Install Steam
login
|
language
简体中文 (Simplified Chinese)
繁體中文 (Traditional Chinese)
日本語 (Japanese)
한국어 (Korean)
ไทย (Thai)
Български (Bulgarian)
Čeština (Czech)
Dansk (Danish)
Deutsch (German)
Español - España (Spanish - Spain)
Español - Latinoamérica (Spanish - Latin America)
Ελληνικά (Greek)
Français (French)
Italiano (Italian)
Bahasa Indonesia (Indonesian)
Magyar (Hungarian)
Nederlands (Dutch)
Norsk (Norwegian)
Polski (Polish)
Português (Portuguese - Portugal)
Português - Brasil (Portuguese - Brazil)
Română (Romanian)
Русский (Russian)
Suomi (Finnish)
Svenska (Swedish)
Türkçe (Turkish)
Tiếng Việt (Vietnamese)
Українська (Ukrainian)
Report a translation problem
Discussions Rules and Guidelines
Report this post
Maybe when choosing mods for your run, you disabled everything first and didn't think about it?
Yup, I can see that from the limited info available. You did forget the whole time thing though. I really don't have time these days to read everything I would like to! ... but believe it or not I am a pretty darn good developer, if I do say so myself! :D
That is exactly what I did now that I think back on it! ... <doh!>
I am part of a fairly small development team at a financial services firm. And we have all increased our productivity by around 30% using ChatGPT.
You do need to know how to prompt it properly to get the results you want, and the free version kinda sucks - but your comment is just wrong. It is super useful for many things, including information gathering and aggregation, especially in gaming. It is rarely wrong with most things... but ya, it did screw the pooch pretty wildly in this instance. So the takeaway is something we always do anyway: Verify that ♥♥♥♥ before you paste it in production! :D
There are things ChatGPT can be useful for, but truth isn't its forte.
To make sure cheese sticks to pizza one can use a bit of non-toxic glue underneath it, if you can recognize that meme.
Also funny how spending 5 minutes on looking through the patch notes for 2.0.7 could've saved dozens of hours on building the factory to rocket in the base game XD
One you can't disable, but it's still in the mod list.
The likely reason for it not to be on by default could be because you had other mods turned on that caused an error during loading and you clicked on "disable all mods".
Especially when asking it about information that is very fresh, like how game mechanics work in an expansion that has only been released a few days ago. It just didn't have enough time to learn about Factorio: Space Age yet, but it still spouts wrong information with the same confidence as every other piece of information. That's why it's always important to double-check, which usually isn't much work and can still be done faster than researching stuff the old fashioned way. But don't just trust it blindly.
Using ChatGPT for software development is actually one of the worst things to use it for, as programming languages, while very structured languages, are as far from nominal written human language as possible - second only to pure math.
And LLMs perform horribly at it, because they're in essence nothing but glorified 8-balls running on a stochastic prediction model of what is the most likely next word in normal written language, that the end-user would like to see printed back.
For instance, you used to be able to ask ChatGPT what the answer to 1+1 was, and it would adamantly try to convince you that the answer was 3. Its staff fixed that type of rote arithmetic to make it 'understand them' (aka; pushed in a few hard overrides, probably) - and afterwards it would indeed answer 2 -- until you'd ask it: "2, are you sure it's not 3?" And it would again try to placate the end-user by answering: "why, yes -- you're right; it should be 3." Because LLMs are dumb-as-brick parlor tricks. They can't actually apply logic or reason. They'll just assume that what you directly tell them is correct. (Worse still; they're trained on a corpus of data that's largely scavenged from the public Internet. Which is full of idiots who don't know what they're talking about. Think about that for a bit!)
Now combine that with programming. Where you have to be meticulous about every little thing happening while performing arithmetic; or happening while manipulating the innards of a complex data structure according to the recipe defined by an algorithm; recipes which often are counter-intuitive at face value.
You're asking an AI ('artificial incompetence') to dredge up the answer to complex algorithmic problems you don't know how to solve. And you can't rely on your intuition to verify whether what it returns is actually correct or is a case of cyber-meth induced hallucination.
Yeah-- that will end just great.
Now compound that with the fact that another major part of actual real-world software development relies on third-party libraries, frameworks, APIs, etc. All of which represent a rapidly shifting landscape. I.e. an LLM trained on information for a framework or API's prior version from a year ago, is probably going to be worthless. Worse-- it does not comprehend actual concepts such as "software versions" and thus if you train it over multiple versions it will start to mix and match both.
(I've seen this happen in practice btw. With colleagues struggling to make sense of the nonsense it was outputting to them while hamfisting two conflicting APIs together. In the meantime, I actually pulled up the then-current documentation; read through it; produced a working sample; and handed it to them, while they were still trying to make sense of ChatGPT's tossed salad mess.)
so glad I skipped, kinda, to the end /s