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

Re: The L in "LLM" Stands for Lying

#381

Post by wepple » Thu, Mar 05, 2026, 5:52 PM UTC

LetsGetTechnicl wrote: post #265

wepple wrote: post #196

Earlier quoted context omitted.

> It is there to reduce our agency, to make it easier to fire us, to put us in even more precarious position Could be. It could also end up freeing us from every commercial dependency we have. Write your own OS, your own mail app, design your own machinery to farm with. It’s here, so I don’t know where you’re going with “I’m unhappy this is happening and someone should do something”

It could also end up freeing us from every commercial dependency we have. Write your own OS, your own mail app, design your own machinery to farm with. Lmfao LLM's can barely count rows in a spreadsheet accurately, this is just batshit crazy. edit: also the solution here isn't that every one writes their own software (based on open source code available on the internet no doubt) we just use that open source software,…

Batshit crazy?

3 years ago LLMs couldn’t solve 7x8.

Now they’re building complex applications in one shot, solving previously unsolved math and science problems.

Heck, one company built a (prototype but functional) web browser

And you say it’s crazy that in the future it’ll be able to build a mail app or OS?

Re: The L in "LLM" Stands for Lying

#382

Post by moffkalast » Thu, Mar 05, 2026, 5:54 PM UTC

simianwords wrote: post #29

What the author and many others find hard to digest is that LLMs are surfacing the reality that most of our work is a small bit of novelty against boiler plate redundant code. Most of what we do is programming is some small novel idea at high level and repeatable boilerplate at low level. A fair question is: why hasn’t the boilerplate been automated as libraries or other abstractions? LLMs are especially good at fuzz…

It has been automated as much as possible, boilerplate is the result of people being terrible at designing programming languages. The whole idea of increasingly higher level ones was just that all along. Is there any point in writing a billion ADDC commands in assembly by hand when it takes one line in python?

LLM type systems are the final level of abstraction that lifts it up to literal natural language. Any dev with decent self awareness would admit they were just copying shit from stackoverflow half the time before LLMs anyway, high level languages and libraries just streamline that process with canonical implementations.

The value we provide is turning "person with problem" -> "person with solution to said problem" with as few caveats as possible. A programmer is that arrow, we solve problems. The more code we have to write to solve that problem, the worse we are at our job.

Re: The L in "LLM" Stands for Lying

#383

Post by noemit » Thu, Mar 05, 2026, 5:55 PM UTC

kevstev wrote: post #356

salad-tycoon wrote: post #181

Earlier quoted context omitted.

Agreed, I think the good gained by wool mills is greater in that little Timmy is less likely to lose a leg to frostbite than the bad loss of my scarf not passing through a ring. Long term though, I’ve always wondered if the Amish turn out to be the only survivors.

I am kind of lost here on this whole scarf through a ring thing as well. This is just a function of the thickness of the scarf? My wife went through a scarf phase about a decade ago, and I am pretty sure a Pucci scarf could easily fit through a typical sized ring meant to go on a finger? Its entirely possible that old manufacturing methods produced things that are different, but I would be entirely surprised if they…

A full size wool scarf cannot go thorough a ring. You are probably thinking of a silk scarf. I have a wool scarf next to me from Kashmir and it went down about 25 cm. The full scarf is a bit over a meter.

Looked up Pucci - looks like a designer that makes silk scarves. Silk is a totally different material. The Luddites were wool and cotton weavers.

Making wool thin enough for a meter long scarf to go thorough a ring requires the individual strand of wool to be very thin. Both making it thin and weaving that thin strand is the craft that was lost. Go look at wool yarn next time you are at a store and see how thin they can get it.

As for "Are they better?" Yes. Thinner wool is incredible, soft. High quality merino wool is one of the most expensive fabrics. Look up this brand "Made in Rosia Montana" if you are curious. It's not like what the Luddites made, but its as good as it gets in the modern world. Getting stuff from the Kashmir region is difficult - I got mine because I knew someone who ran a school in the area. Most "Cashmere" stuff in department stores is fake/chemically processed for fake softness which makes it nice but it doesn't last. Real quality wool lasts a lifetime. The chemically processed stuff is ok if you want to see how it "feels"

EDIT: also, wool is naturally waterproof! I can walk in the rain with my scarf from kashmir on my head, its pretty thin but absolutely no water goes through even in heavy rain. it has to do with the springiness of the fibers and its natural oils. I will stop nerding out on fibers now!!

Re: The L in "LLM" Stands for Lying

#384

Post by FpUser » Thu, Mar 05, 2026, 6:01 PM UTC

raincole wrote: post #310

FpUser wrote: post #230

Earlier quoted context omitted.

>"well perhaps except the function vs object one" If this is what I think it is, I consider it very lopsided view, failure to recognize what model fits for what case and looking at everything from a hammer point of view

I think function is the fundamental unit and object is an extra level over it (it doesn't mean there is no use for object). Thinking objects/classes are the fundamental/minimal level is straight up wrong. Of course it's just my opinion.

My opinion: Fundamental levels are data and operations (your functions). Not my view that class is a foundation. It is a representation convenient for some cases and not so much for other

Re: The L in "LLM" Stands for Lying

#385

Post by bluefirebrand » Thu, Mar 05, 2026, 6:02 PM UTC

eucyclos wrote: post #42

simianwords wrote: post #29

What the author and many others find hard to digest is that LLMs are surfacing the reality that most of our work is a small bit of novelty against boiler plate redundant code. Most of what we do is programming is some small novel idea at high level and repeatable boilerplate at low level. A fair question is: why hasn’t the boilerplate been automated as libraries or other abstractions? LLMs are especially good at fuzz…

I wrote a book a while back where I argued that coding involves choosing what to work on, writing it, and then debugging it, and that we tend to master these steps in reverse chronological order. It's weird to look at something that recent and think how dated it reads today. I also wrote about the Turing test as some major milestone of AI development, when in fact the general response to programs passing the Turing t…

I would argue that chatbots still barely pass the turing test

They have such obvious patterns and tells that humans have already picked up on them and they can eventually sus out that they're talking to an LLM

For instance I heard recently about someone talking (verbally) with an AI voiced customer support. They were very convinced, so they asked the support agent to calculate the product of two large numbers, and it replied with the result instantly

I would argue that fails the chinese room

Re: The L in "LLM" Stands for Lying

#386

Post by spacecadet » Thu, Mar 05, 2026, 6:03 PM UTC

pixl97 wrote: post #277

spacecadet wrote: post #159

Earlier quoted context omitted.

Demand full automation. Demand universal basic income. Notice how the later is nearly absent from the conversation. Another distraction is AGI that which is a danger to humanity- the only danger is people...

> the only danger is people... Simply put, no it is not. But on the reverse, the first danger with AI is people. Over the longer term it will look like this. The rich 'win' the world by using AI to enslave the rest of mankind and claim ownership over everything. This will suck and a lot of us will die. The problem is this doesn't solve the greed that cause the problem in the first place. The world will still be limit…

This is entirely assumptions about a future that has not happened.

Ive worked in "AI" for 20 years, through 2 winters, and run an alignment shop and AIRT... The problem is people. People will use the problem as a scapegoat.

Re: The L in "LLM" Stands for Lying

#387

Post by shadowgovt » Thu, Mar 05, 2026, 6:11 PM UTC

WarmWash wrote: post #345

dec0dedab0de wrote: post #271

Earlier quoted context omitted.

No, it's simply untrue. Players only object against AI art assets. And only when they're painfully obvious. No one cares about how the code is written. This reminded me of a conversation about AI I had with an artist last year. She was furious and cursing and saying how awful it is for stealing from artists, but then admitted she uses it for writing descriptions and marketing posts to sell her art.

Everyone is in it for themselves. The world makes waay more sense when you really internalize that. It doesn't necessarily mean people are selfish, large groups often have aligned interests, but when an individuals interest alignment changes, then their group membership almost always changes too. I'd bet she has a bunch of pirated content and anti-copyright remarks from the golden age of piracy as well.

If she's a practicing artist, she almost certainly cut her teeth doing tracing at some point. And if a digital artist, she almost certainly used a cracked copy of a tool.

The big eye-opener for me in college was taking a class that put me up-close with artists and learning that there were, in the whole class, a grand total of two students who hadn't started doing 3D modeling on a cracked copy of Maya (and the two, if memory serves, learned on Blender).

Re: The L in "LLM" Stands for Lying

#388

Post by htek » Thu, Mar 05, 2026, 6:12 PM UTC

banannaise wrote: post #354

tanjtanjtanj wrote: post #232

Earlier quoted context omitted.

While there are many Roguelikes with procedural generation, I think the most popular ones do not. Slay the Spire, Risk of Rain 2, Hades 1/2, BoE etc are all handmade stages with a random order with randomized player powers rather than procedurally generated.

I've seen a couple roguelike developers report that they played around with procedural generation, but it was difficult to prevent it from creating dungeons that were bad, unfun, or just straight-up killscreens. Turns out it's often easier to simply hand-draw good maps than to get the machine to generate okay-to-good ones. Procedural generation is good when variety matters more than quality, which is a relatively rar…

That says more about the developer than procedural generation as a whole. Using procedural generation IS difficult, it requires understanding how to set up constraints on your p-random generated elements and ensuring the code validates that you have a "good" level/puzzle/whatever before dumping the PC into it.

Re: The L in "LLM" Stands for Lying

#389

Post by alexpotato » Thu, Mar 05, 2026, 6:35 PM UTC

whywhywhywhy wrote: post #215

dannersy wrote: post #113

Earlier quoted context omitted.

You're cherry picking. The open world games aren't as compelling anymore since the novelty is wearing off. I can cherry pick, too. For example, Starfield in all its grandeur is pretty boring. And the users may not care about code directly, but they definitely do indirectly. The less optimized and more off-the-shelf solutions have seen a stark decrease in performance but allowing game development to be more approachab…

Startfield is boring because of the bad writing and they made a space exploration game where there are loading screens between the planet and space and you don’t actually explore space. They fundamentally misunderstood what they were promising, it’s the same as making a pirate game where you never steer the ship or drop anchor. You can prove people are not bored with the concept as new gamers still start playing fall…

This is why Sid Meier's Pirates [0] remains such a great game.

It was really a combination of mini-games:

- you got steer a ship (or fleet of ships) around the Caribbean

- ship to ship combat

- fencing

- dancing (with the Governors' daughters)

- trading (from port to port or with captured goods0

- side quests

Each time I played it with my oldest, it felt like a brand new game.

https://en.wikipedia.org/wiki/Sid_Meier%27s_Pirates!

Re: The L in "LLM" Stands for Lying

#390

Post by dannersy » Thu, Mar 05, 2026, 6:36 PM UTC

mexicocitinluez wrote: post #235

dannersy wrote: post #113

Earlier quoted context omitted.

You're cherry picking. The open world games aren't as compelling anymore since the novelty is wearing off. I can cherry pick, too. For example, Starfield in all its grandeur is pretty boring. And the users may not care about code directly, but they definitely do indirectly. The less optimized and more off-the-shelf solutions have seen a stark decrease in performance but allowing game development to be more approachab…

> Starfield in all its grandeur is pretty boring. And yet "No Mans Sky" is massively popular. > ny software engineer worth their salt knows intimately that more immediate results is usually at the expense of long term sustainability. And any software engineer worth their salt realizes there are 100s if not 1000s of problems to be solved and trying to paint a broad picture of development is naive. You have only seen 1…

No Man's Sky got better as they were more intentional with their content. The game has more substance and a lot of that had to be added by hand. It is dropped in procedurally but they had to touch it up, manually, to make it interesting. Let's not revise history.

I don't think it has anything to do with ego. There are studies on the topic of AI and productivity and I assume we have a way to go before we can say anything concretely. Software workflows permeate the industry you're in. You're putting words in my mouth, I said nothing about what people are doing is wrong or not useful. I said the claim that generative AI is making engineers more productive is an unfounded one. What code you shit out isn't where the work starts or ends. Using expedient solutions and having to face potentially more work in the future isn't even something that is a claim about software, I can make that claim about life.

You need to evaluate what you read rather than putting your own twist on what I've said.