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

What it feels like to work in AI right now

robotic.substack.com

Post reply on HN
Print view View original

81–90 of 768 posts

Re: What it feels like to work in AI right now

#81

Post by pixl97 »

tayo42 wrote: post #70
karmakurtisaani wrote: post #55

Earlier quoted context omitted.

A PhD student I recently talked to was complaining how in academia they will never have the resources it takes to train models like GPT. And how now some academics are only researching the input/outputs to these LLMs. Seems like it's pretty dire or at least uninspiring to all but very few at the moment.

why is that? don't like those academic particle researchers get tons of money to build wild experiments to measure waves or collide particles? unless i get the funding part wrong? are those not academic projects?

The issue here is speed.

If it takes 10 years to get funding on particle physics, generally it doesn't matter "that much" because that part of the science scene moves pretty slowly.

If it takes 10 years to build a billion dollar cluster for AI today, what the hell is the AI world going to look like in 10 years anyway? Building a cluster to study AI ethics might be meaningless because a terminator may be pointing its laser rifle at our head by then telling us to be good little human subjects because Microsoft and OpenAI decided to move fast and break things.

Re: What it feels like to work in AI right now

#82

Post by nickthegreek »

lm28469 wrote: post #71
voz_ wrote: post #67

Earlier quoted context omitted.

This question is in poor faith.

Not really, I've been digging for a while and I didn't find any ground breaking projects It's all nice and cool but most of the projects are toys and/or junior level generated boilerplate It's good for summarizing, rephrasibing and other neat features but that's about it, a far cry from the oracle some people seem to worship

Oracle to worship? There are no serious people doing any such thing. I don't even know of clowns doing that. Is this twitter discourse, a strawman or more poor faith?

Re: What it feels like to work in AI right now

#83

Post by pixl97 »

the_lonely_road wrote: post #43
hvs wrote: post #26

Earlier quoted context omitted.

It has helped me learn things more interactively because I can ask questions and get answers without having to read 10 blog posts or a book before understanding things. For my personality, it would be hard to recreate without having a living mentor available 24/7 that had knowledge of everything I wanted to ask of it. Does it get things wrong? Sure. But generally the subtle ways it might get stuff wrong would be the…

You must have missed the article posted this morning where ChatGPT complexity invented a sexual allegation against a law professor including citing a fictional news article that never existed. I wouldn’t call that a subtle misunderstanding.

So it's the Weekly World News... And yea, we've had rags before AI was made too.

Re: What it feels like to work in AI right now

#84

Post by pixl97 »

__loam wrote: post #38
hvs wrote: post #26

Earlier quoted context omitted.

It has helped me learn things more interactively because I can ask questions and get answers without having to read 10 blog posts or a book before understanding things. For my personality, it would be hard to recreate without having a living mentor available 24/7 that had knowledge of everything I wanted to ask of it. Does it get things wrong? Sure. But generally the subtle ways it might get stuff wrong would be the…

It's shocking to me how many people are treating chatGPT as a reliable source given how much we know it hallucinates.

Because very few people ever stop and take account of how monstrously unreliable people are.

At the same time I can ask GPT to write a script, reflect on its output, and it takes about 95% of the work I need to do out of script writing.

Re: What it feels like to work in AI right now

#85

Post by pixl97 »

vinyl7 wrote: post #15
bakugo wrote: post #8

Earlier quoted context omitted.

Pretty much, you can take any article hyping up AI/ChatGPT/etc and replace those words with crypto stuff and it's virtually indistinguishable from the crypto spam of yesteryear. I wonder what the next "get rich quick" techbro fad will be.

Replace crypto with cloud replace cloud with big data replace big data with web scale It's all just a hype and crash cycle

The cloud crashed? I'm sure AWS would like to know about that.

Re: What it feels like to work in AI right now

#86

Post by _hao »

I won't try to predict the future as far as such ML models will take our jobs or not. However, I'd like to point out to most people that working on and studying fundamentals won't be a time wasted. Even if programming disappears as an employable job, people will still need to understand and learn things. You'll still need to know the math/physics/philosophy/history/literature/art/music etc. It might turn out that it's going to be more important than ever for us to be more confident and strict in our own skills and knowledge. And I'd also suggest to broaden our horizons. We should become good at a couple of different things like the polymaths of old.

IF you have to spend hours debugging ML generated code for something critical you better know exactly what you're doing. If you have to generate a Harry Potter Balenciaga meme video (the whole Balenciaga meme and all the different versions are hilarious btw) you better have some visual/art and musical sense in order to make it funny and engaging.

I think the end of the world is greatly exagarated. We actually have much bigger problems than "AI".

Re: What it feels like to work in AI right now

#87

Post by medvezhenok »

lm28469 wrote: post #71
voz_ wrote: post #67

Earlier quoted context omitted.

This question is in poor faith.

Not really, I've been digging for a while and I didn't find any ground breaking projects It's all nice and cool but most of the projects are toys and/or junior level generated boilerplate It's good for summarizing, rephrasibing and other neat features but that's about it, a far cry from the oracle some people seem to worship

ChatGPT is quite good at organizing text into taxonomies, error correction, anomaly detection, data cleaning, etc.

I bet there are a lot of companies out there where that's a non-trivial percentage of their workflow (I know that in my case, we have reconsidered several internal projects in light of ChatGPT doing a better job).

It already does translation on the level / better than Google Translate, without being specifically trained for it (as one example). And it can play chess without a specialized model.

Re: What it feels like to work in AI right now

#89

Post by paulryanrogers »

JumpCrisscross wrote: post #54
UncleOxidant wrote: post #25

Earlier quoted context omitted.

Why do you think OpenAPI is so far out in front? It's not like there's a lot of secret sauce here - most of this stuff (transformers,etc.) is all out there in papers. And places like Google & Meta must have a lot more computing resources to train on that OpenAI does thus they should be able to train faster. Do you think OpenAI has discovered something they haven't been open about?

> Why do you think OpenAPI is so far out in front? There is a network effect forming around its models. The strengths of its kit speak for themselves. (It also cannot be understated how making ChatGPT public, something its competitors were too feeble, incompetent and behind the curve to do, dealt OpenAI a massive first-mover advantage.) But as others note, other models are in the ballpark. Where OpenAI is different i…

Doubtful "feeble and incompetent" are fitting considering much of the research behind OpenAI was pioneered and published by those predecessors. More like "unwilling", and perhaps for good reason. Time will tell.

Re: What it feels like to work in AI right now

#90

Post by letitgo12345 »

UncleOxidant wrote: post #25
fdgsdfogijq wrote: post #12

I work on a research team in FAANG. What it really feels like is one company made everyone else obsolete. And we are going to work working on NLP models that underperform ChatGPT by a huge margin. Twiddling my thumbs and keeping quiet while no one wants to recognize the elephant in the room. Also, there is no "working in AI", a few thousand people are doing real AI at most. The rest of us are calling an API.

Why do you think OpenAPI is so far out in front? It's not like there's a lot of secret sauce here - most of this stuff (transformers,etc.) is all out there in papers. And places like Google & Meta must have a lot more computing resources to train on that OpenAI does thus they should be able to train faster. Do you think OpenAI has discovered something they haven't been open about?

They have been collecting human feedback data for 2 years + probably have a lot of data from Copilot + are training with large context models + have invested a ridiculous amount in curating pretraining data -- the kind of stuff that won't get you a ton of pubs (so you won't see Google researchers having focused on it a lot) but apparently turns out to be super important for a good LLM
Post reply on HN