I think the current code I see being generated is generally “good enough”. I’m not comparing it to perfect: I’m comparing it to people.

If this were true, then open source projects would have much less of an issue with pull requests from sloperators.

Have you seen output in which satirical code is actually included?

I wouldn’t expect to see it. Satirical code requires more thought than an LLM is capable of putting into its writing - you need to understand what is expected of whoever you’re satirizing, and then you have to take that expectation and take it a step further into the absurd. Without having that context of something that is specifically being satirized, what you have instead is just incorrect code. And again, the LLM is incapable of valuing proper code over intentionally wrong code, so it’s going to poison the database to some extent.

And LLMs don’t drop big chunks of copy-pasted code from Stack Exchange like an intern would. They work one token at a time. (Which is why trying to get them to understand that quotations need to be all in one piece is a futile endeavor.)

Besides, ‘satirical code’ is just one example of the many things that can poison the training. I couldn’t even begin to enumerate all the things that could mess with it, and honestly I’m surprised that LLMs do as well as they do considering they likely have all sorts of cross-language screwball connections (which may be why it has such a tendency to make up libraries; it doesn’t necessarily understand that a common PHP library doesn’t exist in Java).

do you not believe that either (a) these types of trivial issues would be caught by a person whose job was just to audit output or even (b) this type of issue could be caught by specially trained domain limited AIs designed to check output?

These issues could be caught by someone whose job it is to audit code, sure. The problem is that sloperators often don’t audit their own stuff well enough. They leave it to the open source repo’s admins. When pull requests from overeager noobs were infrequent, it wasn’t the problem; they could gently correct them, the repo would stay high-quality, the noob would learn, everyone is fine. But now, sloperators are dumping low-quality pull requests on the repos faster than the admins can sort through them - because it now takes less time to produce slop code than it takes to determine whether or not the slop is worth including. The admins are swamped, because they can’t sort the wheat from the chaff fast enough.

A domain-limited AI designed to check output would be useful - if it could be trusted. Open-source project admins are some of the best coders out there, and they vastly outstrip the capabilities of LLMs. You’re suggesting that we replace THEM with an agent. They are in that position because they’re right far more often than they’re wrong when it comes to understanding the code as it exists, and how incoming code would impact it - or at least they’re right often enough to keep the project alive. LLMs will be worse at that job, I guarantee it. They’d be fast, but they’d be wrong too often. This is the primary issue with LLM agents.

You lose reach when you’re swordfighting if you have the high ground, and reach is a much more significant advantage. It’s also far easier to guard your head than it is to guard your ankles. Given the choice between the two, I’d take the low ground.

This changes significantly with mixed arms. If you have spears, you want the high ground. If you have bows, you REALLY want the high ground. If you’re duelling someone, sword to sword, you do NOT want the high ground. And this is the voice of personal experience.

When you’re duelling, what you actually want is the sun at your back, and solid ground behind you - and for your opponent to be on uneven ground, without good options to back up safely, and the sun in their eyes.

You say that, but you have to remember that LLMs produce the average output of their training materials. Not the best, but the average. And there’s a lot of code out there that is simple. Only the outliers have the magic combination of conciseness AND quality AND complexity.

LLMs also have no understanding of context outside the immediate. Satire is completely opaque to them. Sarcasm is lost on them, by and large. And they have no way to differentiate between good and bad output. Or good and bad input, for that matter. Joke pseudocode is just as valid in their training corpus as dire warnings about insecure code.

I read a comment once that still rings true - “Hallucinations” are a misnomer. Everything an LLM puts out is a hallucination; it’s just that a lot of the time, it happens to be accurate. Eliminating that last percentage of inaccurate hallucinations is going to be nearly impossible.

LLMs have made it so that it takes longer to determine whether content is bad than it takes to make bad content. The solution SHOULD be to demand that people examine the content themselves and turn it into high-quality content, but that’s not going to happen so long as it is possible for anyone to submit pull requests. The only solution that will actually work is to restrict who is allowed to submit content to your projects.

I’ve exhausted other avenues of exploration. I’ve long ago hit the point of diminishing returns on my important stats. Upgrading my primary loadout is prohibitively expensive at this point, and switching weapons isn’t helping my damage output.

I have the Claw, but it only lets me parry specific attacks - if I can remember to use it at the right time, and time it just right. Which I need to learn, which is difficult when phase 1 is such a slog.

Usually I can power through and rely on gradually gaining muscle memory against the bosses, but for whatever reason Malaketh is the exception. He’s my kryptonite, and I really, really, REALLY hate that first phase (and the fact that it doesn’t teach me anything about how to fight phase 2).

I wanted to beat him myself. I’m starting to think that summoning may be my only solution.

Yup.

There are very niche situations where a free market actually works - situations where there is no hidden information and no barrier to entry, where monopolies can’t arise due to the nature of the specific market. By the nature of these restrictions, nothing of any importance will ever be supplied by these markets.

The difference is the amount of effort expended on the video. If you actually make the video yourself, you want a good return on your time. That means - usually - that you don’t want the majority of your comments to be about how your facts are wrong, because you want people to come back and watch more of your videos.

AI channels, on the other hand, don’t care at all. They don’t need subscribers. They don’t need likes. They’re just putting out high volumes of crap content with clickbait titles to get whatever scraps they can sucker out of people. They know they won’t ever get recurring visitors, so they don’t need to fact-check anything. And AI has made it so that they can churn out this content ridiculously fast and make it nearly indistinguishable from quality content at first glance, at minimal cost.

If it takes you hours to make your video, I’m more likely to want to hear what you’re saying. AI videos are an attack against our attention; it takes less effort to make than it does to consume, now, and the quality of the content reflects that.

So the thing that I learned - which really improved my time with World (I haven’t played Wilds more than ten minutes because I play on PC, and it was TERRIBLE there) was that MH’s combat is all about positioning and early reading of enemy tells. You need to make sure you’re in the right place at the right time to get off a good set of hits without getting punished too badly, and you aren’t going to be able to dodge attacks like you’re used to doing in DaS games. Making good use of your slinger and the environment is also way more important than you might initially think.

There is also, however, the fact that it throws a bajillion mechanics at you without good opportunity to absorb them. It didn’t really click with me until I went through it with a seven-year-old.

LttP was absolutely top of its genre for years and years. It was a drastic refinement of the formula that LoZ pioneered.

OoT was a completely new beast, and had to innovate practically all of its gameplay - gameplay that, I will mention, has been refined since then even more drastically than LttP’s gameplay. It’s rough, and it shows. Unless you played it around when it came out, the adjustment after playing modern games would be … difficult.

all of the main bosses don’t put up that much of a fight

Try fighting Malaketh with strength/faith. I’ve hit a solid wall. It takes forever to get through the first phase, to the point that it’s become boring - and then I last about six seconds in the second phase because I can’t actually learn it when I have to slog through the first phase for so long.

And there never will be. Not so long as it is possible to hide information from the consumer, and any sort of barrier to entry exists for market competition to spring up.

Reviled by tech-savvy users, not suits. They get a lot more money from suits, so we can hang for all they care.

It happens in the new series too. Only one of the nephews manages to pick up the skill.

The underlying architecture for Windows has changed like three times since those systems were created. The fact it works at all is a minor miracle.

HDMI audio depends on a proprietary license. The Linux drivers for it are, predictably, less robust.

Don’t punish someone for practicing their art imperfectly. Instead of improving their craft, that only incentivizes them to stop practicing.

Yes. I use Kindle or Moon Reader +, depending on whether or not I can easily get it outside Amazon’s ecosystem. Actual ebook readers are vastly superior to using a browser or PDF.

Mostly I use my phone because it’s always right there. I’ll spend my lunch break getting another handful of pages through my current book while eating.

As an American…

I agree with every single bullet point.

If it’s something that you don’t really care about others seeing, that’s a prime candidate for cloud storage and more power to you.

This topic is about password lockers. I’m pretty sure you don’t want some schlub who happens to work at Cloud Password Lockers Inc. to be able to get at your PayPal account.

The problem is that it’s fighting against gluing a giant iPad to the dash, which just hides all the complexity behind a pane of glass so that it -looks- simple when the car is off.

With this setup, once you learn how it works, you can do just about anything with only a cursory glance away from the road to double-check your proprioception.