I could learn new strategies from it. Knowing you're going to lose allows you to learn from all moves and all mistakes.
Against Magnus, I'd have no chance. I'm barely over a 2000 rating. Magnus could play with queen odds and drunk off his ass while blindfolded and would wipe the floor with me.
One thing that might not be apparent to non-chess players is that an experienced human (particularly a GM or Super-GM) with an engine can often beat the same engine or another engine that lacks human assistance. There are some positions, particularly in closed games where this can become more of a factor. It'll be interesting to see if a similar logic plays out in other fields. I imagine that some companies may be quick to automate away roles to save money, however, if you follow what we've learned about chess, there's likely an opportunity to make a bet (start a company) on AI-assistance outperforming full AI automation in some domains.
Centaur Chess (the term for this) used to be better but outside of some positions that I don't think have ever occurred in an actual game, no human can help out a modern chess engine.
We obviously wish for this to be true so we must be careful with our own bias. It's not clear at all that "centaur" teams are better at chess than computer alone, are there actual tournaments where this is tested?
I remember this was a topic in early 2010's and then it was said that the human team member was already not contributing much anymore. In a blitz game it would most likely be detrimental.
Chess is a closed system w/ finitely many rules. I don't think success in closed domains transfers to open ones & this is the usual error optimists are making about AI & its success in closed domains like chess. Continuous learning is necessary for open domains b/c real world distributions are constantly changing & unless deployed systems can keep up w/ the changes their performance will continuously degrade until they're worse than random. Chess has been the same for millenia & that's why chess is essentially solved.
> Code review becomes your post-game analysis. Magnus reviews his games with engines to learn from their superior analysis. You review LLM code to ensure it's actually correct.
If AI coding assistents really worked like Chess engines, they would review your code, pointing out issues and suggesting improvements.
I guess the purpose is difference? You play chess because you want to play chess, as a fun thing (and also sometimes more serious thing), but a lot of us write code for something else, the purpose is not the code itself, but what it accomplishes, the code just happens to be a tool, so it's not necessarily the activity you want to do, it just unlocks/improves some other activity.
I think it'd be fairly easy to create a project that uses an LLM to effectively review your code, if that what's you want. But personally I'd rather do the opposite, describe what sort of code I'm looking for, how it's supposed to work, and what the goal is, and then not having to do the actual typing itself.
Yes, chess engines review your games and point out blunders. They also suggest moves you'd never consider. Like when you're analyzing a position and the engine recommends a move that flips the eval from -1.5 to +1.8. Similarly, coding assistants might suggest a solution you'd never considered.
Fair point but Magnus wasn't born Magnus. He became world champion partly by doing what the article describes: using engines for post-game analysis. A 1200-rated chess player who reviews their games with an engine improves faster than one who doesn't.
If I'm understanding this right his ELO was 2064 in 2001 when he was 11. I don't think he was using engines for post-game analysis at 11 or maybe he was, but he was certainly born better at chess than most people.
> Carlsen showed an aptitude for intellectual challenges at a young age. At two years, he could solve 50-piece jigsaw puzzles; at four, he enjoyed assembling Lego sets with instructions intended for children aged 10–14. His father, a keen amateur chess player, taught him to play at the age of five, although he initially showed little interest in it https://en.wikipedia.org/wiki/Magnus_Carlsen#Childhood
I don't know you could go as far to say he was born better at chess, but he certainly seemed to be smart early on, and with his father's influence and with a supporting environment that smarts got aimed at chess.
Some people have that talent, others do not. And then those who do have that talent and put in the hours studying theory can go very far without ever looking at a computer to help them with analysis.
Carlsen is that once-per-generation exception, someone who has a talent that somehow allows him to see a little further down the game tree than others. I grew up next to someone who was nearly that good, but there are so many people in that bracket that outside of NL he's not very well known. The distribution of skill at the high end of chess is highly non-linear.
everyone is the Magnus Carlsen of their own life, though. and humans are irreplaceable. sure, budget decisions are made that cause people to have to go find another employer. but there is no civilization without people in it.
> budget decisions are made that cause people to have to go find another employer
When those decisions are made en masse, incomes and even chosen careers can change dramatically. This is true even when there isn't massive changes in technology. I started my career during the time when many ended up leaving tech and not returning due to how competitive the job market became.
It's irrelevant who you are to yourself if you're nothing but a row in a database to the people looking for the earliest opportunity to replace you with a cheaper option. Those people are not acting in the best interests of long-term civilization.
Mildly OT: but which one is the best (offline) frontend for Stockfish? Or just what do HN people use? There are so many options [0] from free to paid I wish there was a big comparison table somewhere
This is like saying automated delivery drones haven’t replaced Saquon Barkley, so they won’t replace you (delivery driver).
Imagine if you will that chess engines don’t exist but apps do. A chess game company might employ a small army of excellent chess players to play games against their customers. You load up your app, request a medium difficulty game, and get connected with James, who’s playing 10 games at once for $20/hour. It’s a decent gig. Customers pay a fee for the experience but it’s worth it to them.
But now let’s say Google publishes a groundbreaking paper that lays out the foundation of an automated chess engine. Should James be worried about his job? Carlsen’s going to be fine. I’m not so sure about James.
If people wanted to see "perfect" chess they would be watching chess engine championships. These feature some of the deepest and most thought out moves ever performed in chess. They also get a few hundred live viewers.
Magnus is objectively an inferior player to top chess engines. If there was a technological problem that requires the best chess abilities, nobody would be paying Magnus to solve that problem, they would be using a chess engine.
The analogy does not work at all, because Magnus is not paid to be a good chess player in absolute terms, but a good chess player relative to other human top chess players.
Not everything that draws a crowd is a performance. I suspect Magnus Carlsen would play chess even if nobody watched. Just look at Mikhail Tal, Max Euwe, Garry Kasparov and all those others. None of them did it because of the crowds. They did it because the game itself fascinated them.
I agree with the conclusion but the premise is bunk. AI won’t replace you because it’s unsustainable and grossly overhyped, not because it’s as effective at general tasks or code review to the same extent that computers can play chess.
And to what end? I sometimes get the impression software people want to optimize out everything so in the end, when there's nothing left to do, they can lay down and die.
It wasn't that long ago that if you worked at an engineering firm, you needed a whole floor or floors of skilled draftsmen to produce the engineering drawings of your bridge, or your widget. Now an intern can do that in seconds with modern CAD software.
The same phenomenon is already occurring with AI. The concern is not that it will replace people, it's that it will make people so productive that it will create an oversupply of workers and thus kill a lot of jobs.
While I agree that AI wont replace you, the analogy is BS. Magnus Carlsen is not replaced by a chess engine because he is not paid solely because of his raw skills. As any professional player of any massive sport, he is an entertainer first. He gets paid only because people want to watch him play.
My go to phrase for the equivalent observation is "If they ever make a machine that can make a coffee at the press of a button then all the coffee shops will have to close" /s.
This is a poor analogy. Magnus Carlsen stays because chess consumers decide to pay for humans even though they are inferior to Stockfish. BigCorp will always pick machine over you if they can.
I agree, that was a weak analogy. Magnus stays employed because chess fans value watching humans compete, not because engines didn't replace his capabilities.
I've updated the post title to "Train with coding assistants like Magnus Carlsen trains with chess engines" to focus on the main point: the methodology. Magnus uses chess engines as sparring-partners to improve his game after matches. Same can be done by developers who will use coding assistants to level up their skills.
Thanks for calling this out.
Yeah. Coding is not a sport. Even if it is (Leetcode competitions or something), as in chess, it is the top 100 or so that can make money and survive.
Companies automate the parts that are commodity. On messy product work (drifting specs, integration, liability), human + AI + good process > AI alone. The machine proposes; the human sets goals, constrains risk, writes/reads tests. That combo ships faster and with fewer costly mistakes than letting an ai free-run.
Productivity replaces people, if you get more done from a team of 5 than your old team of 10 you generally fire 5 people.
Programmers have significantly higher unemployment than the general workforce today, but it’s hitting a wide swath of white collar jobs and that’s not going away. The industrial revolution replaced manual labor, so people moved to more mentally challenging jobs but AI can eventually replace anybody from CEO’s on down.
Or you keep your team of 10 and produce more things.
Demand isn’t doubling economy wide any time soon. So you might keep a team of 10 if you’re outcompeting a different team of 10 and getting all of them fired.
Even if you’re keeping your job expect huge downward pressure on wages.
I've never worked at a place where I have any shortage of work to do. Usually the roadmap is several years out.
Meanwhile, I’ve been let go several times for finishing major projects. Steady long roadmaps depend on slow moving projects.
> BigCorp will always pick machine over you if they can.
But, people might not always prefer BigCorp over humans, if they can?
> even though they are inferior to Stockfish
They're not.
Magnus's peak rating was 2882.
Stockfish is currently rated 3644.
Lc0/AlphaZero was estimated to be rated 3800.
Stockfish would destroy Magnus even with queen odds.
Computer ratings are kind of random because there's no meaningful sample of them playing top players. They're obviously much stronger, but specific numbers are kind of meaningless beyond comparing them to other software which they have competed against.
But Stockfish is not going to beat Magnus with queen odds. Nakamura has beaten Lc0 with knight odds in blitz (though he got crushed overall), and queen odds in bullet, all while chatting on stream. And fast time controls favor computers simply because they're practically impossible to flag and will almost never miss a tactical idea.
Who would you rather play: Magnus or Stockfish?
And what made him inferior again?
Magnus.
With him being human, there's a 0.00000000000000000000000000000001% chance that he might miss something or miscalculate deep into a game.
With Stockfish, those odds are quite literally 0%.
Magnus could make a mistake. Stockfish is literally programmed to not make a mistake.
So you would enjoy playing Magnus more than playing Stockfish. Because it would actually give you a chance. And that's the whole point of playing games: to enjoy them, and to both have a chance at winning, even if that chances is remote. Stockfish takes that joy away. For me, that makes Magnus far superior to Stockfish. If I want to lose a game playing with myself I'll play solitaire. Or better yet, I'll go read a book. But the chance to interact with a human at the peak of their skill would be unbeatable for me in terms of enjoyment, far to be preferred over playing a game with a computer program. Though I can see the satisfaction in programming a computer to be that good, I would not enjoy the game because the other side would not enjoy it either.
>So you would enjoy playing Magnus more than playing Stockfish.
Oh, no, I wouldn't enjoy playing him more. I'd enjoy playing Stockfish infinitely more because I could learn new strategies from it. Knowing you're going to lose allows you to learn from all moves and all mistakes.
Against Magnus, I'd have no chance. I'm barely over a 2000 rating. Magnus could play with queen odds and drunk off his ass while blindfolded and would wipe the floor with me.
With respect to winning a game of chess?
With respect to playing a game of chess.
What do you mean they are not?
One thing that might not be apparent to non-chess players is that an experienced human (particularly a GM or Super-GM) with an engine can often beat the same engine or another engine that lacks human assistance. There are some positions, particularly in closed games where this can become more of a factor. It'll be interesting to see if a similar logic plays out in other fields. I imagine that some companies may be quick to automate away roles to save money, however, if you follow what we've learned about chess, there's likely an opportunity to make a bet (start a company) on AI-assistance outperforming full AI automation in some domains.
Centaur Chess (the term for this) used to be better but outside of some positions that I don't think have ever occurred in an actual game, no human can help out a modern chess engine.
We obviously wish for this to be true so we must be careful with our own bias. It's not clear at all that "centaur" teams are better at chess than computer alone, are there actual tournaments where this is tested?
I remember this was a topic in early 2010's and then it was said that the human team member was already not contributing much anymore. In a blitz game it would most likely be detrimental.
I think this was true before AlphaZero; now, I'm not sure there's any benefit to having a human + engine versus just an engine alone.
From my understanding the AlphaZero based engines (i.e. neural nets) do not really benefit from having a human in the loop.
Chess is a closed system w/ finitely many rules. I don't think success in closed domains transfers to open ones & this is the usual error optimists are making about AI & its success in closed domains like chess. Continuous learning is necessary for open domains b/c real world distributions are constantly changing & unless deployed systems can keep up w/ the changes their performance will continuously degrade until they're worse than random. Chess has been the same for millenia & that's why chess is essentially solved.
This is advanced/cyborg chess in a nutshell (https://en.wikipedia.org/wiki/Advanced_chess)
> Code review becomes your post-game analysis. Magnus reviews his games with engines to learn from their superior analysis. You review LLM code to ensure it's actually correct.
If AI coding assistents really worked like Chess engines, they would review your code, pointing out issues and suggesting improvements.
I guess the purpose is difference? You play chess because you want to play chess, as a fun thing (and also sometimes more serious thing), but a lot of us write code for something else, the purpose is not the code itself, but what it accomplishes, the code just happens to be a tool, so it's not necessarily the activity you want to do, it just unlocks/improves some other activity.
I think it'd be fairly easy to create a project that uses an LLM to effectively review your code, if that what's you want. But personally I'd rather do the opposite, describe what sort of code I'm looking for, how it's supposed to work, and what the goal is, and then not having to do the actual typing itself.
That's definitely possible, just not perfect but they can catch/point out some bugs.
Yes, chess engines review your games and point out blunders. They also suggest moves you'd never consider. Like when you're analyzing a position and the engine recommends a move that flips the eval from -1.5 to +1.8. Similarly, coding assistants might suggest a solution you'd never considered.
Both teach something new.
AI won't replace you ... if you are the Magnus Carlsen of your field.
How you will get experience to grow into the Magnus Carlsen of your field is an open question, however.
We're not all Magnus Carlsen. Some of us are hacks or very ordinary 'players', and can be replace without a ripple on the pond.
I'd be happy just being Hikaru, whose busy seal clubbing his way to the spot on the Candidates.
Fair point but Magnus wasn't born Magnus. He became world champion partly by doing what the article describes: using engines for post-game analysis. A 1200-rated chess player who reviews their games with an engine improves faster than one who doesn't.
He was a grandmaster at age 13. In 2003.
If I'm understanding this right his ELO was 2064 in 2001 when he was 11. I don't think he was using engines for post-game analysis at 11 or maybe he was, but he was certainly born better at chess than most people.
https://ratings.fide.com/profile/1503014/chart
I love the original article. Hey guys you just have to be the best programmer in the world to not be replaced by AI!
> Carlsen showed an aptitude for intellectual challenges at a young age. At two years, he could solve 50-piece jigsaw puzzles; at four, he enjoyed assembling Lego sets with instructions intended for children aged 10–14. His father, a keen amateur chess player, taught him to play at the age of five, although he initially showed little interest in it https://en.wikipedia.org/wiki/Magnus_Carlsen#Childhood
I don't know you could go as far to say he was born better at chess, but he certainly seemed to be smart early on, and with his father's influence and with a supporting environment that smarts got aimed at chess.
Some people have that talent, others do not. And then those who do have that talent and put in the hours studying theory can go very far without ever looking at a computer to help them with analysis.
Carlsen is that once-per-generation exception, someone who has a talent that somehow allows him to see a little further down the game tree than others. I grew up next to someone who was nearly that good, but there are so many people in that bracket that outside of NL he's not very well known. The distribution of skill at the high end of chess is highly non-linear.
everyone is the Magnus Carlsen of their own life, though. and humans are irreplaceable. sure, budget decisions are made that cause people to have to go find another employer. but there is no civilization without people in it.
> budget decisions are made that cause people to have to go find another employer
When those decisions are made en masse, incomes and even chosen careers can change dramatically. This is true even when there isn't massive changes in technology. I started my career during the time when many ended up leaving tech and not returning due to how competitive the job market became.
It's a big problem if they can't find another employer fast enough or if the new employer doesn't pay enough.
It's irrelevant who you are to yourself if you're nothing but a row in a database to the people looking for the earliest opportunity to replace you with a cheaper option. Those people are not acting in the best interests of long-term civilization.
[dead]
Nobody is paying money to watch you code
Mildly OT: but which one is the best (offline) frontend for Stockfish? Or just what do HN people use? There are so many options [0] from free to paid I wish there was a big comparison table somewhere
https://official-stockfish.github.io/docs/stockfish-wiki/Dow...
https://github.com/mbiesiad/awesome-chess
This is like saying automated delivery drones haven’t replaced Saquon Barkley, so they won’t replace you (delivery driver).
Imagine if you will that chess engines don’t exist but apps do. A chess game company might employ a small army of excellent chess players to play games against their customers. You load up your app, request a medium difficulty game, and get connected with James, who’s playing 10 games at once for $20/hour. It’s a decent gig. Customers pay a fee for the experience but it’s worth it to them.
But now let’s say Google publishes a groundbreaking paper that lays out the foundation of an automated chess engine. Should James be worried about his job? Carlsen’s going to be fine. I’m not so sure about James.
Magnus Carlsen is a performer.
If people wanted to see "perfect" chess they would be watching chess engine championships. These feature some of the deepest and most thought out moves ever performed in chess. They also get a few hundred live viewers.
Magnus is objectively an inferior player to top chess engines. If there was a technological problem that requires the best chess abilities, nobody would be paying Magnus to solve that problem, they would be using a chess engine.
The analogy does not work at all, because Magnus is not paid to be a good chess player in absolute terms, but a good chess player relative to other human top chess players.
> Magnus Carlsen is a performer.
Not everything that draws a crowd is a performance. I suspect Magnus Carlsen would play chess even if nobody watched. Just look at Mikhail Tal, Max Euwe, Garry Kasparov and all those others. None of them did it because of the crowds. They did it because the game itself fascinated them.
> I suspect Magnus Carlsen would play chess even if nobody watched.
Maybe, but it wouldn't be his job then.
Even though cars are faster, there are still human runners.
I agree with the conclusion but the premise is bunk. AI won’t replace you because it’s unsustainable and grossly overhyped, not because it’s as effective at general tasks or code review to the same extent that computers can play chess.
And to what end? I sometimes get the impression software people want to optimize out everything so in the end, when there's nothing left to do, they can lay down and die.
It wasn't that long ago that if you worked at an engineering firm, you needed a whole floor or floors of skilled draftsmen to produce the engineering drawings of your bridge, or your widget. Now an intern can do that in seconds with modern CAD software.
The same phenomenon is already occurring with AI. The concern is not that it will replace people, it's that it will make people so productive that it will create an oversupply of workers and thus kill a lot of jobs.
While I agree that AI wont replace you, the analogy is BS. Magnus Carlsen is not replaced by a chess engine because he is not paid solely because of his raw skills. As any professional player of any massive sport, he is an entertainer first. He gets paid only because people want to watch him play.
Crazy cope lately , understandable human emotions though
My go to phrase for the equivalent observation is "If they ever make a machine that can make a coffee at the press of a button then all the coffee shops will have to close" /s.