Researchers at Apple have come out with a new paper showing that large language models can’t reason — they’re just pattern-matching machines. [arXiv, PDF] This shouldn’t be news to anyone here. We …
This isn’t news. We’ve known this for many, many years. It’s one of the reasons why many companies didn’t bother using LLM’s in the first place, that paired with the sheer amount of hallucinations you’ll get that’ll often utterly destroy a company’s reputation (lol Google).
With that said, for commercial services that use LLM’s, it’s absolutely not true. The models won’t reason, but many will have separate expert agents or API endpoints that it will be told to use to disambiguate or better understand what is being asked, what context is needed, etc.
It’s kinda funny, because many AI bros rave about how LLM’s are getting super powerful, when in reality the real improvements we’re seeing is in smaller models that teach a LLM about things like Personas, where to seek expert opinion, what a user “might” mean if they misspell something or ask for something out of context, etc. The LLM’s themselves are only slightly getting better, but the thing that preceded them is propping them up to make them better
IMO, LLM’s are what they are, a good way to spit information out fast. They’re an orchestration mechanism at best. When you think about them this way, every improvement we see tends to make a lot of sense. The article is kinda true, but not in the way they want it to be.
Are they a serious researcher in ML with insights into some of the most interesting and complicated intersections of computer science and analytical mathematics, or a promptfondler that earns 3x the former’s salary for a nebulous AI startup that will never create anything of value to society? Read on to find out!
I’m a Software Engineer at Amazon, and have worked on compositional and large language models for the last four years. I’m not a researcher, but I do work in applied and research science. My work is mostly to facilitate the science work and create scalable ways to make experiments work.
Aside from that, I joined fully aware of how cutthroat it is, and naively assumed that because I had succeeded in other places I’d be fine here. The churn is 100% real.
I’m still grateful to have worked here. It sounds cool to say I work in AI, I have enough money saved to take 2+ years off if I wanted, and it’s been good to work on something that millions of people use - regardless of my own thoughts on if it’s useful/good or not.
Outside of big tech and quants, yes. In the UK especially, salaries vary wildly. Some amazing software engineers might be lucky to make £50k, while someone in big tech might make that in RSU’s alone in a given year, with a salary twice as big again.
With that said, I’d find a middle ground to be nice, decent pay for decent WLB.
The average salary for London is ~£60k, and substantially less outside (I’m not from London).
Salaries here are much lower than that of the US, and lower than some European countries. Despite all of this, it also means that if you save enough from a high-paying job you can supplement your income from your interest and pursue other things.
Tax is also very high, but the other side of this is that healthcare is free, and worker protections are higher than in the US.
IMO, moving to the UK or parts of Europe is something many people in tech should do once they’ve earned seven figures in tech in the US. Despite lower income, they could probably live off of their interest and retire somewhere with an arguably equal quality of life.
You may laugh, but there’s very real applications that are quite interesting, particularly in how AI can help someone with a severe speech impediment, or in understanding someone who might not speak/write in English as their primary language. Persona agents are quite interesting, and might actually be a useful application for LLM’s.
“sigh”
(Preface: I work in AI)
This isn’t news. We’ve known this for many, many years. It’s one of the reasons why many companies didn’t bother using LLM’s in the first place, that paired with the sheer amount of hallucinations you’ll get that’ll often utterly destroy a company’s reputation (lol Google).
With that said, for commercial services that use LLM’s, it’s absolutely not true. The models won’t reason, but many will have separate expert agents or API endpoints that it will be told to use to disambiguate or better understand what is being asked, what context is needed, etc.
It’s kinda funny, because many AI bros rave about how LLM’s are getting super powerful, when in reality the real improvements we’re seeing is in smaller models that teach a LLM about things like Personas, where to seek expert opinion, what a user “might” mean if they misspell something or ask for something out of context, etc. The LLM’s themselves are only slightly getting better, but the thing that preceded them is propping them up to make them better
IMO, LLM’s are what they are, a good way to spit information out fast. They’re an orchestration mechanism at best. When you think about them this way, every improvement we see tends to make a lot of sense. The article is kinda true, but not in the way they want it to be.
Are they a serious researcher in ML with insights into some of the most interesting and complicated intersections of computer science and analytical mathematics, or a promptfondler that earns 3x the former’s salary for a nebulous AI startup that will never create anything of value to society? Read on to find out!
do i have to
Welcome to the future! Suffering is mandatory!
as a professional abyss-starer, I’m going to talk to my union about this
I’m a Software Engineer at Amazon, and have worked on compositional and large language models for the last four years. I’m not a researcher, but I do work in applied and research science. My work is mostly to facilitate the science work and create scalable ways to make experiments work.
Is thereA specific reason you picked Amazon?
Money, mostly.
Aside from that, I joined fully aware of how cutthroat it is, and naively assumed that because I had succeeded in other places I’d be fine here. The churn is 100% real.
I’m still grateful to have worked here. It sounds cool to say I work in AI, I have enough money saved to take 2+ years off if I wanted, and it’s been good to work on something that millions of people use - regardless of my own thoughts on if it’s useful/good or not.
Did they pay more than competitors? How did it compare?
Outside of big tech and quants, yes. In the UK especially, salaries vary wildly. Some amazing software engineers might be lucky to make £50k, while someone in big tech might make that in RSU’s alone in a given year, with a salary twice as big again.
With that said, I’d find a middle ground to be nice, decent pay for decent WLB.
Wait 150k-200k is considered good in the UK? I’d have thought you’d be getting offers in the millions/year.
The average salary for London is ~£60k, and substantially less outside (I’m not from London).
Salaries here are much lower than that of the US, and lower than some European countries. Despite all of this, it also means that if you save enough from a high-paying job you can supplement your income from your interest and pursue other things.
Tax is also very high, but the other side of this is that healthcare is free, and worker protections are higher than in the US.
IMO, moving to the UK or parts of Europe is something many people in tech should do once they’ve earned seven figures in tech in the US. Despite lower income, they could probably live off of their interest and retire somewhere with an arguably equal quality of life.
this but with extra wasabi
*trying desperately not to say the thing* what if AI could automatically… round out… spelling
It hurts when they’re so close!
You may laugh, but there’s very real applications that are quite interesting, particularly in how AI can help someone with a severe speech impediment, or in understanding someone who might not speak/write in English as their primary language. Persona agents are quite interesting, and might actually be a useful application for LLM’s.
Preface: repent for your sins in sackcloth and ashes.
Buh bye now.
while true; do fortune; doneis a good way to spit information out fast.This is why people don’t like Lemmy.