But when your boss tells you that you have to keep doing it this way, then you don’t have much choice in the matter. You either keep asking AI for new code and hope it gets it right, or you have to actually delve into the code and spend your time correcting it.

The 1 million lines of code is just untenable, assuming they want code that actually works.

LLMs are - by the nature of how they work - only able to achieve 90-95% accuracy. That’s the theoretical best they can do, according to the people behind OpenAI. And worse, it will be presented as 100% accurate, even going so far as to make up sources wholecloth.

That’s an insane and completely unacceptable error rate for any system even pretending to be mission critical.

Can you imagine sending people to space with a system that has a 1 in 20 chance of just being completely unfit for service?

But when they don’t pass, then you have to dissect a bunch of AI pasta, right?

I dunno man, I tried coding a simply http listener with an LLM one time in python (a language I’m unfamiliar with). Just something to sit on a port, listen for a request, and run a script.

I ended up spending more time troubleshooting the maybe two dozen lines of code than I would have spent just looking up a tutorial online.

That was me, and my point was that we already had a suitable, regulated, and relatively safe system in place, ie taxis, so there’s no need for innovation that will needlessly endanger people.

Or, we managed to create a society where people are upset about new, unregulated, and, according to the article in question, potentially dangerous technology being implemented with little regard to the public good.

Until it’s multiple self driving cars getting stuck in that intersection

Yea, so what was wrong with Taxis anyway? At least we regulated those

Is it brown people? It’s brown people, isn’t it?

If I had an extra 300 tb I’d do it.

You could cut off your search around the time AI tracks started to appear. Not sure when that was, maybe 2023. You’d miss a lot of recent stuff, but you’d filter out a lot of spam too

That scope will narrow as time passes.

Now it’s brown poor people. Soon it will be trans people. Maybe next brown rich people, or Muslims, or Socialists. Y’all know the poem.

They’ve already “declared” Antifa a terrorist group and fentanyl as a WMD, that’s all the justification previous republican presidents have needed for starting wars and civil terror campaigns.

I heard a lawyer argue something like this once in court, regarding the the fourth and fifth amendments:

These laws are not meant to protect the innocent, they are meant to protect criminals. The founding fathers who penned it were traitors and seditionists who fought a war against their own country. They wrote these laws so that guilty people would be able to avoid punishment if proper procedures aren’t followed, and certain rights aren’t upheld.

I’m not sure how much I agree with that, but it was definitely an interesting take.

Some things are just super obvious that even if we can, we should not

I use it for coding advice sometimes, as an amateur hobbyist it’s really useful to point me in the right direction when facing problems I’m unfamiliar with. I often end up reinventing the proverbial wheel, just worse, but LLMs can help point out standards and best practices that I, as an outsider to the industry, am unaware of.

Movie recommendations is my biggest thing, personally.

And lots of other purposes. Just because a ton of people are misusing this tool and treating it like GAI doesn’t mean that it isn’t a useful tool. Even something as simple as proofreading a letter has massive utility for some people.

Lets use LLMs for things LLMs are useful for. It is not a panacea, and it is not appropriate for every use case