Writing and Speaking
I got a lot of comments about the two videos I posted last week suggesting that I have nailed the art of public speaking.
I don’t know about that, I am my harshest critic.
But I do believe that writing regularly makes it so much easier to speak publicly in unscripted situations.
Writing forces you to work out your views and articulate them clearly and concisely.
Then when you are asked a question related to those views, you have already worked out the answer.
It is in the brain, waiting there to come out crisply and concisely.
I’ve been writing daily for going on fourteen years so that is a huge body of work, opinion, thought, and insight to be able to pull from.
My views have evolved over the years and so not all of that content is relevant at this point, but most of it is.
So if you have to speak publicly a lot, particularly in unscripted situations, I would suggest you write publicly regularly as well.
They work incredibly well together.
This exact thought struck me when watching your interview video from a couple days ago. These interviews we see with you don’t give us readers much that is new. They rather reinforce what we know from reading the blog. I see them as an extension of this blog. Further, the interviewer will always come unstuck when looking for an outrageous comment from you because you have clarified your thoughts so consistently over the years of blogging. There is a lot of wasteful blogging on the web. A lot of rubbish. This one is not rubbish, and the practical application is clear in your interviews.
Extemporaneously endowed, an expert is.
Bingo. It’s like writing your speeches in anticipation of saying them later.Whoever thinks that great speakers don’t write their speeches first is wrong. I do that all the time. If you want to wing it, it shows.But all this requires that you have a deep knowledge of a topic or something. Then, you become more natural at it. Saying less is more impactful than saying more.
The true art of speeches comes from unrelenting practice. I practice my speeches over and over and over again so that when it is time to get up in front of people, it seems like it is flowing naturally.
As do I.The dynamics changes with the size of the crowd and the cut off from the audience.As long as there are dynamics, knowledge and passion is enough.As soon as you are disconnected and are speaking to large groups ‘out there’ beyond the lights, you are in the realm of performance.Even practice without true talent is not enough at times.
“Saying less is more impactful than saying more” E = mc2
I believe your counter-culture, balanced approach (e.g. Facebook) and high specialized early market ventures such as Cryptocurrency is invaluable to counter the mundane and obvious that exist on the web.
Worth sharing among the people who do not speak too often and never really write. Probably they will get more speaking opportunities when they start writing – it helps clarity of thought as you said, and it adds to the personal brand too.
So true with commenting as well. I find I’m much better at firing off pithy retorts that restate and amplify someone else’s insights in conversation as a result. 😉
Fred–I am sure you are correct.But need to say, that the reason I’m here, is less about the clarity of expression and more about the breadth of your exposure that feeds the funnel of your thinking.One is skill. The other is a perspective truly unique and special.
I thought the Doriot lecture you gave last month at MIT was a good example of public speaking.Curiously, I didn’t think of the two videos last week as public speaking…that they were 1×1 public conversations.Great point on the writing.The Feynman Technique is a simple and wonderful tool to learn that involves writing and then explaining. Read-Write-Explain-Iterate. A great technique to identify gaps in understanding and go back to fill them in.
.The research-write-read-revise approach is what makes for good writing. Even fiction has to be converted into the spoken word to see if it “sounds” correct.When you read dialog you have written, you are seeing it in its native and intended method of delivery — verbally.JLMwww.themusingsofthebigredca…
Starting in about 5th grade, I went through a process with my son for all of his writing assignments for school. He would write a draft, then read it out loud to me. He would often find his own errors by reading it out loud. Then he would write a second draft and repeat. We would repeat this for as many times as was necessary for him to write without errors.As he grew older and better at it (less drafts to find the errors), I added in editing. I would go through each draft and write comments about word choice and to note where his thoughts were unclear, then we would go through the draft together and he would rewrite it.Some common themes were “Be more specific!” and “You need more details about this idea. Add another paragraph to explain it with more detail.”This continued until he went to college. It was painstaking and mind-numbing work for me most of the time, especially because my son hated it and he often fought me on it. Writing was never his strength but I insisted he learn it because I knew he would need it in life.After his first semester in college he thanked me for it. He said that many of his peers in college were struggling with knowing how to write good papers and he felt that he was lucky that I had worked with him on it all those years. It took a lot of time and lot of years but it’s worth it.
Oh yeah, and I also gave him writing assignments all summer long when he wasn’t in school so that he would not loose momentum.
Great story! This is what countless Moms do every day, fully knowing that their kids will hate them in the moment but is the absolute right thing for them in the long term.
.One can’t fully appreciate a child until they become a teenager and have command of a fuller slice of the language with which to bludgeon you.Then, miraculously, they get to 25 and they remark how much more intelligent their parents have become since the child has gotten on a first name basis with all of their bills.JLMwww.themusingsofthebigredca…
Your son is a lucky young man
Susan,If only mothers like you, Kirsten, panterosa, Donna, Anne and others showed up at AI workshops when male PhDs say silly things like “Babies don’t need a lot of training. They can take in a lot of data and just do unsupervised learning. That’s what our algorithm can do … like babies” my life would be easier.Because those male PhDs have obvious amnesia about how they acquired data, knowledge and context as babies.The SCIENCE, the facts and the truth is that women CO-CREATE the basis for the intelligence of human babies from the moment of conception. Even our habit of talking to our baby bumps is a form of training them to recognize voices, understand language and make sense of presence and emotions.Then when those little boys become adults and want to conquer the world of work, intelligence from their mothers / wives / sisters / female colleagues mentor, steer and sanity-check them. https://uploads.disquscdn.c…
LOL! I’ve been showing up and disputing men with silly beliefs my whole life. Never going to acquiesce!
@cavepainting:disqus — You asked me what I do.I’m doing this on 20 June.https://www.eventbrite.com/… https://uploads.disquscdn.c…@JLM:disqus @InformationShield:disqus @philipsugar:disqus @ccrystle:disqus @SixgillBlog:disqus @wmoug:disqus — There’s a lot of TALK about diversity problems and data bias problems in AI. Meanwhile, I do “HANDS-ON, get your hands dirty with the code in the sandpit” workshops.Frankly, the “AI has a sea of dudes problem,” “AI is sexist, racist, biased,” “AI biases against your ability to find employment opportunities” headlines are boring and uninspiring.And we all want to be reading less of those types of headlines next year.And we certainly don’t want our kids to be held back by silly biases that are completely fixable in the systems by male and female developers and UX designers, alike.
Great. I will try to attend this if possible. I also live in the bay area and would love to catch up some time.
Look forward to seeing you there. Sign up!
Awesome. Good luck Twain!
.You know a lot of interesting and useful stuff. I doubt two people on this blog have any idea what the Feynman Technique is. I’ve been using it since I was in college though I didn’t know the name then.Feynman, a brilliant physicist, is one of the most interesting people to have walked the earth. Maybe the smartest physicist in the history of physics. He made a perfect score on almost every physics test he ever took.When being admitted to Princeton, he was the subject of anti-Semitic discrimination and it is hard to believe such a discourse actually happened. He was admitted because he “didn’t look Jewish.”After his death, his notebooks were transcribed and they were filled to overflowing with grammatical and spelling errors. There is a series he published, can’t remember the name, but it’s something like the Feynman Lectures on Physics.You could teach yourself physics by reading that book. It was really Physics for Dummies.Well played, Girish Mehta.JLMwww.themusingsofthebigredca…
I use the Feynman Technique to help myself understand a book better. Usually that means I have to go back and read portions of the book again…and again.There is a story…don’t know if true but its the kind of story that would be true of a mind like his and a man like him.One of his colleagues asked him to explain a concept/equation (don’t remember what it was, something about quantum statistics). He said he would prepare a freshman lecture on it instead. Some days later he came back and was empty-handed.He said “I couldn’t write it at a freshman lecture level. That means I don’t really understand it”.2 minutes of Feynman’s genius – Names Don’t Constitute Knowledge.https://www.youtube.com/wat…
.Great stuff. Thanks. Well played.JLMwww.themusingsofthebigredca…
I cannot tell you how much I like that video.
And he’s totally right. The AI and Cognitive Science crew are knotting themselves in all sorts of pretzels about names of things: perceptrons, general adversarial networks, consciousness etcetcetc.Just because we can give names to electronic processes of the machine “brain” doesn’t mean we know anything about the human mind or human intelligence, language and values.I wish Feynman was around now. He’d be someone I’d seek out as a mentor.
That is the best thing I have seen this week, this month and this year.
> There is a series he published, can’t remember the name, but it’s something like the Feynman Lectures on Physics.Just checked his Wikipedia entry , it is just that:https://en.wikipedia.org/wi… https://en.wikipedia.org/wi…IIRC I read this other book of his, some years ago:Surely You’re Joking, Mr. Feynman!https://en.wikipedia.org/wi…Pretty interesting, it was.
.Almost nothing as entertaining as a good physics book, no?JLMwww.themusingsofthebigredca…
Yes, I enjoy The Feynman Lectures on Physics — in places he gives better explanations than I’ve seen elsewhere. IIRC he explains the importance of entropy as an invariant. Okay.But, IIRC, he says that if there is a particle and we don’t know anything about it, then it has uniform probability of being anywhere. Tilt! Nope! Can’t do that! Can’t be!In probability, if events A and B are disjoint, thenP(A union B) = P(A) + P(B)More generally, if the eventsA_i, i = 1, 2, …(here using A_i as A with subscript i, borrowing from D. Knuth’s TeX) are disjoint, pair-wise disjoint is the same, thenP(union_i A_i) = Sum_i P(A_i)where we are working with the countably infinite sequence of events (A_i). The short description is that probability is countably additive. Uh, it’s not uncountably additive!Okay, for that particle we know nothing about, consider the volume of the universe. Assume that that volume can be covered by disjoint regions of one cubic foot each, say, regions A_i, i = 1, 2, … countably infinite.Then by the uniformity claim, suppose that for each iP(A_i) = aIf a = 0, then the probability the particle exists anywhere in the universe is 0. However if a > 0, then the probability the particle is in the universe is infinite, in particular, greater than 1 which violates the axioms of probability.Or, Feynman was making a claim: For some positive integer n, there are only i = 1, 2, …, n disjoint regions of volume one cubic foot each. Then, with the claim of uniformity, the probability the particle is in the 1 cubic foot on the lab table is exactly 1/n. So, if we can estimate that probability, then we can estimate n, the size of the universe! Or, for each particle of unknown location, the probability of its being in the 1 cubic foot on the lab table is from the volume of the universe.I doubt that Feynman was thinking of this second.Uh, Richard, in probability theory, there is no probability density function that is uniform on a set of infinite length or volume! Sorry ’bout that.For quantum mechanics, I’m still trying to find a presentation that doesn’t make me upchuck!(1) I bought a famous book. Early on he says that the wave functions are continuous and differentiable and form a Hilbert space. Wrong! Can’t be! They can be points in a Hilbert space, but with the usual inner product, norm, metric, and topology, they can’t be all of a Hilbert space because they are not complete, that is, so that every Cauchy convergent sequence converges as required by completeness.Note: By definition, a Hilbert space is some vectors and has an inner product plus some other nice properties.The short and correct definition of a Hilbert space is a vector space that has an inner product and is complete with this inner product — is a complete inner product space.For the inner product part, likely you’ve seen that before in high school algebra, physics, and engineering as the dot product — it is central to talking about length, distance, and angle. We get to talk about the cosine of an angle and get the Pythagorean theorem and generalizations of it.E.g., in statistics the statement that the total sum of squares is the regression sum of squares plus the error sum of squares is the Pythagorean theorem because that statistics for some positive integer n and the set of real numbers R is working in the Hilbert space R^n.The easiest example of a Hilbert space is the real line, that is, the set R of real numbers. There, for two real numbers x and y, their inner product (x,y) is just xy, that is, their product.What’s exciting about a Hilbert space is not the one dimensional real line but the examples that are multi-dimensional, finite or infinite, even uncountably infinite dimensional. What’s exciting is that a lot that works in the real line, the plane, and the 3-space it usually appears that we live in carries over to the other examples. So, we still get to have a version of the Pythagorean theorem, length, angle, converging sequences, linearity, convexity, etc.The most common inner product is the one in linear algebra and there in matrix multiplication or, more commonly, for positive integer n, a system of n equations in n unknowns.To keep the notation short, that’s the same as Ax = b where A is an n x n matrix, x is n x 1 (a column vector) and b is also n x 1. Then in the Ax part, each row of Ax is the inner product of that row of A with x. So, if A = [a_ij] and x = [x_j], i, j = 1, 2, …, n, thenSum_j a_ij x_j = b_iand the left side here is the inner product of row i of A and x.Or, even simpler, but with less context, for vectors x = (x_i) and y = (y_i), i = 1, 2, …, n, their inner product (x,y) isSum_i x_i y_iThis inner product is so common in computing for math, physical science, engineering, and even social science that it is the most important vector instruction in a computer, e.g., an old Cray, and apparently in any Intel processor since 2006 with the SSE4 instruction set extensions.This inner product is by far the largest consumer of computer arithmetic in linear algebra and multi-variate statistics and, thus, in regression analysis, classification and regression trees, AI and ML.Uh, for a little more, the vector space of real valued random variables X so that E[X^2] is finite, with the usual inner product (X.Y) = E[XY] actually is complete and a Hilbert space. Amazing but true. (2) Watching some MIT lectures, (A) he says that every wave function is differentiable and also continuous. Of course it’s continuous you boob: Every differentiable function is continuous. (B) He mumbles and fumbles with the Fourier integral (transform). Uh, guy, there’s a beautiful, totally clear, rock solid, perfectly precise, beautifully done chapter on the Fourier integral in W. Rudin, Real and Complex Analysis. Go do some math homework, and let’s try again.My tests in college physics were all perfect until I started to lose respect for the subject due to the sloppy math of physics. That sloppy stuff is why I majored in math instead of physics; took enough courses to major in physics but one of the courses was a reading course in mathematical physics instead of a specified course.The sloppy math in physics continues to bother me. By now I’ve gotten enough solid math to be able to clean up a lot of the sloppy math of physics.Sure, when I’m done with my startup, back to mathematical physics, clean up some more of that stuff, and then start to attack the main questions!Gee, it’d be neat if the probability a particle was in a box on the lab table could be used to measure the volume of the universe, but I don’t believe it.Here’s a clue I might want to follow: Have a weak light source that puts out a photon every now and then. Run that through a beam splitter so that half of the beam goes through and the other half goes off at 90 degrees. Using a mirror for the first half, have it go in a direction opposite to the second half.Have two photo detectors, one light hour apart and positioned so that each half the beam hits one of the detectors.Okay, the beam splitter sends half of the wave function to one detector and the other half to the other one.So, if the detectors are really good, then get exactly one detection. So, in effect, when one detector gets a detection, it sends a signal much faster than the speed of light signal to the other half of the wave function telling it to disappear and not cause a signal at the other detector. There might be a clue here about more that is going on behind the scenes in the universe.Next, look into the fact that each piece of the wave function both feels and has gravity — hmm.
.I want Fred to evaluate your answer and decide how many credit hours we can all get. Nice answer.JLMwww.themusingsofthebigredca…
See Einstein’s gravitational waves detector:https://phys.org/news/2017-…
let’s also not forget many of us live in a 140 char world. brevity and directness is the price of entry.
i don’t think everyone living in a 140 char world is a good thing. just because we can, doesn’t mean we should.
Practice writing everything as haiku. One of my favorite things to do. It helps create clarity. When I was in college, my roommate and I would practice conversing only in haiku every evening at dinner. Someone would say something, then there would be silence while the other person sat thinking out the response in perfect haiku. She and I still often send each other haiku messages all these years later.
I’ve been an AVC fan since 2007(?) – and every week I’ve told myself to start blogging everyday… sigh…I am presenting Innovation to the Chamber of Commerce in Philly on June 2nd on behalf of UnitedHealthCare, if there are any fellow AVC’ers around, would love some friendly faces in the crowd…
I’ll show up. I’ll be in MN on Wednesday. I used to work with Phil McKoy the CIO. On calls it could get confusing on which Phil somebody was referring to. So he said I needed to be referred to by my last name and I said he could just be called “the real”
I’ve noted to myself your writing to be particularly well-articulated as of late.’It’s so artsy, you could sell it on Etsy’ — that just came to mind, sorry :-)More seriously, the moral, I suppose, is assiduously putting in one’s 10,000 hours pays off.
.There is a difference between a prepared speech and an interview in which one answers questions.Fred has turned into a “good interview” not by what he says — which is at the top of the food chain based on experience alone — but what he does not say. The traps and rabbit holes he doesn’t follow.When the delivery methodology is visual, the world is filled with way too much bullshit. When someone knows their subject and uses fewer words to answer a question, its authority is enhanced by the economy of the message.I always remember Gen Colin Powell’s answer as to what the plan was to drive the Iraqi army out of Kuwait — “First, we’re going to cut them and then we’re going to kill them.”There is a superior message, exquisite clarity, and retention in fewer words.Years later, I had a chance to visit at length with the MG (later 4-star) who commanded the 101st Airborne Division in that fight. The 101st made the longest airmobile insertion in the history of warfare and set up shop in the Iraqi’s backyard whereafter they unleashed their artillery and combined arms to destroy anybody within range.The Iraqis never considered you could move a division 700-800 miles in one jump. Therefore, they were not deployed to counter that threat. It was a brilliant bit of soldiering.The 101st Abn Div CO said that phrase was repeated a million times at planning sessions, during operations, and when explaining the mission of the 101st to their troopers.In brevity is both wit and wisdom.JLMwww.themusingsofthebigredca…
Couldn’t agree more. I actually minored in creating writing back in college. Since then I have written many hundreds of presentations and whitepapers. (Not nearly with the proficiency of Fred.) It does help force a person to organize thoughts. And more than once writing a “positioning” paper led to some creative product ideas.I am concerned for our kids. I think both writing and interpersonal skills are suffering in the age of twitter and Facebook.
Don’t worry, AI will write and communicate everything for everyone.https://www.youtube.com/wat…
AI will even write our poems. There’ll be no need for Creatives or creativity. We’ll push a button or say “Alexa, write me a poem” and the machine will think it all for us. https://uploads.disquscdn.c…
.Please do not tell anyone in my writer’s group. Reading some of their submissions is worse than a barbed wire enema.JLMwww.themusingsofthebigredca…
Everyone in AI knows the Natural Language Understanding problem remains the “Holy Grail” that no one has cracked. https://uploads.disquscdn.c…I joked with a friend that when AI can differentiate between Aristotle’s writings and Turing’s and understand it, then we might get somewhere.AI researchers have not even thought of doing this thought experiment yet. I have, tho’. https://uploads.disquscdn.c…
I did my Masters Thesis at MIT on Natural Language Processing and AI (in 1985). Back then we thought AI would be replacing doctors and drivers by 2000. Still a way to go. (Thank goodness)
I’m so happy you call it what it is Natural Language Processing rather than Natural Language UNDERSTANDING.The same NLP algorithms used today have been around since the 1950s. The only difference is more data and faster processes.However, this still doesn’t mean the machines understand language.That would require some fundamental re-engineering of 2000+ years of Aristotelian logic, Descartes’ deductive-inductive logic trees and Bayesian probability which underpin data classifications, economic models and perceptron structures for ANN.And none of the leading thinkers in AI have the will or the knowhow to do that.
Truthfully, I am thrilled when my voice-to-text works with about 90% accuracy. I think AI has really made some nice inroads. But I agree that “understanding” is not likely to happen.Back then we were using rule-based engines and LISP programming. (((hell)))) Neural networks were just getting started. I think I missed the fun by about 30 years.
It’s now really easy to code voice apps. My second Alexa skill just got certified and released on Amazon app store.As with my Da Vinci quotes app, I was shocked to discover that in 12,500+ certified apps no one had thought of coding an Ada Lovelace quotes app!Well, sure I know all the words of Adele, Ed Sheeran and other contemporary geniuses but surely Da Vinci and Ada Lovelace should have voices too, right?!* https://www.amazon.com/dp/B…https://uploads.disquscdn.c…
Unbalanced parentheses!Rule based, hell? Right. Nonsense.
> I joked with a friend that when AI can differentiate between Aristotle’s writings and Turing’s and understand it, then we might get somewhere.You are getting at a point I tried to make: Natural language is analogous just to communications; the understanding, even any reasonably powerful processing, has to be in the destination of the communications.E.g., for Aristotle and Turing, there’s no hope, ever, of understanding the writing, the language. by itself. Humans don’t do that, either. They can’t. Nothing can. Nothing ever will.Instead what is just crucial, 90+% of the work of the understanding is what knowledge, information, understanding, thinking ability — call it intelligence — the receiver — human, computer, ET — already has that permits the understanding.Or, the understanding of language is the work of the intelligence of the receiver.Result: Until can program some real intelligence and give it background information, e.g., about Aristotle and Turing, there’s no hope for natural language understanding. The bottleneck is not parsing, lexical scanning, syntax, parts of speech, etc. but real intelligence.
The bottleneck in machine intelligence wrt language:https://uploads.disquscdn.c…Descartes created the conditions that led Turing and everyone after him down the wrong paths. https://uploads.disquscdn.c…@InformationShield:disqus — Don’t be sure “understanding” won’t happen in our lifetime. We can’t get there with existing NN (Deep Learning, Reinforcement, Generative Adversarial Networks, LSTM, Sequence2Sequence, Wavenet, attention gates, emotion units and all that hype jazz) and theorems+methods spanning everyone from Shannon to Hinton because they all fell into Descartes’ trap.Meanwhile, someone’s already cracked the Turing-Lovelace paradox, :*) — whilst Silicon Valley and Professors of AI are asleep at the wheel and not even aware such a paradox exists.
I don’t see that Descartes said anything wrong or created a “trap” in that statement. The statement doesn’t take us very far toward NLU or machine intelligence, but it is less objectionable than a lot of stuff parts of K-12 and four year college wanted me to swallow!For Turing, okay for what he said there. For Ada, sorry, honey, but they named a nice computer programming language after you!Apparently what NL, NLP, NLU, AI, and ML are trying to do is just some short term, low fruit, opportunistic, utility progress exploiting essentially just versions of empirical curve fitting. There may be some utility there, but clearly it’s not what any of humans, toddlers, cats, kitty cats, owls, killer whales, or even cuttlefish do; and there is no chance that such work will lead to anything close to real NLU. Again, once again, over again, yet again, the U part needs actual intelligence, and parsing, curve fitting, etc. are at best distantly related.
Wrt your “I don’t see that Descartes said anything wrong or created a “trap” in that statement,” you are in the same “can’t see boat” as Professors of AI and AI researchers en masse.Ironically, the problem with Descartes’ statement is exactly what you refer to in “the intelligence of the receiver”. Descartes’ statement is about EXTERNALLY OBSERVED events of the communicator and the receiver.His Scientific Method makes no provision for the INTERNALLY INTERPRETED understanding of the communicator or the receiver.Everything the human mind does involves the integration of externally observed events AND internally interpreted understanding of the individual and then the people that individual interacts with.
If we disagree about Descartes here, then there’s something wrong:First, I doubt that it’s fair to say that that statement was Descartes describing some “scientific method.”Second, the statement seems simple enough, nothing to argue with,Third, he was saying, say, if want to understand stars, then get data from stars and realize that the sights, sounds, etc. we get might be distorted by our receiving equipment and might not be all there is.Indeed, he was correct: Without a lot of scientific discoveries and equipment, we can detect with our eyes only visible light, with our skin infrared light but only quite intense, and with our skin ultraviolet light but only after we see we have a sunburn. But by now a lot that we know about stars has come from nearly all the electro-magnetic spectrum from long radio waves (from rotating neutron stars) to gamma rays (from gamma ray bursting stars).Fourth, I didn’t take Descartes as really talking much about what really goes on between our ears.Fifth, maybe Descartes saying, say, that if we want to understand a beautiful sunset, the sunset itself, outside of us, then we should concentrate only on the data we can receive, maybe by whatever means, and f’get about our personal, poetic, artistic, romantic emotional reactions. Okay. But that doesn’t mean forgetting about any understanding, testable theories, guesses, etc. between our ears. So, white light comes from the sun, and dust at high altitude in the earth’s atmosphere lets through the red light but scatters the blue light, etc.Moreover, for an explanation of sunsets, come up with that independent of any humans seeing it and an explanation that would work on Venus, Mars, and the planet of ET with no humans around at all, indeed, would work before humans existed.Sixth, I doubt that Descartes was thinking about how to program a general purpose digital computer to have the intelligence of a human.And for that programming, would very much like to know how the heck the natural brains — cuttlefish to humans — on earth work.How a current digital computer works, as we know far too well, we need to partition the workings into largely separate levels, e.g., start at least as low as transistors, work up to simple circuits, AND, OR, NOR, work up to, say, an adder circuit, then central processor, main memory, and I/O, then an operating system, then applications programs, and within an applications program lots of additional partitioning with subroutines, algorithms, data structures, and, then, finally beginning to focus on what the heck the purpose of the applications program is.So, doing something similar for natural brains, finally at whatever is analogous to an applications program, with subroutines, algorithms, and data structures we will have our first shot at how the heck Mother Nature does so well programming natural intelligence.But, so far the neurologists, geneticists, etc. can’t read the applications software source code if only because the in-line comments were omitted!Again, once again, over again, yet again, one more time, as we’ve discussed before, I suggest we start like Mother Nature does, roughly from the old remark that ontogeny recapitulates phylogeny, or how intelligence works is illustrated by how it grows from conception through”Daddy, I need a new dress for the 7th grade class party!”.”Yes, dear. Here’s my credit card. You and your mom go to the mall and get everything you need and want, shoes, gloves, hats, dresses,skirts, blouses, sweaters, scarfs, coats, rain coats, purses, …, far beyond my meager understanding as a male to describe — have a good time and don’t give a single, tiny thought about what it costs! If you can’t find what you want at the mall, the we will charter a plane and spend next week shopping in NYC, with our own 24 x 7 limo, side trips to Tiffany’s, Lincoln Center. And I want to get your ring size so that I can give a father’s dear daughter a father-daughter promise ring, a 2 caret blue, round brilliant cut sapphire in a white gold Tiffany setting.”Ah, VERY intelligent daughter!Hmm.So, start at toddlers or before. Same for kitty cats, puppy dogs, mice, …, cuttlefish.For any of that, I see no real connections with what poor, old Descartes, some many score years behind in technology, said.
The point is that Descartes’ deductive-inductive logic does not map to human intelligence or to language.And his work directly influences today’s AI and computing structures. https://uploads.disquscdn.c… https://uploads.disquscdn.c… https://uploads.disquscdn.c…
> The point is that Descartes’ deductive-inductive logic does not map to human intelligence or to language.Maybe it does “map” but not easily! A field of wheat “maps” into flour and a Sacher Torte in Vienna, but not easily! There is a LOT to a Sacher Torte!”Descartes’ deductive-inductive logic” — fine. Just like other good tools, the question is what can be done with the tools.E.g., what did Vivaldi, Julia Fischer, Stradivari and other Cremona violin makers do with some spruce, maple, varnish, cat gut, horse hair, and tree sap rosin?https://www.youtube.com/wat…Then along came Bach and his Chaconne, Busoni to arrange for piano, Valentina Lisitsa, to play ithttps://www.youtube.com/wat…and get a top, center crown jewel of civilization, from some maple, spruce, black marks on white paper, etc.Again, once again, over again, yet again, one more time, you have found some good, maybe even great. simple, low level tools and critically observed that they are not ready made solutions for some astoundingly difficult, so far unsolved problem.Uh, some people can do more with black marks on white paper and horse hair than others.And you have observed that other people are floundering with this problem. But their floundering is no indication of anything wrong with Decartes’s tools.Are AND, OR, and NOR gates also no good because they are not ready made solutions to NLU and other people use these gates and flounder? Nope.Pay attention to music for a while and begin to conclude that those two women have just exquisite insight into that music. That is not the only performance of the Chaconne I like, and in some places and in some respects overall I like, say, Heifetz better, much better, but those two women are just darned good — crown jewel of civilization quality.And there’s some more by Heifetz I like, e.g., by Bruch and Sibelius.And basically it’s all from just really simple means — black marks on white paper.Descartes’s tools are fine. IMHO, solving NLU with those tools is difficult; without those tools, impossible.If you want a solution to the NLU problem, then, fine. Get to work. Have some good, new ideas. Here you will no doubt make good use of a lot of simple, old tools, including the ones from Descartes. But if you get a good solution, then that will be to your credit, not Descartes. But for solving the NLU problem, I believe you have a lot of work to do.I can’t recommend attacking such a difficult problem, and definitely not via “go for broke” — I saw an astoundingly brilliant women die trying that. But if you want, then go for it, and good luck.I need to get back to my startup: There I’ve already been successful with my original, crucial, core technical work, and the rest is routine but needs to be done.
S = k log W
When the AI can write and communicate in all the forms we can (scientifically, technically, prosaically, for business, in multiple languages) without loss of meaning and resonance, then Natural Language Understanding will have been solved.A few stanzas from my poem ‘Orbit’ … https://uploads.disquscdn.c…
.Nice poem. Very talented. Well played.I won my 5th grade poetry prize at Our Lady Star of the Sea. My older sister and my mother deserved co-writer chops. It is not true that I was only a competent scrivener, but I admit to needing a transfusion of creative energy.Sister Mary pinched my left cheek at the award ceremony. It was my first published combination of words. In the last year, I have had a number of short stories published by respectable journals.I have never written another poem since then.I can still recite that poem from memory. Shall I? Think of it like Purgatory.As you may know, I have a deep seated animosity toward poets. It is unexplainable. Might be Sister Mary’s fault.JLMwww.themusingsofthebigredca…
I just snorted, I laughed so hard!
Do we want that? I guess I thought one of the more optimistic and utopian ideals of AI and automation more generally was to free humans up to do things like make art. Why have humans around at all if AI will be writing our poems too? Will AI also free us up from the labor of “consuming” the poetry “content” on our behalf too?
They’ll also pay us with tokens since they’ll be the issuing authority.Haha, I kid you not, there are AdTech folks in Bay Area looking into the AI writing all the copy and matching it to video image recognition.
That is fascinating.
good poetry flows from the depth of one’s soul. AI-driven poetry might sound good on first read, but is unlikely to be real, for it is some combination of what it has already assimilated. There are no authentic feelings or thoughts behind this jumble of words.An approach that combines the two might be fascinating. i.e. a human augmented by an AI poetry assistant that suggests alternate words, phrases, examples, etc. In artistic and creative domains, AI that assists, advices, and augments can be more effective than doing something by itself.
Yes, I agree with AI complimenting natural human intelligence and creativity. Interestingly, Walter Isaacson noted this difference between Alan Turing and Ada Lovelace’s versions of machine intelligence.”This inspired Turing to develop the imitation game. As no computer has passed the Turing Test thus far, Isaacson points to Lovelace’s symbiotic theory of human/machine collaboration. While Turing set standards by which a computer would be able to surpass humans, Lovelace foresaw a future where people and machines make each other better. This is a fundamental difference in the Lovelace approach versus the Touring approach:”A different way of looking at the way the computer age evolved is sort of Ada Lovelace’s way which is that computers and humans will evolve symbiotically. They’ll be partners. We will get more intimately connected to our machines and the machines will amplify our intelligence and our creativity will amplify what the machines could do. And we don’t need to try to create robots that’ll work without us. It’s kind of cooler to create this partnership of humans and technology or as she put it the humanities and engineering.”http://bigthink.com/think-t…
Pls see my slide comparing Turing with Lovelace and how Descartes sent Turing down the wrong paths in reply to sigmaalgebra.
“I think both writing and interpersonal skills are suffering in the age of twitter and Facebook.”Yup. The world badly needs more face to face, and less screen time. Lots and lots of bad things for society are going to happen if the trend continues unchecked.
I worry about this too. This is why I implemented my own personal writing program with my son when he was growing up (I wrote about it above under a JLM comment).
I could not track down your other post. But it would be cool if there was a web site/app/or guided tutorial on setting up writing for kids. I think they would be totally into self-publishing via the web too.
I sometimes wonder if what they really need is in-person guidance. So much happens on a screen these days that the personal interaction is less and less.
Agreed. What I meant was some guidance on what to teach when, etc. I am making it too complicated. Just read and write with them in person. Some of the best things we can do. (Several studies suggesting that reading with your kids is perhaps the single biggest precursor to academic success.)
True enough. But I think they need guidance on how to continually improve their reading and writing skills. I also worked with my son on critical reading throughout all those years as well. It was a daily practice.Thankfully, his middle school required daily reading and journaling. The students were allowed to select whatever texts they wanted as long as it was at or above their reading level. That set the foundation for future reading+writing exercises.Each summer in high school he was required to read a certain amount of books — same deal, his choice. I added my own rules: half the books were his choice, and half were from a list I provided — I provided a big list so he could select his choices from my list.I read all the same books at the same time and we used them as kickoffs for dinnertime conversations and I also required him to keep journaling on each book. Some of the books he selected from my list were: Grapes of Wrath, Farenheit 451, Brave New World, The Prince, Catch-22….
Nuff said …https://uploads.disquscdn.c…https://uploads.disquscdn.c… https://uploads.disquscdn.c… https://uploads.disquscdn.c… https://uploads.disquscdn.c… https://uploads.disquscdn.c…
None of us knows what we truly think until we write it down. (apologies to Joan Didion, E.M. Forster, Flannery O’Connor and probably others who have said some version of this). It also helps to work out the nuances and connections to other things. That’s why you can speak more articulately + concisely and why you sound more intelligent when you’ve written it first — it clarifies to yourself what you think. Also makes it much easier to deliver presentations, lectures, elevator speeches, or answer questions about anything.
Joan Didion is a joy to read. I believe she may be one of the most under-read writers of our time. I prefer her journalistic pieces and essays to her larger works. She’s the female Hunter S. Thompson.And, man, who can’t love her here: https://uploads.disquscdn.c…
That’s so 1960s. Pic must be 50 years old. My aunt has a Stingray of about that vintage.
I know. Totally badass!
At least a bit ahead of her time. Women weren’t that confidently aggressive back then. Or, were you talking about the car?
The whole photo, the whole Didion is completely badass!
I’m trying to guess whether you’re a fan of the persona, the writer, or the whole package. I do like the quote — have often been surprised to discover what I know or think after writing it down.
Well, originally I fell in love with her writing. Then I fell in love with her person and persona. Then “The Year of Magical Thinking” came out and she was so brutally honest and raw and lyrical and complex, I fell in love with her writing all over again. And her risk taking. She blew up her own style and wrote something completely original. Only two books have ever brought me to tears. And this was one of them. (The other one was Wild Decembers by Edna O’Brien who is also a badass). I hope she keeps blowing things up and creating new things out of the shards.
My fave. http://www.pbs.org/pov/thir…
listening, speaking, reading, and then writing.
Hmm I was wondering this just yesterday, and there to my amazement is an answer in my inbox… Thanks Fred!
One of Dale Carnegie tips on public speaking- Know the material. Writing forces you to know the material…
It’s not only writing and speaking, but writing and thinking at the end.
Yes, I wish Dad had given me some advice: If you need to have a serious conversation with someone, boss, girlfriend, vendor, wife, child, judge, neighbor, etc. first write out nearly everything you’d like to say. Then revise that, over and over, for a few days.Results:(1) Will clarify what the main issues, facts, and points are.(2) Will come up with phrasing and wording that are much better than before this effort.(3) Will think about what the other person is thinking and help them resolve their fears, concerns, commitment to fairness, etc.(4) Will likely do better with light and some good results instead of heat and a fight.(5) Likely will not use all the phrasing, wording, facts, and points you reviewed, but really, usually no telling just what of those will need and the ones that are needed will have much better impact, that is, in place of some expression, presentation, or explanation cooked up at the moment.
When I teach my graduate students how to write, this is the first thing I play for them. The quality of this clip isnt that good, but its gets the point across. Your answers on politics to that reporter were definitely in this vein. :)https://www.youtube.com/wat…
Writing, speaking… and doing.When these three are coherent it is an unbeatable combination, which is the case. 🙂
It also helps that you are an investor, who needs to spend most of the day digging through vast quantities of information and stories to find out what the big point is that is actually hidden in those slides (or not).
Nice post. I’ve always thought of writing as a way for improving my thinking, but not necessarily my speaking. You make a good point though, that by improving and clarifying your thinking you are in turn improving your ability to speak about these thoughts when required. At least, when speaking you only have to think about the words, not the ideas on the fly.
I enjoyed your blog.