Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

How should we talk to AIs?

blog.stephenwolfram.com

Post reply on HN
Print view View original

11–20 of 46 posts

Re: How should we talk to AIs?

#11

Post by jjaredsimpson »

gohrt wrote: post #9
bashedly wrote: post #2

I'm hoping someone closer to the WolframAlpha team and work can shed some light on this - but is the extensive use of the first person warranted here? There's an awful lot of "I" and "I've". Is the work being done really so directly attributable to Stephen Wolfram or is there a army of hard working individuals behind the scenes not being referenced here? I'm not suggesting they list everyone by name or anything, but…

Wolfram employeees sign contracts giving Dr Wolfram credit for their work, in exchange for their pay.

https://www.youtube.com/watch?v=77Y6CIyyBcI

That's what the money is for

Re: How should we talk to AIs?

#12

Post by bashedly »

gohrt wrote: post #9
bashedly wrote: post #2

I'm hoping someone closer to the WolframAlpha team and work can shed some light on this - but is the extensive use of the first person warranted here? There's an awful lot of "I" and "I've". Is the work being done really so directly attributable to Stephen Wolfram or is there a army of hard working individuals behind the scenes not being referenced here? I'm not suggesting they list everyone by name or anything, but…

Wolfram employeees sign contracts giving Dr Wolfram credit for their work, in exchange for their pay.

I know this is overly idealistic, but it seems to me that paying for "credit" (especially in terms of research) seems "wrong". I understand paying for ownership of the results of someone's work, or paying to state that the broader organization takes credit for the body of work produced by the employees. But for one individual to pay to personally receive credit for the output of another seems blatantly dishonest.

Re: How should we talk to AIs?

#13

Post by zamalek »

He goes on ad nauseum about how fitting the language is. He's right.

Back in university I'd frequently help seniors out with projects. The one project involved battleship AI in mathlab. In maybe 3 hours I got it as good as it was probably going to get pitted against a truly random opponent. However, PNRGs aren't truly random. It was so quick and so effortless to grab stats on the biases the PNRG had (in order to optimize the guess selection process). It was the first time I ever used the language and I can honestly admit it had very little to do with my capability and so much more to do with how damn fluent and intuitive that language is.

It's inconsistent, messy and evolutionary but, wow, what an absolute masterpiece.

Re: How should we talk to AIs?

#14

Post by joesb »

hacker_9 wrote: post #10

Interesting that the Wolfram Language is very much a LISP dialect, with the syntax swizzled a bit here and there to be a bit more human readable. Also his Riffle example is thinking about the problem the wrong way round. It looks like he tried to turn code into english, when it should be the other way around. Here is an actual english version: 1. Start with a purple square 2. Clone the square, rotate it by 0.1 radian…

> All these ambiguities we can solve

We can solve it by specify what each word means. That's no different than programming.

Also, your step 4 has bug because "repeating the clone process" does not include "rotating and scale it to fit" part. It'll just keep cloning the square all days and never get the square to be the size of a dot.

Also, If you screen is not a square and you did not start in the center of the screen. Then the square can be as small as a dot, but it won't be "in the middle of the screen".

Re: How should we talk to AIs?

#15

Post by gooseus »

TeMPOraL wrote: post #4
bashedly wrote: post #2

I'm hoping someone closer to the WolframAlpha team and work can shed some light on this - but is the extensive use of the first person warranted here? There's an awful lot of "I" and "I've". Is the work being done really so directly attributable to Stephen Wolfram or is there a army of hard working individuals behind the scenes not being referenced here? I'm not suggesting they list everyone by name or anything, but…

It's Standard Wolfram Style. You can see the same pattern in his "New Kind of Science" book - an interesting piece of writing, but the amount of "I", "I've", "my", etc. is plain distracting.

I understand so little of the topics of that book or this review, yet still very much enjoyed reading the takedown:

http://bactra.org/reviews/wolfram/

"So much for substance. Let me turn to the style, which is that of monster raving egomania, beginning with the acknowledgments..."

Re: How should we talk to AIs?

#17

Post by hacker_9 »

joesb wrote: post #14
hacker_9 wrote: post #10

Interesting that the Wolfram Language is very much a LISP dialect, with the syntax swizzled a bit here and there to be a bit more human readable. Also his Riffle example is thinking about the problem the wrong way round. It looks like he tried to turn code into english, when it should be the other way around. Here is an actual english version: 1. Start with a purple square 2. Clone the square, rotate it by 0.1 radian…

> All these ambiguities we can solve We can solve it by specify what each word means. That's no different than programming. Also, your step 4 has bug because "repeating the clone process" does not include "rotating and scale it to fit" part. It'll just keep cloning the square all days and never get the square to be the size of a dot. Also, If you screen is not a square and you did not start in the center of the scree…

Joesb if anything you are proving my point! As another human you saw large ambiguities in my instructions but could still figure out 1. the problems (you knew no one would clone the square infinitely as it wouldn't 'make sense') and 2. the required solutions (i probably meant scale+rotate as well to be included in the cloning process, otherwise no interesting visual change would happen).

Tell me what programming language / compiler can do that by just specifying word meanings? Context, sentence structure, your understanding, etc all play into it as well.

Re: How should we talk to AIs?

#18

Post by hacker_9 »

zamalek wrote: post #13

He goes on ad nauseum about how fitting the language is. He's right. Back in university I'd frequently help seniors out with projects. The one project involved battleship AI in mathlab. In maybe 3 hours I got it as good as it was probably going to get pitted against a truly random opponent. However, PNRGs aren't truly random. It was so quick and so effortless to grab stats on the biases the PNRG had (in order to opti…

Any specifics as to what you mean? How it was more elegant that a more mainstream language?

Re: How should we talk to AIs?

#19

Post by prahladyeri »

bashedly wrote: post #12
gohrt wrote: post #9

Earlier quoted context omitted.

Wolfram employeees sign contracts giving Dr Wolfram credit for their work, in exchange for their pay.

I know this is overly idealistic, but it seems to me that paying for "credit" (especially in terms of research) seems "wrong". I understand paying for ownership of the results of someone's work, or paying to state that the broader organization takes credit for the body of work produced by the employees. But for one individual to pay to personally receive credit for the output of another seems blatantly dishonest.

Copyright law deals with the situation you've mentioned in different ways in different countries. For instance, it is just illegal in Europe to transfer the credit of one's labor (or copyright). In the Americas, however, companies blatantly mention in the `TOCs` that they own rights over employee's all work.

Of course, the extent to which the law is actually applied in Europe depends on the conditions of employment and whether the employees have the time and resources to wage a legal war against their employers.

Re: How should we talk to AIs?

#20

Post by winter45 »

Concerns about consciousness, intelligence,and Stephen Wolfram's use of the personal pronoun aside, I think it reasonable to presume that communication with early stage AI systems will desirably include the development of a vocabulary and syntax structured for 'precision'.

It does seem reasonable that if AIs evolve they will, in their evolution, develop constructs of 'reality' that Wolfram calls "Post Linguistics Emergent Concepts"(1) and that, if they do, until human languages develop deeply "precise" 'words' for each of these PLECs, Spock will have to translate for Kirk; Spock, talking to an AI, will be (mostly?) unintelligible to Kirk.

This raises the question can one reasonably believe that humans are capable of more than 'bare bones' precision in words? Can one be both Kirk and Spock?

On that note it is encouraging that Wolfram appears to think our systems may be (become?) a driving force in our evolution; we may evolve to a point the use of a precise or imprecise language is a moment-to-moment choice.

It may be that these precise and imprecise languages will develop into a single language.

Before that happens, if you were such a dual-linguist, which language would you prefer? Under what conditions would you switch? What would communication be like in a world where the commonly spoken language was a meld of precision and imprecision? Which of the two would have the most influence on the other? Past a certain stage are there generally distinctions in character?

(1) see: https://www.youtube.com/watch?v=TMviBl46dXg

Post reply on HN