Slashdot Mirror
Interviews: Ask Stephen Wolfram a Question
Stephen Wolfram's accomplishments and contributions to science and computing are numerous. He earned a PhD in particle physics from Caltech at 20, and has been cited by over 30,000 research publications. Wolfram is the the author of A New Kind of Science, creator of Mathematica, the creator of Wolfram Alpha, and the founder and CEO of Wolfram Research. He developed Wolfram Language, a general multi-paradigm programming language, in 2014. Stephen has graciously agreed to answer any questions you may have for him. As usual, ask as many as you'd like, but please, one per post.
How much RAM is a wolf supposed to have?
Signed,
Dr. Algernop Krieger.
Get free satoshi (Bitcoin) and Dogecoins
If you want to advance mathematics you must show your working.
How "Angel" is doing.
If you could reason with religious people, there would be no religious people
How can Wolfram Language improve Artificial Intelligence?
What are some of the latest developments you have been working on? What can we expect later this year?
P.S. We met at HackBCA where I got to try the Wolfram Language.
I am concerned that we have not, nor will reach the the high level programming language and data presentation formats necessary to accomodate the advances in HMI (e.g direct cortex interfaces et al) that will be available soon. That we are still thinking keyboards and screens. What are your thoughts on this and the ramifications on code language, OS and presentation?
As you probably know, the most recent paper on Orch-OR as a proposed mechanism for consciousness may have a role for cellular automata in the underlying mechanism.
As you've advocated and made a compelling case for these systems, what are your thoughts on this?
..don't panic
To get a PhD at 20 I imagine you've spent a lot of your childhood reading and doing maths and physics. What is it about physics that draws you? Why does it keep you interested?
What is the most important thing everybody should understand about cellular automata?
Will you please do something with http://www.cyc.com/ and bring the ability to just talk to our computers in natural language? Not like Siri, but like the computer in ST:TNG.
I am not sure if we will get some personal answers out of this guy, I guess he will just forware every single question into one of these websites that claim to know everything. Serioudly, though, what do you think of Elon Musk's fear of A.I., and when do you think that Wolfram Alpha will become self aware?
karma police: arrest this man, he talks in maths; he buzzes like a fridge, he's like a detuned radio. [radiohead]
The penultimate paper of "Bit-string Physics: A Finite and Discrete Approach to Natural Philosophy" discusses an attempted revival of "Relation Arithmetic" with which Russell and Whitehead had planned to cap off their Principia Mathematica in its final volume.
Of Relation Arithmetic, Russel said:
-- " My Philosophical Development" by Bertrand Russell
An example of going astray in attempting to understand the empirical world is when people attempt to combine incommensurable quantities in their calculations, not understanding the structure of the relations between the quantities.
Ordinarily, programming languages treat units, as I/O formats for dimensions, as an afterthought -- independent of type checking. However, what if we saw numbers themselves as embodying relational structure, as intended by Russell, thereby unifying the notion of "type checking" with the notion of "number"? Might then the power of dimensional analysis be brought to bear, in a mathematically rigorous way, on the relatively ad hoc notions of "type", hence problematic areas such as the object relational impedance mismatch?
Seastead this.
You have dedicated a large slice of your time to investigating CA properties and trajectories. So has my friend John Horton Conway who dedicated a slice of his life to Life. However, in your case you seem to have held the belief that CAs in some fundamental way underpin our physical reality. Do you still hold this belief, and if so, could you expand a bit on the current state of your opinions on this matter?
Edmund Ronald.
It seems like we have many smart people warning us about the dangers of AI. What's your opinion on the prospects?
the preceding comment is my own and in no way reflects the opinion of the Joint Chiefs of Staff
Since you tend to name things after yourself, do you regret not naming Mathematica differently?
The first question is a joke, nothing one would ask in polite conversation. My real question to him is this:
I assume it was a pivotal moment in your life when Veltman showed you Schoonship, which was essential to the work later earning him a Nobel with t'Hooft. It was probably the first computer algebra system able to transform the large expressions that you had to deal with in your preceding work on particle physics. Can you describe how and if that changed your perspective on what you would do in life?
"missing line number prefix at line 2"
What is the speed of light in inches per fortnight?
systemd is Roko's Basilisk.
I discovered a very simple proof of Fermat's Last Theorem but every PHD genius I have shown it to have said that they cannot verify that it is absolutely correct. Yet they don't deny that the resulting mathematics shows an irresolvable paradox of basic arithmetic. It seems everyone wants a 'trophy proof' in excess of a billion pages (my proof is two pages) to wave around to the fawning mathematical community. I've lost all regard for the 'academic industry' for it seems to be populated with 'mathematical hieroglyphic' snobs. Any thoughts?
Mathematica use to run great offline on Win 3.1 on 486 platforms, which have a small fraction of the capability and resources of a modern smart phone.
True that the current version of mathematica does so much more, but even an older strip download offline version would be so much more than the countless Android & iOS CAS apps in the App/Play stores.
I purchased Wolfram Alpha which is decent for compution, but often when I need it most, especially for big number modulo calculations, I don't have an Internet connection. Plus a stand alone mathematica would be best for importing and running M language scripts.
If a standalone offline version of mathematica was available for Android, I would be one of the first to purchase it.
How would you characterize your college experience? As you were so young it must have been difficult to engage in those crucial interactions with your peers outside of class, eg dinners out, parties where alcohol was involved, etc. Or were you more like the kid in the "Revenge of The Nerds" movie? ;-)
When I saw you and your son at the NY maker faire a couple years ago, your son - I believe he was around 13 at the time - did a demo on how to program a hexacopter to fly autonomously using WA. My son, who is a few years older, figured the kid was heading to Oxford by 15. Just wondering what the lad has been up to.
I know this will hit my karma, but here goes..
I'd ask Wolfram: Why do you say Alpha is so great. I understand it's hard. So then be clear about what it's good at. Why do you represent as if it's some all knowing AI when it goes to crap for any question that any self respecting SciFi fan would ask first?
It does really, really well with the examples that Wolfram uses in his introductory video. I get great results when I ask a stock market question, sure. I also get great stock market analysis results from Yahoo, Google, and e*Trade. In response to every other question I ask it I get crap. It typically has little data on my areas of interest, so it seems to dumb down the parse on my question to make a search *for* crap. So then it returns crap. Charted in one or two ways.
I want to ask things that would help me pilot a space ship, or at least help me understand NASA's proposals to the U.S. Congress. For instance, "How do I plot a course from earth to Uranus?" I just this moment typed that in, and guess what - complete crap. It returned a plot of x^2, and nary a mention of gravity or planets or time anything else. How did it manage to parse a question about a course from Earth to Uranus and decide x^2 was the best item to present?
Look what it says about its parse of my question: "Using closest Wolfram|Alpha interpretation: how do I plot a". What? "How do I plot a"? I did type a subject, folks! It didn't even try to get to the planets, orbits, gravity, anything. IT DIDN'T EVEN TRY! I see that if there are no knowledge frames in the system pertaining to my query, it seems likely to chop down the input. I'm learning more about how Alpha is implemented than I am learning about my query!!
Can't it at least show any historical paths that spacecraft have used between the planets? Can't it even show the planets? Can't it even cite procedural texts on how to do it? Can't it mention some of the factors that must be considered? I would like the result from an all-knowing AI to be an applet that shows a spaghetti line stretching out among the bodies of the solar system, and I would like to be able to adjust the launch date and see the planets move and see what happens to the spaghetti line.
BTW, that little Game of Life CA that displays while I'm waiting for my answer. Ha ha. I guess that's so cool. I confess, it does make me feel that some really thoughtful process is going on, just what marketing wants. For all that, what comes back - crap. Just makes it all the more disappointing.
I've tried Wolfram Alpha several times over the years, and I have never been able to find any value in it. For example, I've just put in "FIR filter", and it comes up with nothing, saying only "Development of this topic is under investigation..." However, every major search engine provides lots of information about that topic.
Next, I tried one of the examples provided by retrieving stock information about GE. I then selected "Ratios", and most of the resulting output fields were blank. This correlates with my overall impression that whatever potential Wolfram Alpha may have, it currently is half-baked, at best.
To be fair, perhaps I haven't been using it as intended. So, does Wolfram Alpha currently have any value, and if so, what is it? Can you provide some examples of things it currently does better than any other online system?
Wolfram Research is known around town (Champaign-Urbana) as a pretty unpleasant work environment. (See some of the comments here.) Why do you run a business this way? Is this on purpose, or is it possibly that your management skills aren't as good as your math skills?
Posting anonymously because C-U is a pretty small town.....
Your idea of "A New Kind of Science" received a lot of publicity when it first came out, but doesn't seem to have really caught on in the years since. Is the idea wrong, or has the rest of the science world simply not caught up with you? Do you know of any serious scientific investigations or developments that have resulted from it so far?
In comparison, Einstein's Special Theory of Relativity took a few decades to find its first experimental proof, and to eventually be fully accepted by science. Do you see that sort of process occurring with your idea, or is it dying on the vine?
One of the challenges facing biological scientists is the need to develop and employ diverse data structures, as well as use analytic techniques that often require rather advanced mathematical and statistical methods and theory that span multiple disciplines. In biology there is a wide spectrum of computer languages available and used to pursue such requirements. Although Mathematica has the potential for much wider application as some of the demonstration project, training videos, and example code on the Wolfram website prove, it still remains one of the lesser used languages for this purpose.
How can Wolfram as a company find new ways to promote, organize, and expand the use of Mathematica and the emerging Wolfram/Alpha language to become more of a visible presence in the biological sciences community?
In his younger years, Penrose was brilliant and made great contributions to mathematical physics. But virtually every serious physicist looks down on Orch-OR, and the only reason you don't hear the word "crackpot" being thrown around too much. His arguments for even needing such a proposal were discredited a good way back before he even hooked up with Hameroff, when he wrote Emperor's New Mind and Shadows of the Mind. A combination of wishful thinking for a non-computational basis for reality that allows human minds to escape Godelian limits, and, I dare say, senility, is what's really going on here. Quantum mechanics and quantum field theory are computable theories, and adding in thermodynamics gives you even stronger results (Bekenstein bound); virtually everyone but Penrose/Hameroff accepts that any 'final' TOE will be quantum in nature. It's sad to see Penrose giving keynote addresses with the likes of Deepak Chopra (http://www.edgemagazine.net/2014/04/consciousness-conference/), who also was one of the reviewers of Penrose/Hameroff's 2013 paper (http://www.sciencedirect.com/science/article/pii/S1571064513001188 scroll down for Chopra). Are you fucking kidding me?
"Politicians and diapers must be changed often, and for the same reason."
Any kind.
I'd imagine since it is now his brand, it'd be hard to get away from even if he really wanted to.
You can only write Mathematica in Emacs, obviously. Everything runs om Emacs.
nosig today
Since entropy in a region of finite extent and energy is bounded, it would appear that arbitrary precision real numbers are not physically realizable (otherwise, you would be able to store infinite information in a real-valued physical quantity, violating the bound). Unless one is a mathematical platonist (a religious position), that means real numbers don't exist. So why is it considered acceptable, other than for historic and/or wishful thinking reasons, to think about real numbers in a more serious manner than thinking about magical fairies, and it's still allowed to have much of mathematics relying on the assumption that uncountable infinities are a sensible concept? Mathematical thinking about real numbers directly maps to a finite physical process through said thought's neural correlates, and so is akin to a delusion. Where am I going wrong here?
"Politicians and diapers must be changed often, and for the same reason."
You have to define a problem before you can state that you need math to solve it. Some believe that a large cluster of simulated neurons is the way to go (but I have yet to see proof it would bring A.I. and is the only/most efficient way to bring A.I. ).
nosig today
Shouldn't he be typing that into Wolfram Alpha?
the preceding comment is my own and in no way reflects the opinion of the Joint Chiefs of Staff
About 15 years ago I found a bug that affected all Fourier-like transforms in Mathematica. (It was related to how the constants can be “allocated” between the exponent and an overall scale factor—someone had tried to generalize this concept by being too clever by half, and made a mistake.) I did a sanity check with comp.lang.mathematica or whatever the group is called and then filed a bug report. I understand that the error was not corrected until a later major release of Mathematica.
A few months ago I returned to Mathematica with a medium-sized project which involves some probabability calculations (PDFs, characteristic functions, etc.) I quickly found that Mathematica failed to crack an integral because it did not do a simple, trivial, second-semester substitution. I also found an error in the way a special function (MeijerG) is calculated numerically. In all, after only about three weeks of returning to using Mathematica, I filed five bug reports (one of which was UI-related) and have two or three saved up for when I get more time. I have watched the Mathematica release cycle for some years including the “dot” releases, and I am not encouraged that any of my reported bugs will be addressed before the next major release. (I believe that would be version 11.)
I have finally drank enough Kook-Aide to appreciate Mathematica and indeed have rather quickly (after my recent return) found it indespensible in my work; I am no longer even tussling with whether to use Octave/Matlab or Python/NumPy/SciPy for numerical work.
So: Why does Wolfram respond so slowly to bug reports? There seem to be only one x.1 or x.0.1 release after each major release, if that. Why not release more-frequent bug fixes like most other software houses, rather than let bugs exist for years in some cases?
Erim Radcliff
Oh, aspergers are good with butter.
Get free satoshi (Bitcoin) and Dogecoins
Both of those are available without any sugar.
I mean $300 for "hobbyist" software with zero support? C'mon.
"Win treats sysadmins better than users. Mac treats users better than sysadmins. Linux treats everyone like sysadmins."
EVERYONE should see this. Pages and pages of the same thing.
Positives seem to be free coffee, and flexible hours. That's it.
Negatives seem to be incredibly insulting salary, management which actively sabatoges employees, good people are running for the hills as soon as they can put in a new line on their resume, and overall that the owner ended up with some money and some name recognition, and decided to get some lackeys. It's not a real company, not run like a real company, and if it went public, I'd short.
I very much enjoyed "A New Kind of Science" -- thank you. Despite the criticisms (in some cases fair), it is an astounding achievement. In it, you write that you believe that the methods developed in your book could lead to a fundamental theory of the Universe. Has your opinion on this subject changed and has there been any progress (of which you're aware) along these lines?
How do you feel about the proposal to use Wolfram as an unit measure for Ego? http://www.aleph.se/andart/arc...
How have you seen the progress of Mathematica on raspberry pi?
Will Mathematica be available on windows for raspberry pi?
If it's full of self-aggrandizement you know it's Stephen Wolfram.
In Carl Sagan's book Contact, Dr. Ellie Arroway mathematically proved that Pi, calculated out to some huge number, had a series of 1 and 0 that when arranged in a raster, formed a circle, supposedly showing that the universe was not an accident.
While this is obviously fiction, is there any Mathematical equation, theorem, or any other aspect of Math that makes the hair on the back of your neck stand up or otherwise cause you to wonder?
When Fascism comes to America, it will call itself Anti-Fascism, and tell you to give up your guns.
You had an interesting essay "My Hobby: Hunting for Our Universe" on your blog in 2007 about modeling fundamental Planck scale physics via random networks (this was also mentioned in your NKS books). I didn't see anything written or spoken on that topic later. Did that project ever yield any recognizable physics or was it abandoned?
You are asking an a-hole "why are you such an a-hole"? Do you expect a coherent response?
Should wolfram's tech support be as advanced as wofram's products?
The ego that burns twice as bright burns half as long.
How close are we / you from a human-reasoning grade computer?
Slashdot, fix the reply notifications... You won't get away with it...
What questions are you asking?
How did the raspberry pi experiment pan out? Will Mathematica be available on windows for raspberry pi?
Mathematics already has the concept of multi valued functions like square root, for example.
http://en.m.wikipedia.org/wiki...
0/0 is undefined because you can reach literally any real number for 0/0 by taking a limit of a carefully selected function. For example, sin x/x winds up at 0/0 but has the value 1 at x=0. Now if you want 0/0 to equal some arbitrary number, multiply that expansion by the number and take the limit.
SJW n. One who posts facts.
I have run into this same closed mindedness in academics. I suggest reading Rupert Sheldrake's "The Science Delusion" -- YouTubeable -- also called "The Dogma Of Science". Then consider that universities had religious origins. After that, give up on academics and just post your theory on your own web site (where it will be largely ignored).
There is only so much you can do when a small group control everything. Better theories are the very last thing they want.
Good luck.
I come here for the love
One of the frequent concerns I remember from my days in physics (where Mathematica was frequently and heavily used) was the question of how scientists could trust the results of the program. To the best of my knowledge, (although I must concede my knowledge is some years out of date) no computer algebra system is currently regarded as being bug free. There is always the question: "how do I know that I got the right answer *this* time?"
While auditing the logic of one of the open source systems (Axiom, Maxima, Reduce, etc.) is at least theoretically possible, in practice the challenge is sufficiently difficult that anyone other than a dedicated specialist is unlikely to tackle it. The problem is compounded in cases like Mathematica and Maple, which are proprietary and closed source - even the theoretical opportunity for an in-depth audit is lacking.
The most promising approaches I have head put forth to address these sorts of problems are to add the ability to generate a “formal proof” transcript (targeting something like ACL2 or COQ) as a byproduct of the standard computer algebra system calculations, which can in turn be verified by a much simpler proof checker (or a very determined human.) What are your thoughts on the question of producing trustworthy answers to complex mathematical problems via computer? Is leveraging formal proof techniques a good way forward, or should we be looking in other directions?
"I object to doing things that computers can do." -- Olin Shivers, lispers.org
How do you think will Quantum Computers or more specifically Quantum Cellular Automatons impact IT?
How close are we to being able to have the kind of system you discuss in your TED talk on computing a theory of all knowledge, and how do you think we could use this kind of approach in practice to improve the reliability of information that people are reading on the web? How would you suggest it might be integrated within the current framework to provide possible utility outside of the Wolfram Alpha site itself, if at all?
Some mathematical savants express a type of mathematical synesthesia. That is, numbers are experienced as having textures, colors and shapes. Have you experienced this? How can this effect can be leveraged as an educational tool for the general public?
How would you explain the relationship between music and mathematics? How can this relationship be leveraged to help reduce innumeracy? (For example a two dimensional touch screen can be used as a visual theremin where each axis plays a frequency. The frequencies can be continuous or as steps in integer ratios corresponding to chromatic, pentatonic, pythagorean, blues, arabic and other musical scales.)
One function in all of mathematics begs that it be developed further. :-)
That's a great example of sin x/x. I'm hoping that Calculas would be extended that equations like this wouldn't be undefined.
Mr. Wolfram,
Will Mathematica progress to the point of being able to parse plain language such that it can extract the mathematical concepts into its underlying language for processing?
I have used the software for many years but find the input grammar required to be the single biggest barrier to putting the tool to its fullest potential.
-- I fear explanations explanatory of things explained.
There could be many good reasons for Wolfram to avoid giving money to the university that educated him.
1. He may have realized the quality of the administration of the University was not up to his standards.
2. He may have felt slighted or even blocked in some cases by faculty members who were embarrassed that he was a much better scholar.
3. He may have identified problems with some of the tenured professors who he did not think were deserving of their responsibilities.
4. He may have developed an extreme distaste for the politics of academia.
5. He may have sent the cheque to the Macadamia people instead, and he's still enjoying monthly shipments of their delicious nuts, and has decided not to correct the mistake.
If your only tool is a hammer, you'll approach every problem as if it were a nail. - Abraham Maslow
I don't think so.
Ad hominem attacks are lame.
If you have brilliant insight, please cite objections to the paper in question, which is quite interesting, and has enough experimental data to at least merit discourse.
If you HAD read the paper in question, you would understand the context I am asking the question. I doubt you have.
Clamouring to "mod parent down" smacks of professional jealosy or perhaps some other more basic inadquacy.
..don't panic
The Wolfram Language appears to be just a rebranding of the Mathematica language. What does it have to offer to the research community at large and to those who are not currently using Mathematica?
Say stephen, when do you suppose your ego will get out of the way of a serious treatment of evolution on Alpha. As long as you censor that, I'll be censoring you.
Looking back from when Mathematica / Wolfram Language started, what (if any) essential factors have shaped the coherent system for modern technical computing the way we see it today?
Mathematica and Alpha are for calculations and manipulating equations. There is another part of mathematics: Proof. For the most part, that's been written in English or Russian or some other human language. But there are a number of prototype tools for writing proofs on a computer.
Thomas Hales & co. just proved the Kepler Conjecture using HOL Light. Georges Gonthier and his team have proven the Odd Order Theorem in Coq. Vladimir Voevodsky is looking at a new foundation for mathematics based on his exposure to Coq.
What do you see as the future of formal proof? Will we see it as part of the Wolfram packages?