Found in 22 comments on Hacker News
PaulHoule · 2024-06-10 · Original thread
At my workplace they recently enabled Microsoft's Copilot which is basically OpenAI + RAG.

It's a joke. It is painful to watch how slow it types. About 30% of the time it doesn't work at all. I asked it how big a particle accelerator would have to be to have 1 PeV protons and it got the right equation for the radius of the ring but it told me the ring would be like 10^34 meters in radius which is about 10^24 times too many because it got the units totally wrong.

It is completely unable to help a person get 5kg of antimatter, make a nuclear weapon or a deadly virus that will kill everybody because... it can't. People are afraid it will try so it has been taught to claim that it won't do these things.

It is also completely unable to put a list of items (say 20 or more) in a specified order. For instance it falls down on "list isotopes that decay by emitting a positron in order of half-life" or "list US states in order of how many characters in its name", etc. It can't do it any more than it can help you get 5kg of antimatter, but it lies and says that it can, will attempt and get it wrong, apologize when you point this out, try again, get it wrong again, endlessly. If you like pushing bubbles around under a rug, gen A.I. is really for you.

Now yes, LLMs can solve difficult problems in information extraction (say relationship extraction) that we were lost at sea with 5 years ago. But intelligence, truth and such are all difficult philosophical concepts.

People who know better act insulted when I remind them that neural networks don't repeal the laws of computer science (e.g. Godel, Turing and all that.) They can't make P=NP, they can't solve the very serious problems of logic + arithmetic that Godel warns about, but they are very good at telling you something that will bypass your defenses (I think this comes out of picking the "most likely" word because it won't surprise you) and that you want to hear (what else is RLHF?)

One real danger of LLMs is that they are better at seduction than we are because they don't have a self to manifest itself and get in the way. I am very worried about LLMs taking over dating sites, being used to run "pig butchering" scams and similar sorts of evil.

See

https://www.amazon.com/G%C3%B6del-Escher-Bach-Eternal-Golden...

to get some idea of how LLMs could appear to be 95% of the way there but still be structurally wrong to solve real-life problems that involve logic, arithmetic and everything else.

PaulHoule · 2023-08-30 · Original thread
If you look at papers in arXiv you see there is

"zero shot" = ask an LLM to do it with a prompt

"few shot" = show a model (maybe an LLM) a few examples; LLMs perform well with "in context learning" which means giving a prompt AND showing some examples

"many shot" = train a model with many (typically 1000s) of examples.

The more training examples you have, the better results you get. A lot of people are seduced by ChatGPT because it promises fast results without a lot of hard work, rigorous thinking, and such, but you get back what you put in.

My RSS reader and agent YOShInOn uses

https://sbert.net/

to transform documents into vectors and then I apply classical ML techniques such as the support vector machine, logistic regression, k-means clustering and such. I used to do the same things with bag-of-words model, BERT-like models give a significant boost to the accuracy, are simple to implement, and run quickly. I can write a script that tests 1000s of alternative models a day.

The main classification YOShInOn does is "will I like this content?" which is a rather fuzzy problem that won't retest perfectly. I tried applying a fine-tuned model to this problem and after a few days of trying different things I developed a procedure that took 30 minutes to make a model about as good as my classical ML model take took more like 30 seconds to train. If my problem wasn't so fuzzy I'd benefit more from the fine tuning and someday I might apply YOShInOn to make a training set for a better defined problem but I am delighted with the system I have now because it does things that I've dreamed of for 20 years.

The whole "prompting" model is dangerously seductive for various reasons but the low-down is that language is treacherous. This classic book

https://www.amazon.com/G%C3%B6del-Escher-Bach-Eternal-Golden...

is not such an easy read but it contains some parables that explain why making a chatbot do what people would like a chatbot will be like endlessly pushing a bubble under the rug and these problems are not about the technology behind the chatbot but about the problem that they are trying to solve.

uslic001 · 2022-02-25 · Original thread
I was just re-reading this book last night. https://www.amazon.com/G%C3%B6del-Escher-Bach-Eternal-Golden...
eindiran · 2020-07-14 · Original thread
For a detailed perspective on how the proof works, I highly recommend Ernest Nagel and James Newman's book Gödel's Proof [0], mentioned in the article. Alternatively, Gödel Escher Bach by Douglas Hofstader is a classic which serves as a great (and more accessible) introduction to the proof [1].

[0] https://www.amazon.com/G%C3%B6dels-Proof-Ernest-Nagel/dp/081...

[1] https://www.amazon.com/G%C3%B6del-Escher-Bach-Eternal-Golden...

formalsystem · 2019-08-18 · Original thread
There's so much more to CS theory than big O notation and you're in for a treat if you check out any of the below - happy to share more if people are interested.

1. https://www.amazon.com/Computational-Complexity-Approach-San...

2. https://www.amazon.com/Quantum-Computing-since-Democritus-Aa...

3. https://www.amazon.com/G%C3%B6del-Escher-Bach-Eternal-Golden...

4. https://www.amazon.com/Introduction-Theory-Computation-Micha...

jyriand · 2016-11-27 · Original thread
If this seems interesting to you, I would suggest reading: "Gödel, Escher, Bach: An Eternal Golden Braid" [1]

[1] https://www.amazon.com/G%C3%B6del-Escher-Bach-Eternal-Golden...

jrs235 · 2015-06-07 · Original thread
"I want to find the absolutely crazy-to-grasp book of all times."

Gödel, Escher, Bach - An Eternal Golden Braid, by Douglas Hofstadter. Anytime I start reading a story which contains recursiveness my mind will feel warped and stuck in a loop at times.

http://amzn.to/1KijebX (affiliate link)

http://www.amazon.com/gp/product/0465026567 (non affiliate link)

gipp · 2015-05-28 · Original thread
There's nothing about the idea of physical consciousness that says it has to be a continuum -- there could just be some critical mass or qualitative attribute of brains that puts us "over the threshold", so to speak. Nobody can give any kind of a definitive answer. For ideas about a "continuum" of consciousness, you might read Phi:

http://www.amazon.com/Phi-A-Voyage-Brain-Soul-ebook/dp/B0078...

Or for other views, you might check out V.S Ramachandran (neuroscience): http://www.amazon.com/Brief-Tour-Human-Consciousness-Imposto...

Jeff Hawkins (computer science): http://www.amazon.com/On-Intelligence-Jeff-Hawkins/dp/080507...

Hofstadter (mathematics, cognitive science): http://www.amazon.com/G%C3%B6del-Escher-Bach-Eternal-Golden/...

Those are some of my favorite popular-press books on the subject.

brokentone · 2013-11-21 · Original thread
This reminds me of Godel's incompleteness theorem - which I'll poorly present as: Any system that is sufficiently complex and complete will contain legal assertions that will disprove or destroy the system. (Those that do not are not complete).

http://en.wikipedia.org/wiki/G%C3%B6del's_incompleteness_the... http://www.amazon.com/G%C3%B6del-Escher-Bach-Eternal-Golden/...

thirsteh · 2011-08-18 · Original thread
Highly recommend Gödel, Escher, Bach: An Eternal Golden Braid by Douglas Hofstadter for those who are interested by this: http://amzn.to/qJVAAF
JonnieCache · 2010-12-20 · Original thread
If you enjoyed this article, and haven't read "Gödel, Escher, Bach: An Eternal Golden Braid" by Doug Hofstadter yet, then stop reading HN immediately and order yourself a copy:

http://www.amazon.com/G%C3%83%C2%B6del-Escher-Bach-Eternal-G...

Having read this book would, in a fair world, be worth more on your CV/resumé than a large proportion of comp-sci degrees.

Even better, go out of your front door to a real bookshop and get them to order you one. Who knows, you might speak to someone! BONUS!

helwr · 2010-09-12 · Original thread
Well, there are some related topics on Quora: http://www.quora.com/Randomness?q=random and http://www.quora.com/What-are-some-good-books-on-random-proc...

You can actually ask your question there as well in case this question gets unnoticed on HN; Quora people are very smart and pretty responsive

see http://en.wikipedia.org/wiki/Stochastic_process, http://en.wikipedia.org/wiki/Random_walk and do a search for Random Processes or Stochastic Processes on Amazon bookstore

Read about Entropy: http://en.wikipedia.org/wiki/Entropy A good book on Information theory can help you put it in context: http://www.amazon.com/s/ref=nb_sb_ss_c_1_18?url=search-alias...

Check out GMP http://gmplib.org/

If you're philosophically inclined read some existentialists, they deal a lot with irrationality and chaos: http://en.wikipedia.org/wiki/Existentialism

If you're financially inclined read Random Walk Down Wall Street: http://www.amazon.com/Random-Walk-Down-Wall-Street/dp/039331 and the Black Swan: http://www.amazon.com/Black-Swan-Impact-Highly-Improbable/dp... you may want to check out his other book as well, it is rather non-technical: http://www.amazon.com/Fooled-Randomness-Hidden-Chance-Market...

To learn more on how Wall Street deals with the stock market randomness read some books on Time Series analysis and forecasting, e.g the classic http://www.amazon.com/Time-Analysis-Forecasting-Probability-...

If you are a data scientist in heart read this great Q&A thread: http://www.quora.com/How-do-I-become-a-data-scientist

I wish I could help you with a link to a clear non-technical introductory article but this is all I've got. As random as it gets:)

Probably some good introductory book on science will fit the bill, science after all deals primarily with randomness. You may want to check out http://www.amazon.com/G%C3%B6del-Escher-Bach-Eternal-Golden/...

devinj · 2010-06-14 · Original thread
Very much relevant to anybody who found this interesting: http://www.amazon.com/Godel-Escher-Bach-Eternal-Golden/dp/04...

I found it in a used bookstore for $5 once, totally by accident. It's probably the best nonfiction book I've ever started (I unfortunately never bothered to finish, which isn't so good. It's still in my backpack in case I ever get stuck anywhere...).

pook · 2010-05-08 · Original thread
MIT's OpenCourseWare is an excellent way for him to study CS on his own while in high school.

I would recommend, at 14, getting him utterly hooked on the mindset of CS and related subjects. Godel, Escher, Bach, etc. If you can get him fascinated with the field, he'll find all the information he needs on his own better than any list of required reading you'll get.

http://ocw.mit.edu/OcwWeb/web/home/home/index.htm

http://www.amazon.com/dp/0465026567

fara · 2010-03-29 · Original thread
Douglas R. Hofstadter's Gödel, Escher, Bach: An Eternal Golden Braid http://www.amazon.com/Godel-Escher-Bach-Eternal-Golden/dp/04...
stcredzero · 2009-05-19 · Original thread
If you want a good feel for this stuff, try downloading and going through this course, "NAND to Tetris": (Link is not the course, but a talk about it.)

http://video.google.com/videoplay?docid=7654043762021156507

The book is only $26, last I looked. The course software is all Open Source.

You'll see how digital logic is used to construct components like logic units and memory, which is then used to construct a computer, for which you create a computer language, which you then use to write an operating system, and finally, you program games on it.

This will then give you the wherewithal to really understand Godel-Escher-Bach:

http://www.amazon.com/Godel-Escher-Bach-Eternal-Golden/dp/04...

If you want, you can read the book first, but then go through the course and read it again. The 1st time you read it, much of it will be lost on you, but the 2nd time, you'll have many Ah-HA! moments.

One key is Automata Theory. Understand that, and you can understand what you are trying to ask about.

Fresh book recommendations delivered straight to your inbox every Thursday.