Hacker News new | ask | show | jobs
by mns 711 days ago
I keep seeing these posts on HN and thinking, man, these are some smart people. Training LLMs, doing all this amazing AI stuff like this guy with the email agents and the other guy with the dropping of hats, and then I open the posts and it's just some guy making API requests to OpenAI or some similar disappointment.
5 comments

When “altcoins” took off I spent a while racking my brain trying to figure out what special tech I could offer, how I could build my own blockchain, incentivize miners…

When I realized it was just dudes copy-pasting a “smart contract” and then doing super shady marketing, it was already illegal in my jurisdiction.

Nowadays, an "AI Expert" is someone who knows how to download an AI client lib and prompt the AI to perform tasks. These are people who are not even technical and have no idea how all this works, but they can at least follow a Youtube Tutorial to get a basic website working.
As someone who actually has a university degree in Artificial Intelligence, I feel like this is always how it's been. Before, an "AI Expert" was someone who knew how to use Tensorflow, PyTorch or Keras. Before that, an "AI Expert" was someone who knew how to write a Monte Carlo simulation, etc etc.

You could of course say the same for frontend engineer or backend engineers. How many frontend engineers are simply importing Tailwind, React, etc? How many backend engineers are simply importing apache packages?

Where do you draw the line? Can you only be an AI expert if you stick to non-LLM solutions? Or are AI experts the people who have access to hundreds of millions of USD to train their own LLMs? Who are the real AI experts?

I would liken it to cars. There is a difference between engineers, mechanics, and mechanics that know a certain car so well that they fabricate parts that improve upon the original design.
Good comparison. Engineers who build cars and understand their intricacies oftentimes just work on one small thing at a time, even in teams. Like a team just working on breaks. The mechanics can piece the stuff together and keep it working in a real world setting. But nowadays a self-declared "AI Expert" in that metaphor might be just some person who knows how to drive a car.
If you think back to when cars were introduced, knowledge of how to drive a car was actually a rare skill! People weren't born with that inherent knowledge, so someone who could operate a vehicle (and do some basic maintenance) was an expert.

Nowadays, that would be laughed at. But AI is more comparable to cars from 1900 than modern vehicles.

I used to work on breaks, but then I realized I was more productive when I actually stopped and walked around a bit.
i draw the line at people claiming to be experts in something they have only done for a year
Also in most cases they were a "crypto expert" just two months ago.
And they were a leadgen/SEO expert a few years ago. These technogrifters just move from one hot topic to the next trying to make whatever buck they can smooth talk people into giving them.
Someone who can get a website working is actually technical.
Business as usual iow. Used to be scrum masters, then javascript "experts", then crypto bros.

Snake oil salesmen we called em back in my day ;-)

well, no one's going to be talking about the secrets behind LLM while the market is paying billions to own their slice of the pie.

And in reality, most software work is 1) API calls and 2) applied math. If you're not in cutting edge private tech or acedemia, your work probably falls into 1 or both categories. Modern "Software engineers" is more a matter of what scale of APIs you're wrangling, not how deep of domain knowledge you have.

You thought wrong, that's all. Those things aren't remotely hard. They're just simple things people don't bother doing.
It’s more about being on the front of the hype train and being endlessly positive versus competence.
I can't see this working long term though. Being endlessly positive and ignoring your actual competence sounds like a recipe to eventually bite off more than you can chew.
Oftentimes this is fervor is channeled into personal brand building, which rarely has any sort of feedback mechanism that is tied to actual competence.

It's a calculated move on their part.

Brand building actually sounds good and productive to me, as long as it doesn’t approach fraud.

If your audience likes your brand and doesn’t distinguish between your services and services done by more competent providers, then you’ve found your niche. So: snake oil is not fine; but Supreme branded brick sounds ok to me, even if I wouldn’t buy it myself.

I guess the author will find followers who enjoy that approach to software and product growth. If spamming wasn’t part of it, I’d be ok.