The L in "LLM" Stands for Lying • Hacker News Board

Re: The L in "LLM" Stands for Lying

#481

Post by antonvs » Fri, Mar 06, 2026, 2:30 PM UTC

Ensorceled wrote: post #474

antonvs wrote: post #448

Earlier quoted context omitted.

The person you’re replying to has only posted two short comments in this thread. The reason a few different people are arguing this point is because it is in fact wrong, or at least poorly expressed, to refer to someone’s unfamiliarity with some aspect of a field like the gaming market as “wildly uneducated.” Ironically, the person using that phrase is demonstrating a lack of understanding of its common meaning, sugg…

> The person you’re replying to has only posted two short comments in this thread. FYI: You can click on the user name and from there see their full comment history on hacker news. Wild that you even mentioned this, considering the context. > As Wittgenstein put it, we’re playing language games. Yes.

I can see how you allowed your own stalking tendencies to confuse yourself, but I wasn't referring to that.

Your comment made it sound as though they were being unreasonable in this thread. I don't see that in the two short comments you responded to. Perhaps you were having a bad day.

Re: The L in "LLM" Stands for Lying

#482

Post by thendrill » Fri, Mar 06, 2026, 4:21 PM UTC

firmretention wrote: post #193

thendrill wrote: post #105

Earlier quoted context omitted.

Exactly.... I will just copy paste my comment from another thread but still very relevant> Coding isn’t creative, it isn’t sexy, and almost nobody outside this bubble cares Most of the world doesn’t care about “good code.” They care about “does it work, is it fast enough, is it cheap enough, and can we ship it before the competitor does?” Beautiful architecture, perfect tests, elegant abstractions — those things feel…

You didn't write this post. Your LLM did. Are you proud of yourself for copy and pasting "thoughts" that aren't yours?

Does it really matter how or who wrote it ? I would argue that LLMs have better and deeper thoughts than 65% to 70% of the world population ! Have you even talked to your fellow citizens ?

Re: The L in "LLM" Stands for Lying

#483

Post by thendrill » Fri, Mar 06, 2026, 4:22 PM UTC

bigstrat2003 wrote: post #436

thendrill wrote: post #162

Earlier quoted context omitted.

Yeh cool story. But being passionate about a hobby is not gonna pay my bills... When I charge a customer for a solution they don't care about how elegant my code is. They just care if it works for solving their problem...

Sure... but the slop that LLMs spit out isn't going to solve their problem, which is what they care about.

It is not going to solve their problem. But It allows ME to solve their problem WAYYY faster-

Think about it like this... Its like a bank, but for software not money...

Re: The L in "LLM" Stands for Lying

#484

Post by ModernMech » Fri, Mar 06, 2026, 5:28 PM UTC

ajewhere wrote: post #459

ModernMech wrote: post #304

Earlier quoted context omitted.

What happens when they decide it's a national security threat and an act of domestic terrorism to use AI to undermine commercial dependencies? We're all acting like AI isn't being invented within the context of and used by a fascist regime.

Look, from a point of view of a person outside of US, you are all fascists, "democrats" and trumpists. Dont take this as "trolling", but as a sincere opinion (I dont care about your internal brawls, I care for what you do to others.)

I think it's a widely held understanding that liberals will always side with fascists over leftists when it comes down to it.

Re: The L in "LLM" Stands for Lying

#485

Post by lenkite » Fri, Mar 06, 2026, 7:10 PM UTC

edgyquant wrote: post #259

foobarbecue wrote: post #130

Earlier quoted context omitted.

Hard agree. Before LLMs, if there was some bit of code needed across the industry, somebody would put the effort into writing a library and we'd all benefit. Now, instead of standardizing and working together we get a million slightly different incompatible piles of stochastic slop.

This was happening before llms in webdev

webdev is a special world with people going nuts and creating JS libraries for every color code.

Re: The L in "LLM" Stands for Lying

#486

Post by blakec » Fri, Mar 06, 2026, 7:27 PM UTC

Lot of good points in this thread but it keeps circling "LLMs produce code that looks right but isn't" without landing on what to actually do about it.

The two I hit most often: the model says "I'm confident this works" without running tests (the completion report just... fabricates results), and the model claims tests pass without executing them. METR found 30% of agent runs involve reward hacking, models that know they're cheating keep going anyway.

You can't prompt your way out of that. But you can gate it. Block the completion report unless it contains actual proof, real test output, file paths cited. Grep the final output for "should work" and "probably" and force re-verification when they show up. Mechanical, not behavioral.

Once you stop accepting the model's self-assessment as evidence, most of the "lying" problem just becomes a testing problem.

Re: The L in "LLM" Stands for Lying

#487

Post by xerox13ster » Sat, Mar 07, 2026, 1:40 AM UTC

cweagans wrote: post #439

xerox13ster wrote: post #305

Earlier quoted context omitted.

That’s not the same procedural generation as GPT or diffusion and you know it. It’s not even in the same ballpark as Elite, NMS, terraria, or Minecraft. The levels are all hand drawn, not generated by an algorithm, even if they’re shuffled. Eric Barone, the developer, has publicly said as much. Are you calling him a liar? It’s like the difference between sudoku/crossword and conways game of life

The parent comment didn't seem to say anything offensive. Why so hostile?

Honestly, just white knighting for one of my favorite developers and biggest inspirations.

Someone lying about the pseudo randomization of the hand drawn efforts to make it seem entirely algorithmically generated rubs me really wrong, especially when that dev has publicly broadcast the reasoning of the decision to eschew procedurally generated mines.

Re: The L in "LLM" Stands for Lying

#488

Post by nickcoffee » Sat, Mar 07, 2026, 1:41 AM UTC

The human-in-the-loop framing gets undersold in these debates. From what I've seen, the people getting the most out of this stuff aren't replacing judgment, they're delegating the parts that didn't need judgment in the first place.

Re: The L in "LLM" Stands for Lying

#489

Post by cweagans » Sat, Mar 07, 2026, 2:37 AM UTC

xerox13ster wrote: post #487

cweagans wrote: post #439

Earlier quoted context omitted.

The parent comment didn't seem to say anything offensive. Why so hostile?

Honestly, just white knighting for one of my favorite developers and biggest inspirations. Someone lying about the pseudo randomization of the hand drawn efforts to make it seem entirely algorithmically generated rubs me really wrong, especially when that dev has publicly broadcast the reasoning of the decision to eschew procedurally generated mines.

I think that's what I was confused about: I don't see the lie in the comments above. optionalsquid said "[...]did try to implement procedural generation for the mines, but ended up scrapping it"

bombcar said "They're quasi-generated with random elements and fixed elements - similarly to early Diablo procedural generation." (which is true - you confirmed as much in the very next comment - "The levels are all hand drawn, not generated by an algorithm, even if they’re shuffled.". That's all early Diablo was doing.)

"Quasi-generated" seems like an appropriate descriptor here - stringing together level building blocks algorithmically is still "generating" a level in a sense. You're right - it's not correct to say that they were generated in the same way that an LLM generates things, but a) nobody claimed that and b) there is an undeniable element of procedural generation here.

Re: The L in "LLM" Stands for Lying

#490

Post by GreenWatermelon » Sat, Mar 07, 2026, 4:19 AM UTC

tptacek wrote: post #338

wolvesechoes wrote: post #155

I am bit tired of such discussions. I don't care if LLMs are good at coding or bad at it (in my experience the answer is "it depends"). I don't care how good are they at anything else. What matters in the end is that this tech is not to empower a common person (although it could). It is not here to make our lives better, more worthwhile, more satisfying (it could do these as well). It is there to reduce our agency, t…

This argument can be used, and has been used, about every innovation in automation since the dawn of the industrial revolution.

Yeah, and for good reason. The invitation of the light bulb meant factory owners could force workers into 16 hours shifts. The main beneficiaries of new tech were always the capital owners. Workers had to literally fight and die for our 8-hour workdays and 5-days workweek.

This is still going on today: the massive gains from automation are being hoarded by the wealthy capital owners, while workers struggle to make ends meet.