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

The Rise of the AI Engineer

latent.space

Post reply on HN
Print view View original

21–30 of 159 posts

Re: The Rise of the AI Engineer

#21

Post by yacine_ »

cj wrote: post #14

I can see a world where "ML Engineer" (or similar) is someone that's hired to solve a known problem (whether it be with classifiers, LLMs, neural nets, etc), whereas a "AI Engineer" (or whatever the title) is hired to figure out how the hell to capitalize on the AI hype, without a specific problem to solve. IMO right now we're entering the "Peak of Inflated Expectations" in Gartner's hype cycle model. https://en.wiki…

The right way to leverage AI is to use it to do ye olde engineering, faster. Actually, most AI use cases are in fact that! Ye olde engineering - data pipelines, automation. Stuff we've been doing since before I was born.

Everything else - transformative tech, will hit the open domain fairly quickly, is my bet. Similar to databases.

Re: The Rise of the AI Engineer

#22

Post by richardw »

cj wrote: post #14

I can see a world where "ML Engineer" (or similar) is someone that's hired to solve a known problem (whether it be with classifiers, LLMs, neural nets, etc), whereas a "AI Engineer" (or whatever the title) is hired to figure out how the hell to capitalize on the AI hype, without a specific problem to solve. IMO right now we're entering the "Peak of Inflated Expectations" in Gartner's hype cycle model. https://en.wiki…

I’ve been thinking the leading edge is already in the trough of disillusionment! At least on HN etc.

We see lots of reports of limitations on HN. Those in the know don’t trust it nearly as much as the public and all the CEO’s of businesses drooling to install AI and have the money go up. Going to have to dig our way through those issues to get real value.

Re: The Rise of the AI Engineer

#23

Post by dr_dshiv »

cbm-vic-20 wrote: post #7

I'm a natural skeptic, and I believe we're still on the rising edge of the "AI" hype cycle. Five years ago, it was "blockchain", and everyone was trying to ram blockchain into everything, attracting lots of VC and media attention, etc. It seems that blockchain is beyond the honeymoon phase: I haven't seen an NFT or even a Bitcoin headline in HN for a while. So I'm trying to wrap my head around what an "AI Engineer" i…

ChatGPT is useful in a way that block chain never was. There may be inflated expectations, but we are already on “the plateau of productivity.” I don’t think LLMs are overhyped relative to the normal background levels of tech hype.

Re: The Rise of the AI Engineer

#24

Post by spmurrayzzz »

cbm-vic-20 wrote: post #7

I'm a natural skeptic, and I believe we're still on the rising edge of the "AI" hype cycle. Five years ago, it was "blockchain", and everyone was trying to ram blockchain into everything, attracting lots of VC and media attention, etc. It seems that blockchain is beyond the honeymoon phase: I haven't seen an NFT or even a Bitcoin headline in HN for a while. So I'm trying to wrap my head around what an "AI Engineer" i…

> So is the "engineering" part of this finessing the input and massaging the output?

I don't know if I'll ever use the phrase "AI Engineer" myself, but there's plenty of meaningful engineering work in that space that strays pretty far from just calling some provider's APIs. A few that come to mind just for LLMs:

- Custom fine-tuning of foundational models both in the classic sense and with more modern strategies like PEFT/QLoRA

- Data preprocessing pipelines to help automate fine-tuning, vectorization, etc

- Continuous integration suites to evaluate models on standard benchmarks as they change over time

- Vector db / semantic search engineering to help decorate context windows effectively

- Architecting ensemble models infrastructure to accommodate more complex task processing

I think many of those probably going into what folks are calling the "MLOps" bucket, but I think its a more broadly a combination of research, application engineering, and operations engineering.

Edit: for clarity, my position is that the line in the article between AI Engineer and ML Engineer need not be that bright. Just like software engineers today that write/operate their own devops tooling to deploy and manage the apps they build.

Re: The Rise of the AI Engineer

#25

Post by Jtsummers »

supportengineer wrote: post #11

There are a couple types of roles that sounds really interesting to me. One would be taking some proprietary data and training LLM in a format it could use. One company I know has a database of cars. They want to train their LLM with some inventory facts like "We have a Ford Mustang on the lot whose vin is ABC123 and it has the following features...". And then another role would be writing the prompts for API calls t…

> They want to train their LLM with some inventory facts like...

So are they actually intending to retrain their LLM every time the inventory changes? Because, otherwise, how is it going to "know" the current state of the inventory? This is useless after a single sale or a single new delivery without retraining. (And it's likely useless before that anyways.)

And if they already have a database of inventory data with all this then they could just generate a report the "old fashioned" way that's worked for decades.

Re: The Rise of the AI Engineer

#27

Post by sanderjd »

voz_ wrote: post #9
cbm-vic-20 wrote: post #7

I'm a natural skeptic, and I believe we're still on the rising edge of the "AI" hype cycle. Five years ago, it was "blockchain", and everyone was trying to ram blockchain into everything, attracting lots of VC and media attention, etc. It seems that blockchain is beyond the honeymoon phase: I haven't seen an NFT or even a Bitcoin headline in HN for a while. So I'm trying to wrap my head around what an "AI Engineer" i…

Imagine comparing blockchain and AI.

I would submit that we don't yet have enough evidence to say whether the comparison is apt or inapt. I personally think what's going on with these "generative" models seems like a bigger deal than blockchain, but it's fiendishly difficult to know what is or isn't hype while embedded within a hype cycle.

Re: The Rise of the AI Engineer

#28

Post by minimaxir »

cbm-vic-20 wrote: post #7

I'm a natural skeptic, and I believe we're still on the rising edge of the "AI" hype cycle. Five years ago, it was "blockchain", and everyone was trying to ram blockchain into everything, attracting lots of VC and media attention, etc. It seems that blockchain is beyond the honeymoon phase: I haven't seen an NFT or even a Bitcoin headline in HN for a while. So I'm trying to wrap my head around what an "AI Engineer" i…

> Where where these experts hiding five years ago?

You could use GPT-3 in 2020 but it was expensive and difficult to make it behave. Iterations of GPT-3 starting in 2021-2022 allowed it to obey commands (InstructGPT) and made it more feasible to "engineer" with it.

The true inflection point was due to free, accessible and good-enough-quality AI generation in the form of Midjourney and ChatGPT.

Re: The Rise of the AI Engineer

#29

Post by minimaxir »

voz_ wrote: post #9
cbm-vic-20 wrote: post #7

I'm a natural skeptic, and I believe we're still on the rising edge of the "AI" hype cycle. Five years ago, it was "blockchain", and everyone was trying to ram blockchain into everything, attracting lots of VC and media attention, etc. It seems that blockchain is beyond the honeymoon phase: I haven't seen an NFT or even a Bitcoin headline in HN for a while. So I'm trying to wrap my head around what an "AI Engineer" i…

Imagine comparing blockchain and AI.

Unfortunately, web3 bros pivoted to AI, which just adds even more noise to the space.

Re: The Rise of the AI Engineer

#30

Post by sanderjd »

preommr wrote: post #17

Oh God, this is going to be like DevOps all over again isn't it? Where there's people constantly gluing together things from different companies. I foresee layers and layers of abstraction dependent not on technology spec but on some company's poorly maintained docs and apis that you get by calling some service rep.

I have found data engineering to feel that way as well. Just throwing darts at a board of vendor solutions, trying to figure out how to cobble them together into something useful. I am worried this is going to be that as well.

I'd like to learn more about how to self-host systems that are large enough to be useful, rather than doing this cobbling together proprietary APIs thing.

Post reply on HN