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

What we know about LLMs

willthompson.name

Post reply on HN
Print view View original

11–20 of 173 posts

Re: What we know about LLMs

#11

Post by go_elmo »

jannyfer wrote: post #7

Great summary. I’ve been reading a pop neuroscience book called Incognito (2011). In it, the author talks about how the brain is a group of competing sub-brains of many forms, and the brain might have several ways of doing the same thing (e.g. recognizing an object). The author also posited that the lack of AI progress back then was due to the fact that there are no constantly competing sub-brains. Our brains are alw…

In the end - an AI should have these competing subsystems in one system - just as our brains are one system. What I find extremely interesting is how perception and thinking differs from person to person too - it was a "taboo" topic to call this neurodiversity - just as other genetic traits, but AI makes this relevant more than ever imo. Sure, its complicated and much comes from nurture (Nurture vs nature.. as exposure/epigeneticd vs genetics) but there sure are markable differences - the ones starting to stand out are e.g. adhd / autistic people, but Im sure it wont stay just there over time!

Re: What we know about LLMs

#12

Post by washadjeffmad »

dsco wrote: post #4
blitzar wrote: post #2

You had me at "Crypto VCs & ”builders” making a hard left into AI"

As a matter of fact, there’s even more developers making a hard left into AI who have never touched crypto. The interesting follow up question is: what will they actually spend time on? Training new models? Copy pasting front ends on ChatGPT? Fine tuning models? I think many of them will be scared by how much of a hard science ML is vs just spinning up old CRUD apps

I'd imagine something like openBB / BB Terminal with consolidated API access for financial reporting, a platform for insider communities ("chat, forums, and an app!"), etc. Make it a club, and it'll sell itself.

Since investment has been demoed successfully with off the shelf models, I don't think we're waiting on big advancements to be able to build a product. The bar for something like this, short term, is 1) be cool and 2) lose less money than traditional investing, sometimes.

Re: What we know about LLMs

#13

Post by killernap »

ChatGPT was announced November, 2022 - 8 months ago. Time flies.

Question for HN: Where are we in the hype cycle on this?

We can run shitty clones slowly on Raspberry Pi's and your phone. The educational implementations demonstrate the basics in under a thousand lines of brisk C. Great. At some point you have to wonder... well, so what?

Not one killer app has emerged. I for one am eager to be all hip and open minded and pretend like I use LLMs all the time for everything and they are "the future" but novelty aside it seems like so far we have a demented clippy and some sophomoric arguments about alignment and wrong think.

It did generate a whole lot of breathless click-bait-y articles and gave people something to blab about. Ironically it also accelerated the value of that sort of gab and clicks towards zero.

As I am not a VC, politician, or opportunist, hand waving and telling me this is Frankenstein's monster about to come alive and therefore I need billions of dollars or "regulations" just makes folks sound like the crypto scammers.

Please HN, say something actually insightful, I beg you.

Re: What we know about LLMs

#14

Post by usaar333 »

killernap wrote: post #13

ChatGPT was announced November, 2022 - 8 months ago. Time flies. Question for HN: Where are we in the hype cycle on this? We can run shitty clones slowly on Raspberry Pi's and your phone. The educational implementations demonstrate the basics in under a thousand lines of brisk C. Great. At some point you have to wonder... well, so what? Not one killer app has emerged. I for one am eager to be all hip and open minded…

> Not one killer app has emerged

ChatGPT itself is a killer app.

Re: What we know about LLMs

#15

Post by killernap »

usaar333 wrote: post #14
killernap wrote: post #13

ChatGPT was announced November, 2022 - 8 months ago. Time flies. Question for HN: Where are we in the hype cycle on this? We can run shitty clones slowly on Raspberry Pi's and your phone. The educational implementations demonstrate the basics in under a thousand lines of brisk C. Great. At some point you have to wonder... well, so what? Not one killer app has emerged. I for one am eager to be all hip and open minded…

> Not one killer app has emerged ChatGPT itself is a killer app.

About as killer as that twitter clone that was in the news for a minute after forcing people to use it and immediately losing 90% of the captive audience..

They have been losing users. Summer is here, school is out, the kids are back in reality for the moment and apparently when they aren't busy plagiarizing homework the interest is very limited.

Re: What we know about LLMs

#16

Post by reader5000 »

killernap wrote: post #13

ChatGPT was announced November, 2022 - 8 months ago. Time flies. Question for HN: Where are we in the hype cycle on this? We can run shitty clones slowly on Raspberry Pi's and your phone. The educational implementations demonstrate the basics in under a thousand lines of brisk C. Great. At some point you have to wonder... well, so what? Not one killer app has emerged. I for one am eager to be all hip and open minded…

I think 8 months is a little short for the utility of a new tech to be fully realized and utilized. I'm pretty sure there were still horses on the roads long after 8 months after the Model T first went on sale.

Re: What we know about LLMs

#17

Post by viraptor »

FishInTheWater wrote: post #10

Given a set of instructions, an instruction fine-tuned/aligned LLM is able (conditional on size and training quality) to reason through a set of steps to produce a desired output. This is plainly wrong. The model's growing size makes it better at guessing the outcome of a reasoning task, but little to no actual reasoning is performed. It's trivial to prove this as well, as LLMs will still fail miserably at (larger) m…

> The model's growing size makes it better at guessing the outcome of a reasoning task, but little to no actual reasoning is performed.

If there's no observable difference between the behaviours, why not call it as the post did?

> LLMs will still fail miserably at (larger) math problems

They're neither trained on such problems, nor is that a goal for LLMs. They can however tell you how to convert that problem into steps that can be run in an algebra system.

Re: What we know about LLMs

#18

Post by throwuwu »

jannyfer wrote: post #7

Great summary. I’ve been reading a pop neuroscience book called Incognito (2011). In it, the author talks about how the brain is a group of competing sub-brains of many forms, and the brain might have several ways of doing the same thing (e.g. recognizing an object). The author also posited that the lack of AI progress back then was due to the fact that there are no constantly competing sub-brains. Our brains are alw…

How are they similar?

Re: What we know about LLMs

#19

Post by gjm11 »

getmeinrn wrote: post #6

How do you do science on LLMs? I would imagine that is super important, given their broad impact on the social fabric. But they're non-deterministic, very expensive to train, and subjective. I understand we have some benchmarks for roughly understanding a model's competence. But is there any work in the area of understanding, through repeatable experiments, why LLMs behave how they do? Do we care?

I would imagine it's a bit like doing science on human beings, who are also non-deterministic, expensive to train, and subjective. Perhaps there's scope for a scientific discipline corresponding to psychology but concerned with AI systems. We could call it robopsychology.

Re: What we know about LLMs

#20

Post by samwillis »

go_elmo wrote: post #11
jannyfer wrote: post #7

Great summary. I’ve been reading a pop neuroscience book called Incognito (2011). In it, the author talks about how the brain is a group of competing sub-brains of many forms, and the brain might have several ways of doing the same thing (e.g. recognizing an object). The author also posited that the lack of AI progress back then was due to the fact that there are no constantly competing sub-brains. Our brains are alw…

In the end - an AI should have these competing subsystems in one system - just as our brains are one system. What I find extremely interesting is how perception and thinking differs from person to person too - it was a "taboo" topic to call this neurodiversity - just as other genetic traits, but AI makes this relevant more than ever imo. Sure, its complicated and much comes from nurture (Nurture vs nature.. as exposu…

You touch in an important topic here, how our understanding of AI/ML/LLMs will influence our "understanding" of the human brain and intelligence.

My fear is that we will ascribe too much human behaviour to that we see in and understand of our AI inventions, and that this will result in the dehumanisation of people.

So essentially my fear is what we justify doing to each other due to AI, rather than what "AGI" could do to us.

Post reply on HN