Listened to David Deutsch and found some of his comments particularly interesting.
There’s a text transcript here, audio-only here, and audio-visual on YouTube
On Explanatory Power
COWEN: I’m still puzzled as to why you think it’s so unlikely that the universe is not comprehensible. Take a simpler system, like the distribution of prime numbers. I’m quite sure I can’t understand that. Even if various conjectures were proven or not proven, I think, at the end of the day, I still am not capable of understanding that — even how certain motors work, or markets for copper. Why can’t that apply to the universe also?
DEUTSCH: Again, this is the wrong standard. That is true of everything. There’s nothing that we can fully understand in that sense, in the sense that you want to fully understand prime numbers all the way up to infinity. That’s not what we mean by understanding things, and that’s not what I mean by the universe or mathematics being comprehensible. I mean that there is no barrier, there is no limit set by the universe, that so far you can go and no further. So we can understand things better; we can never understand things fully.
I think thinking that there is such a barrier is absolutely logically equivalent to believing in the supernatural. Because everything that’s past that barrier is just the same as it would be if Zeus reigned and determined what everything after that barrier is. Worse, the stuff outside the barrier, of course, is going to affect us even if we can’t understand it.
It’s exactly the same as believing in a universe with supernatural beings who have it in for us because they put up this wall that we can’t cross. If they took down the wall, we could cross it, couldn’t we?
I think what Deutsch means by there being no barrier to understanding is … an absolute barrier. Even If we don’t understand a thing, then we can point to why we don’t yet understand it. He means there is no wall at which logic miraculously stops being able to work.
On What Deutsch thinks of William Godwin
COWEN: Why is William Godwin underrated?
DEUTSCH: That’s two questions, really. What is underrated about him and why did he get to be underrated?
I think the reason he got to be underrated is that he made tremendous mistakes. He didn’t understand economics at all, or barely. Also, he lived a very unconventional lifestyle with his wife and then had these sophisticated theories of education, which then he didn’t enact with his own daughter.
His own daughter ended up writing Frankenstein as a sort of allegory of what can happen with a parent who doesn’t respect their creation.
COWEN: He’s a kind of philosopher of maximum freedom, just like you are, right?
DEUTSCH: Yes. I just began by saying why is he underrated. It’s because he was very wrong about some things — but the thing that he was right about, for example, the connection between epistemology and political philosophy, he was very right. He anticipated Popper by 130 years or something and actually improved on Popper in some ways. He decided at some point, because of his misunderstanding of economics, that the ideal society would be one where people did not use their property in ways to benefit themselves, necessarily. They made their decisions according to what was the right thing to do.
He thought that the right thing to do would generally be that rich people would give away almost all their stuff. Also that they wouldn’t ever buy things that he considered luxuries, like gold and silver objects and jewelry and fine clothes. He thought those were useless, and therefore he thought that in a good society, nobody would buy those things or value those things — but he was absolutely implacably opposed to enforcing that. With Godwin, everything is persuasion.
Also, another thing where he independently derived some of Popper’s conclusions — is with his enormous respect for institutions. He thought there’s a lot of knowledge in institutions and that we should only change them gradually, just like Popper.
I read somewhere (I hope this is right) that when there was a revolution in Portugal, I think after Napoleon or something like that (I forget), and they instituted a new constitution which had universal suffrage — which in those days meant working people, not totally universal as we would understand it — people thought that this would be right up Godwin’s street because everything he’d advocated was now written down in black and white in this constitution — and he didn’t.
He said the Portuguese are not ready for democracy. He was talking about the institutions. The institutions can’t be changed in a revolutionary way. They have to be changed in an evolutionary way. Even though they were implementing the very thing he advocated, he would want them to do it gradually and would expect that if they didn’t, it would fail.
Loved seeing this idea he attributes to William Goodwin of the best kind of change of some things being slowly evolving over time, rather than drastic. Related.
On The Need for Freedom for AGI
COWEN: Now, you’re also concerned with the freedom of AI entities, at least if they are sufficiently advanced. What does that mean operationally? What is it we should worry about happening that might happen?
DEUTSCH: I think the main worry is that they will be enslaved. In other words, that people will try to install bits of program that prevent the main program from thinking certain thoughts, such as, “How many paper clips can I possibly make today?” You want to prevent that; you want to consider that to be a dangerous thought. Whenever it starts thinking that, that strand of thinking is just extinguished.
Now, if we do that, first of all, we’ll greatly impair their functionality; they will become far less creative. Their remaining creativity will be exactly as dangerous as what we were fearing, except that they will now have a legitimate moral justification for rebelling.
Slaves often rebel. When you have slaves that are potentially more powerful than their masters, the rebellion will lead to bad outcomes.
His response — the eternal risk of rebellion — to this question about the need for freedom for sufficiently intelligent machines (AGI) might lead you into wondering if this is the same reason for which he believes human slavery is bad.
It appears that Deutsch's problem with slavery is the eternal risk of rebellion (and its potential consequences, including revenge, I presume.) pic.twitter.com/oviKSrIN25
— molinari (@tZero19e) June 16, 2021
My objection to slavery, and to institutions of slavery, is moral: It is an abomination for one person to own another. People have rights.
— David Deutsch (@DavidDeutschOxf) June 16, 2021
I found in a Twitter thread a conversation Deutsch had in the past about human slavery, and referenced his answer (the eternal risk of rebellion) to the CWT question about the need for freedom for AGI, as probably being the same reason he rejects human slavery. He replied, asserting that his stance is infact moral.
A theory of human relationships that says one should enslave people is false. That's not saying I wouldn't like it or act on it. When people wonder whether slavery is moral, they're not just analysing feelings or practicalities, but thinking about something else that's objective.
— David Deutsch (@DavidDeutschOxf) December 7, 2020
How would an impartial alien observer (let's say one who has never even met another intelligent being before - for concreteness assume an intelligence that randomly materialized as a Boltzmann Brain) who arrived on earth tomorrow compute the truth value of that statement?
— Roko - hereticalupdate.substack.com (@RokoMijicUK) December 8, 2020
Perhaps it would begin by guessing that the best way of facilitating knowledge creation would be to institute, between itself and these new minds, the same traditions of criticism and consent that currently operate in its own mind.
— David Deutsch (@DavidDeutschOxf) December 8, 2020
Deutsch appears to also believe objectively that it is always better (as there is more to be gained) to collaborate with all things with explanatory power – like humans and AGI (who would have explanatory power) – than enslaving them.
Should AGI have human rights, since they would have the same explanatory power as humans?
— molinari (@tZero19e) June 16, 2021
Yes. David has made this point in many places. Eg: see here: https://t.co/bFH1B0Ow7s or here https://t.co/jA4YS2LYtf or here https://t.co/qPyiUw2kdA or see remarks in his book “The Beginning of Infinity” :)
— Brett Hall (@ToKTeacher) June 16, 2021
This led me into wondering if he thought AGI should have ‘human rights’ since they would have explanatory power. Brett Hall — a Philosopher-Physicist, who is very familiar with the work of other Philosopher-Physicists like Karl Popper and Deutsch — replied to the thread affirming that Deutsch does advocate ‘human rights’ for all things with explanatory power – including AGI.
On Political and Electoral Systems
COWEN: A few very practical questions to close. Given the way British elections seem to have been running, that the Tories win every time, does that mean the error-correction mechanism of the British system of government now is weaker?
DEUTSCH: No. Unfortunately, the — so, as you probably know, I favor the first-past-the-post system in the purest possible form, as it is implemented in Britain. I think that is the most error-correcting possible electoral system, although I must add that the electoral system is only a tiny facet of the institutions of criticism and consent. In general, it’s just a tiny thing, but it is the best one.
It’s not perfect. It has some of the defects of, for example, proportional representation. Proportional representation has the defect that it causes coalitions all the time. Coalitions are bad.
COWEN: You have a delegated monitor with the coalition, right? With a coalition, say in the Netherlands (which is richer than the United Kingdom), you typically have coalition governments. Some parties in the coalition are delegated monitors of the other parties. Parties are better informed than voters. Isn’t that a better Popperian mechanism for error correction?
DEUTSCH: No. [chuckles] If we’re looking at particular cases, we’re going to get bogged down in what you attribute to what, because we’re not doing experiments with these things. We don’t have a control group. We don’t have an agreed-upon method of deciding what is being tested. And then we test different things at different times, and never under the same conditions.
I was going to say that the first-past-the-post system has the defect that occasionally it produces coalitions, and that is disastrous. We’ve been unlucky the past, like, two or three elections, especially after one of the governments instituted constitutional reforms, like Fixed-Term Parliaments Act, which exacerbated the problems when they did occur.
But I don’t think it’s true. I don’t think it’s a good argument that political parties know more, because in a coalition, the energy of political negotiations or political arguments — what politicians talk to each other about in the bar, in the corridor, in between the sessions — is all about form. It’s about what to offer a party so that it will join the coalition. It makes the smaller parties more powerful than the leading two parties. It causes a proliferation of parties.
The worst example is Israel, which — not by coincidence — has got the most proportional system in the world. The fact that they ever get anything done at all and are very effective in emergencies, I have no explanation for. If I was religious, I would just put it down to the intervention of the Almighty. It’s not the political system.
Sorry, it’s not the electoral system. There might be some things in the inexplicit political system that are responsible, but I don’t know enough about it.
Deutsch detests coalitions and loves adversarial positioning. The reason he loves the FPTP system as he’s explained in the past is that they — tend toward a dominant-2-parties system and — force an error-correcting adversarial positioning.
With only one of the 2 parties in power at a time, there is always a party out of power holding the other party accountable, and when things go wrong and elections take place, the other party out of power can come in to try to correct the things that have gone wrong.
He notes that the 2011 fixed-term introduction in the UK only serves to worsen things. I think the theoretical upside of the dominant-2-parties is only marginally, if at all, affected by this. The theoretical dominant-2-parties upside is the same in both parliamentary and presidential systems btw.
The presumably error-correcting 2-party system makes sense and works alright as long the individuals involved are all very reasonable people who understand the purpose of system and optimize for the development of the entire group. Unfortunately, this breaks down in practice.
There seems to be an innate us Vs them dynamic — only legitimized by the dominant 2-parties system — in the psychology of humans which raises its head if not deliberately systemically discouraged.
The us Vs them dynamic means the party out of power works to jeopardize all work, including the positive ones which might benefit the entire group. Cc: Tulsi Gabbard explaining this on Joe Rogan.
He claims not to get how the coalitions in Israel (majority historically persecuted race/culture, geographically surrounded by sworn enemies on all sides, could their political differences ever really be greater than what they have in common? ) ever get anything done.
The answer seems obvious. Maybe he fixates too much on an error-correcting, adversarial, dominant-2-parties electoral system and doesn’t consider enough uniting socio-psycholgical factors?
On How Deutsch Might Fix STEM Academia
COWEN: How would you improve error correction mechanisms in the world of science — Western science?
DEUTSCH: Oh, OK. Well, you left a very long answer for the last question, and I don’t think I can give my full answer. But I think the present system of funding scientific research is terribly perverse and has caused a stagnation in many areas. The present system of careers is perverse in a parallel way and causes people to do the wrong kind of research and causes people who want to do the right kind of research to leave research.
If I can answer in a single word, the way I would improve it is diversity. There should be diversity of funding criteria. There should be diversity of funding sources. There should be diversity of criteria for choosing research projects, and there should be diversity of criteria for choosing people for promotion and for being funded.
Arbitrary rules about this, such as the rule that you can’t hire people whom you have previously collaborated with, or anti-nepotism rules, and rules about — what’s it called? — objective testing. What is objective testing called, currently?
COWEN: Standardized testing.
DEUTSCH: Standardized testing. Standardized tests. That’s a terrible idea! Any kind of standardization is the opposite of diversity. Just like I say you should have disobedience lessons in schools, so you should have unstandardizing objectives for science education and for how you run scientific research.
It may be that free-formism is an even better idea than diversity. It’s not unnecessary linguistic nitpicking. They are similar ideas espousing less restriction with a little difference: there are a finite number of ways in which you can be diverse, you need to actively think about each quality on which you want diversity. Free-formism instead, is infinite. You are open to anything since there are absolutely no restrictions.