I'm not sure it's right to characterise Deep Blue or Stockfish as repositories of human chess theory. Fundamentally they were all based on a relatively simplistic function for calculating the value of a board position combined with the ability to evaluate more board positions further into the future than any human possibly could (plus a database of opening moves). That approach seems thoroughly non-human, and represents a victory of tactical accuracy over chess theory or strategy.
However I agree that the games between AlphaGo and Stockfish are really interesting. It strikes me that the AlphaGo version of chess looks a lot more human; it seems to place value on strategic ideas (activity, tempo, freedom of movement) that any human player would recognise.
I think you're right, I meant to say that chess engines usually have book openings built into them which derive off of human chess theory but you're absolutely right in that they don't play in a human form.
It's kind of crazy how AlphaZero has managed the success it has. Stockfish calculates roughly 60 million moves per second and AlphaZero calculates at only 60 thousand per second. Three orders of magnitude less yet its brilliance is mesmerizing, tearing Stockfish apart in certain matches.
I think, in practice, "justice" has always been a mediated form of vengeance. The state maintains its monopoly on violence by extracting enough vengeance from criminals that individuals (i.e. victims or their families) are willing to subcontract revenge to the state. In other words, the "justice" system has to provide enough vengeance to retain the consent of the victims, or those offended on their behalf. The benefits to society are large; total violence is much lower once it's able to break the cycles of private retaliation and avoid endless bloodfeuds. But it is an uncomfortably pragmatic equation.
I don't think there's any other consistent way to look at the issue though. Treating murder as the ultimate crime, as almost all justice systems do, makes sense in terms of the cost to the victim and the desire of societies as a whole to see murderers punished. But most murders are unplanned, and most murderers are no more likely to murder again than any other member of the population. Given the above, long sentences for (unpremeditated) murder just do not make sense; they probably have little value either as deterrence or prevention, and they certainly don't rehabilitate. But would a victim's associates stand idly by if their murderer was free after five years?
In as much as this analogy is useful: when you lie to the lender about your income, spending and ability to repay the loan, you have committed fraud.
But the EU is supposed to be an a collective of allies and partners, not a network of purely financial transactions. In that context surely the shame belongs to the partner that betrays the trust and confidence of its friends?
Not that I'd argue that Greece "deserved" everything that happened. But this line of argument seems to justify a complete refusal of responsibility on Greece's part.
Historically it seems that the highest paid developers in London have been contractors, who can charge £600 ($780) / day for fairly non-niche skills in web-development.
IR35 may change the incentives for employers enough that that kind of money is more likely to be paid in salary form. We're now seeing senior developer salaries being advertised as high as £90k - £100k ($130k). Again, for in demand but non-niche skills (Ruby, React, full-stack JS etc).
I'm sure London won't be seeing SF or NYC level salaries, but you can definitely live comfortably in London on a software engineer's salary.
"As soon as men decide that all means are permitted to fight an evil, then their good becomes indistinguishable from the evil that they set out to destroy."
The UK right to silence is pretty well summarised in the modern wording of the police caution: "You do not have to say anything. But it may harm your defence if you do not mention when questioned something which you later rely on in court. Anything you do say may be given in evidence."
Staying silent is not obstruction of justice or contempt of court. No one is ever required to respond to questioning. But it might lose you credibility in court of you produce an alibi or other defence at trial that you didn't mention during questioning.
The only exception is passwords and encryption keys, which have to be supplied if a court order is obtained and aren't considered testimony.
The right to silence was changed to this in the 1994 Criminal Justice and Public Order Act.
The reason for the change (AIUI) was due to the arrest of IRA dissidents who remained silent and then produced alibis in court that would have been easy to dispel at the time but were not so easily dismissed in court, once they had gathered support from their communities.
Incidentally, the 1994 act also banned music containing "repetitive beats" and legalised anal sex between heterosexual couples, albeit with a number of caveats and conditions.
FFTP is biased in favour of the party with the largest share of the vote, or perhaps another way of putting is that the relationship between vote share and number of MPs is not linear.
And beyond this, the British electoral system has consistently favoured Labour for decades (as Boundary Commission changes fall behind demographic change and population movement)
Thanks for the link. I've read the Independent article, clicked through and read the author's Facebook "footnote", and also read the paper referenced in that.
I can't see anything that disputes the claim "The Tories are the clearest benefactors" of the UK's electoral system.
The author makes a repeated assertion that boundaries are currently biased "15 seats" against the Tories, but I cannot tell where this number is coming from.
The fact remains that the Tories have won the last few elections by receiving a disproportionate number of MPs relative to their vote share. Moreover, polling suggests that they will continue to have the highest vote share.
it is utterly bizarre that you're treating the fact that the party that got the most votes won the most seats as evidence that the electoral system is biased towards them.
Have you considered adding elections when the Conservatives didn't win the most votes to your data set?
That's a disappointing misrepresentation of my premise, which is that the system is biased so that high vote shares confer disproportionately many seats.
I know this conversation has gone slightly stale but I've reread that John Rentoul piece a couple of times, and it's very wishy washy but as far as I can make out his argument is this:
1. The Conservative party has proposed some new boundaries.
2. The new boundaries would have given the Conservatives 15 extra seats, according to the 2017 vote share.
3. The current boundaries are very out of date.
4. The new boundaries follow stricter rules about constituency size variation.
5. The new boundaries lead to a reduced number of MPs.
6. 3-5 imply that the new boundaries are fairer than current boundaries.
7. 2 and 6 imply that the current boundaries are biased 15 seats towards Labour.
The hole is in statement 6; points 3 and 5 are irrelevant to the question of fairness and point 4 is weak as there are plenty of ways to gerrymander the boundaries whilst equalising constituency size. What's "utterly bizarre" is that Rentoul expects many people to swallow such a flawed argument.
'Brogrammer' started as a witticism and has become a social phenomenon that people discuss as though it was based on rigorous observation and not just a joke.
In college I knew a number of "brogrammers" (frat boys, who did fratty things, and banged their head against coding as best they could). None of them are programming today.
However I agree that the games between AlphaGo and Stockfish are really interesting. It strikes me that the AlphaGo version of chess looks a lot more human; it seems to place value on strategic ideas (activity, tempo, freedom of movement) that any human player would recognise.