We've been over this. Gödel's mu-recursive functions were a poor model of computation because it's completely unclear how to physically implement the arbitrary-function minimization operator. So people didn't see how to build a machine that calculates this way. Similarly, there's no clear way how to mechanize lambda calculus. Turing Machines, on the other hand, were instantly obviously mechanizable. It was clear that…
Jürgen Schmidhuber has a track record of complaining about unfairness in the conventionally accepted view of the scientific record, especially in the English-speaking world. For instance, he has claimed for years that his own early research on AI has been unfairly overlooked or ignored by a "conspiracy" (his word, not mine).[a] At one point, the NY Times called him the "Rodney Dangerfield of AI research" because he d…
Turing Oversold?
211–220 of 311 posts
Re: Turing Oversold?
#212Earlier quoted context omitted.
Why do we need idols, though? If there was no narrative, no idols, no celebrities, would people be less motivated to do science? Why do we need to lie to ourselves so? > If you're already deeply interested in computer science, of course the detailed narrative recognizing dozens of brilliant early computer scientists is richer. Of course! Personally I'm mostly uninterested in who did what, but maybe that's just me. It…
> Why do we need idols, though? Because we're flesh and blood, i.e. utterly irrational. > If there was no narrative, no idols, no celebrities, would people be less motivated to do science? Why do we need to lie to ourselves so? Yes, definitely, a huge amount of what motivates scientists is desire for fame, being considered a genius, Nobel prizes, scientific immortality, and so on. It is entirely unrealistic to imagin…
> This may be an accurate description of your personality, in which you're one in a million, or it may be that you're ignorant about the things that actually drive you. The vast majority of people are driven by some kind of desire for fame, recognition, status, upvotes, and so on.
Or it might be that people who are driven by fame and recognition are more likely to become famous than those who aren't, which skews our idea of what motivates people. Given how emphatic society is about fame and money as markers of success, I feel people tend to be mistaken in the other direction: many people think they are, or should be driven by fame or money even when it simply contradicts their personality.
Even if it was indeed the case that most people are motivated by fame, I think those who aren't are more like 1 in 3 or 1 in 4 than 1 in a million. It might be 1 in a million in actually famous people, but not in the population at large.
> Even in proper context, they were exceptional geniuses, just among other, lesser-known, exceptional geniuses.
If I am correct that millions of people had the capability, that would place "exceptional genius" at 1 in 1000, or 1 in 10000. I think that's a reasonable ballpark.
> If we take your view seriously, we remove all human agency and uniqueness, we remove the idea of an "achievement" and we can only give credit to luck, the historical process, and various contingent circumstances.
Whether we acknowledge exceptional geniuses or not, it remains the case that 99.99% of people are not exceptional geniuses. Are you saying these people don't have agency, or that they aren't unique? I think we all have agency, we're all unique, and we all have achievements. Some achievements are more impactful than others, some achievements are more impressive than others, but these are not necessarily the same, and neither is necessarily remembered, because what matters most is not the person or the achievement, but how the story fits in the narrative. In any case, you don't need to care about that narrative to care about or acknowledge agency, uniqueness or achievement.
Re: Turing Oversold?
#213Earlier quoted context omitted.
I'm still a little confused. It seems like Turing came up with something that works, and clearly fulfills precisely what Godel and Church and Turing were all looking for; but it also seems like it's a mathematically inelegant solution. Is it possible that in the future we'll find a way to show that Gödel's mu-recursive functions or Church's lambda calculus also precisely describe 'what a machine can do'? If so, it se…
> Is it possible Yes, the set of functions computable with mu recursion, or with lambda calculus, is the same as the set of functions computable with a Turing machine. (Turing in his original paper showed that the set is the same for his system and Church's, and the proofs for mu recursion and many other systems are well-known.) > I'll totally agree that from a teacher's perspective, the Turing machine is the better…
Thanks, I think I understand now. I thought there was a distinction between the mathematical idea of what computation is, and the engineering we've invented to implement that computation, and so I didn't really get the significance/literalness of what you were saying about mechanization.
It does seem weird to me though that we're letting our engineering limitations determine how we think about these things mathematically.
Re: Turing Oversold?
#214Jürgen Schmidhuber forging on in his never-ending quest to slay historical misattributions and generally right all wrongs.
We can count ourselves lucky this time: at least, the topic isn't himself.
On a positive note, he can't be accused of being inconsistent.
Re: Turing Oversold?
#215This has changed over time. In early computing, the people mentioned were Eckert and Mauchley, who designed the ENIAC, and von Neumann, who figured out how a practical CPU should be organized. Also Shannon, who figured out that relays basically did Boolean algebra and could be analyzed formally.
Bletchley Park work was still classified in those days, so not much was known about that. Much to the annoyance of Tony Flowers, who built the Colossus code-breaking machine but couldn't mention it in post-war job-hunting. (Incidentally, Colossus was not a general-purpose computer. It was a key-tester, like a Bitcoin miner ASIC.)
As I've pointed out before, the big problem was memory. IBM had electronic addition and multiplication, with tubes, in test before WWII. But the memory situation was really bad. Two tubes per bit. Or electromagnetic relays. Or electromechanical counter wheels, IBM's mainstay in the tabulator era. To store N bits, you had to build at least N somethings. Babbage's Analytical Engine called for a memory of a ring of rows of counter wheels, a mechanism the size of a locomotive. Access time would have been seconds.
Much of early computing was about kludges to get some kind of memory. The "Manchester Baby" had a CRT-based memory, the Williams tube. That was the first computer to actually run a program stored in memory. Frederic C. Williams, Tom Kilburn, and Geoff Tootill, 1948.
After that, everybody got in on the act. Mercury tanks and long metal rods were built. Really slow, and the whole memory has to go past for each read, so twice the memory means twice the access time. Then there were magnetic drum machines. Magnetic tape. Finally, magnetic cores. At last, random access, but a million dollars a megabyte as late as 1970. Memory was a choice between really expensive and really slow well into the 1980s.
Turing was involved with one of the early machines, Pilot ACE/ACE. But he quit before it was finished.
Re: Turing Oversold?
#216Earlier quoted context omitted.
I'm still a little confused. It seems like Turing came up with something that works, and clearly fulfills precisely what Godel and Church and Turing were all looking for; but it also seems like it's a mathematically inelegant solution. Is it possible that in the future we'll find a way to show that Gödel's mu-recursive functions or Church's lambda calculus also precisely describe 'what a machine can do'? If so, it se…
I'm convinced that David Harland's Rekursiv[1:] machine _is_ the manner by which lambda et al. might be implemented at the machine level. Unfortunately, Rekursiv seems to have died an ignominious death, with the last Rekursiv chip having fallen off the side of a steamboat (apocrypha; I remember having read this, but I'm unable to find the original citation.) The Rekursiv advantage is its ability to do recursion on th…
I guess I'm confused as to why this is more than simply an interesting formalism. Complexity theory and models of computation are largely based around the Turing Machine model. Lambda calculus is an effective lens to design programming languages and prove equivalence between programs. We know by way of the Church-Turing Thesis that these two models are equivalent. The Turing Machine model is both better studied from a computation perspective and has much more practical realization; what's the point in creating something like this? It feels a bit like silly lambda calculus fetishism, but again maybe the value here is the actual computation formalism itself.
Re: Turing Oversold?
#217Earlier quoted context omitted.
Likewise, people dispute that Ada Lovelace was the first programmer, because Babbage and Menabrea had previously created a few simple example programs. But that downplays her accomplishments too much. She didn't write the "first program" but she was the first to understand what computers would be capable of doing (for example, that by assigning numbers to letters and symbols, computers could do more than simply perfo…
I really enjoyed Stephen Wolfram's mini-bio of her. https://writings.stephenwolfram.com/2015/12/untangling-the-t... I very much recoginized from that that she had the attitude and experience of a "programmer," so I would say she was the first programmer, in the modern sense.
Ada Lovelace's father was Lord Byron?! TIL
Re: Turing Oversold?
#218Earlier quoted context omitted.
>>They're not the markers of exceptional genius we want to think they are, not for Turing or Einstein, but not for anyone at all, really. The point isn't to prove that they're special. The point is that something special happened and these people are designated symbols for that... and they're kind of selected for being good at this. We're not doing this for them, they're dead. The celebrity of Einstein is a deificati…
But what purpose do these idols or symbols serve, exactly? I'm speaking as someone who doesn't care who came up with relativity and doesn't care whether there is a founding person of computer science or not let alone who that would be, and would like to know what others see. Is it an inspiration thing? A motivation thing?
Not necessarily a motivational thing, but events such as these become widespread and allows for easier dissemination of information.
It's easy to see then that such events allow for the eventual "recruitment" of other scientists, and in showing society that "science is working" and "solves important problems".
Both of which serve to enrich the scientific world with new researchers and funding to keep the engine running.
Re: Turing Oversold?
#219We've been over this. Gödel's mu-recursive functions were a poor model of computation because it's completely unclear how to physically implement the arbitrary-function minimization operator. So people didn't see how to build a machine that calculates this way. Similarly, there's no clear way how to mechanize lambda calculus. Turing Machines, on the other hand, were instantly obviously mechanizable. It was clear that…
Jürgen Schmidhuber has a track record of complaining about unfairness in the conventionally accepted view of the scientific record, especially in the English-speaking world. For instance, he has claimed for years that his own early research on AI has been unfairly overlooked or ignored by a "conspiracy" (his word, not mine).[a] At one point, the NY Times called him the "Rodney Dangerfield of AI research" because he d…
>Jürgen Schmidhuber has a track record of complaining about unfairness in the conventionally accepted view of the scientific record, especially in the English-speaking world.
So, he has
>Ultimately, as always, credit will be assigned by future generations of scientists and historians long after everyone who is alive today has disappeared from the face of the earth.
Well, credit for the field of CS is already applied now, and also yesterday and 20 years ago. It's not like this future credit would be the first such, or that it would be more accurat (or that it would not feed of the current popular views).
Re: Turing Oversold?
#220We've been over this. Gödel's mu-recursive functions were a poor model of computation because it's completely unclear how to physically implement the arbitrary-function minimization operator. So people didn't see how to build a machine that calculates this way. Similarly, there's no clear way how to mechanize lambda calculus. Turing Machines, on the other hand, were instantly obviously mechanizable. It was clear that…
Likewise, people dispute that Ada Lovelace was the first programmer, because Babbage and Menabrea had previously created a few simple example programs. But that downplays her accomplishments too much. She didn't write the "first program" but she was the first to understand what computers would be capable of doing (for example, that by assigning numbers to letters and symbols, computers could do more than simply perfo…
>She didn't write the "first program" but she was the first to understand what computers would be capable of doing
Or merely the first to express it? I'm pretty sure Babbage himself, as the inventor, understood well what computers would be capable of doing.