Humans are wired fundamentally to be irrational - our perceptual/cognitive apparatus is deeply flawed - umpteen studies show this - so this is a given.
But, we also discovered a way to think/model which seems to work amazingly - which is the scientific method or reasoning. But this language is not natural to the way humans operate at all. It is a struggle for most of us to think in that manner.
thats why math/science is difficult for most of us, and these were discovered only in the last 2000 years.
LLMs cannot yet represent conceptual relationships deterministically/symbolically. At some point in the future, perhaps they can, but the current generation has a long way to go.
take 3 ideas that are hot: chiplets, cloud, and LLM - remix them into the title of a paper that describes a hypothetical machine.. academia playing catch up and trying to stay relevant in my cynical eye.
prove that there are no non negative numbers less than 3
bullshits an answer with confidence (all llms do this)
stupid monty hall
Suppose you're on a game show, and you're given the choice of three transparent doors...
stupid river crossing
A farmer with a wolf, a goat, and a koala must cross a river by boat....
basically, these LLMs have ingested canned solutions and cant reason with newly defined concepts. Anything "out-of-the-box" and they BS canned answers - like the rote student. The BS is particularly distasteful because of the confidence projected in the answer...
So, they are great for looking-up commonly understood "in-the-box" narratives, but are poor at reasoning where there is some novelty. this is what we can expect from a probabilistic "deep" autocompleting machine. unlike a child which can learn ideas and metaphors from a few examples and anomalies.
failed all the logic puzzles with slight tweaks - including stupid monty hall (with transparent doors). BSs with confidence.
agi is not knocking at the door.
enterprise software architects trying to wedge into this emerging area, and you soon start hearing of: provenance, governance, security postures, gdpr, compliance.. give it a rest architects, LLMs are not ready yet for your wares.
Please provide this reference in your readme / blog as it is the original source for your work... and provides the background for the tradeoff between the 2 approaches: 1) fine-tuning vs 2) Search-ask
imo, it is injustice to present a clean capsule of the subject in a book,
and expect students to digest it - whereas it took hundreds of years of discovery to arrive at the concepts in that book.
i would love books that illustrate the struggles, and problems that drove advancements and how or why the ideas were invented in the first place.
to that end, i think gilbert strang's linear algebra is decent.
The question is ill-posed imo. I would invert the question and ask: "How not to suck at your work" as that would lead to similar conclusions, and is more actionable.
This essay has too many weasel sentences like:
"Boldly chase outlier ideas,"
"Husband your morale"
"Doing great work is a depth-first search whose root node is the desire to. "
"Curiosity is the best guide."
This is woolly-feel-good writing that chatgpt and folks like steve pinker, deepak chopra etc specialize in, ie: a bag-of-words about fuzzy feel-good ideas we all want to hear.
It is not confirmation bias. It is a different tendency.
It is a tendency to absorb whatever we read or hear - regardless of its information content.
IMO: Human languages evolved more for bonding than for conveying information. So, words exchanged were more to soothe the emotions of individuals involved rather than convey useful information about the state of their environment.
Hence we are fundamentally wired not to process the information content critically but to be soothed emotionally by whatever we read or hear
Proving a problem is np-complete should not be news. what should be news is when a problem has a P algo. (example, primes is in P)
my cynical eye sees this as an over-eager grad student rushing out his/her discovery onto hacker news. Next thing you know, an FPTAS for it may rear its ugly head.
how can they make a sale to enterprise IT without a giant architecture diagram ? The next thing you know, java may rear its ugly head as well just when we thought python eliminated it completely for these applications.
But, we also discovered a way to think/model which seems to work amazingly - which is the scientific method or reasoning. But this language is not natural to the way humans operate at all. It is a struggle for most of us to think in that manner. thats why math/science is difficult for most of us, and these were discovered only in the last 2000 years.
LLMs cannot yet represent conceptual relationships deterministically/symbolically. At some point in the future, perhaps they can, but the current generation has a long way to go.