[JSR_FDED]

It’s crazy. Literally the only valid thing you can say about LLM use is “in the hands of an experienced person LLMs can be a force multiplier”.

But no, we have to replace entire companies with it. All the problems of LLMs stem from inexperienced people using it (by inexperienced I mean not skilled in the domain in which the LLM is being applied).

[m463]

> from inexperienced people using it

I think of inexperienced people using power tools. Power tools theoretically make your life easier, but beginners will strip screws or snap bolts right and left. Or with ikea furniture the screw threads in the "wood" get pulled right out.

With screws or LLMs, craftsmen can speed up and do the finer details by hand.

[stogot]

Ive seen skilled specialists use it in their domain (badly)

[vkou]

That's not the problem.

There are two[1] problems. One is that a small group of people will own a critical part to all future economic activity. It's wealth consolidation at an unimaginable scale.

The other is that the reason LLMs produce so much fucking garbage isn't because their users suck at their jobs. These users were producing good to passable work for years before LLM slop started flooding the world.

It's because their jobs - their bosses - can't (or don't care to) tell the difference between good work and fucking garbage. If the computer said it's ok, ship it.

----

[1] These two problems are not an exhaustive list by any means.

[dotancohen]

  > One is that a small group of people will own a critical part to all future economic activity. It's wealth consolidation at an unimaginable scale.

You might want to read about the history of the term sabotage then. This actually is precedented.

  > The other is that the reason LLMs produce so much fucking garbage isn't because their users suck at their jobs ... bosses can't tell the difference between good work and fucking garbage.

The bosses usually can. However, there is no quick, objective way to do that. The boss would likely literally have to do the work over from scratch to know if the current result is adequate.

I'm familiar with a firm that does government required safety inspections for clients - big factories and the like. There is a huge push inside the company to use AI to write the reports. I don't want to tell you how many times the LLM fills in the report with the most common string "everything is fine" even though the specific future report checkbox has been left unchecked. Only a single engineer is fighting this within the company, and fears for her job because of it.

[Nevermark]

> You might want to read about the history of the term sabotage then. This actually is precedented.

Yes, history.

History reveals time and again that dislodging the powerful via lawlessness is not a sweet prelude to a better world. To be precise, even in the best case, at least a generation goes through hell before things stabilize.

I am not a big fan of "solutions" that dramatically introduce much bigger deeper problems.

Corruption is generally solved with more order. Not more disorder.

Cynicism borne out of frustration is not our friend.

[dotancohen]

Thank you for the blunt response. I was not advocating for sabotage - rather I was demonstrating that the current situation has precedent and giving GP a starting point from where to read. Thank you for pointing out that interpretation.

[Nevermark]

Well, I am relieved we are on the same page!

It is all so dicey.

[cyanydeez]

We can also say the world is worse off when its in the hands of the oligarchy unconstrained by regulation.

We can say that, it's very easy and history has proven this many times.