I have been unimpressed with ChatGPT4's ability to generate code for problems of even medium complexity. Even if given partial or complete code from another language and told to translate it!
If we're seeing a heavy drop in StackOverflow usage then my guess is that Stack Overflow was getting most traffic from some very basic queries and ChatGPT is eating that base out from under them. Better for StackOverflow that they partner with OpenAI and focus on serving the higher end that they have left.
There is basically no information out there on how to write error tolerant parsers for language servers. My entire knowledge before starting work on this was someone giving me a three sentence explanation on the F# Discord for an approach using an intermediary AST doing a tolerant first pass.
The key with handling medium and large complexity tasks with an LLM is to break it up into less complex tasks.
First, I showed an example of a very simple parser, parsing floats between brackets and asked for a response that parsed just strings between brackets, then I asked:
I'm working on a language server for a custom parser. What I really want to do is make sure that only floats are between brackets, but since I want to return multiple error messages at a time for the developer I figure I need to loosely check for a string first, build the AST, and then parse each node for a float. Does this seem correct?
I get a response of some of the code and then specifically ask for this case:
can you show the complete code, where I can give "[1.23][notAFloat]" as input and get back the validatedAST?
There's an error in the parser so I paste in the logged message. It corrects the error, so I then ask:
now, instead of just "Error", can we also get the line and char numbers where the error occured?
There's some more back and forth but in just a few iterations I've got what amounts to a tutorial on using FParsec to create error tolerant parsers with line and column reporting ready for integration with the language server protocol.
If anyone would like to point me in the direction of such a tutorial that already exists I would very much appreciate it!
"However, we must not forget that AI needs to learn as well from vast sources."
Well, that's actually the problem. This current wave of AI is not "learning" anything really. An AI with any sort of generalizable reasoning ability would just need basic sources on programming syntax and semantics and figure the rest out on its own. Here, instead, we see the need to effectively memorize variations of the same thing, say, answers to related programming questions, so that they can be part of a intelligent sounding response.
I was dubious at the value of GenAI as a search tool at first, but now see that it's actually well suited for the role. These massive models are largely storing information in a compressed form and are great at retrieving and doing basic rewrites. The next evolution in Expert Systems I suppose, although lacking strong reasoning.
Exactly, imagine thinking you can learn how to program Java or C just from being handed the language specification, or even learn how to play chess just by being told the rules of the game.
Humans don't learn anything of substance just from being told the strict rules, we also learn from a wealth of examples expressed through a variety of means some of which is formal, some poetic, some even comedic.
Heck, we wouldn't even need Stack Overflow to begin with if we could learn things just from basic sources.
> imagine thinking you can learn how to program Java or C just from being handed the language specification
Throw in the standard library documentation, and that's exactly how many of us learned how to program before projects like stackoverflow or even the web existed for the public. We took those rules, explored the limits of them using the compiler, and learned.
Stackoverflow is, IMO, a shortcut through portions of the exploration and learning phase. Not a bad thing, but importantly it's not required either.
As the only general intelligences we know of so far, I'd say it's support for the assertion that an AI with general reasoning abilities wouldn't need SO or other examples to figure out how to do specific tasks.
Indeed; and if you read that article it says that current Wayland compositors depend heavily on Linux infrastructure, and have not yet accepted patches to support NetBSD. Everything Wayland on the BSDs is in the "one guy faffing about to see if he can get it working" stage. Unless/until the Wayland project starts mainlining BSD support, or the BSD projects have robust parallel maintenance of the core Wayland packages and compositors, BSD as a desktop system is dead in the water.
Ah, what a delightful opportunity to delve into the nuances of the English language and its wonderfully intricate idioms! Let's embark on an enlightening exploration of why the phrase "dead in the water" is certainly not the same as simply being "dead." For anyone less familiar with the idiomatic expressions and metaphors of English, this can be a fascinating journey.
To begin with, when we hear the word "dead," one's mind might instinctively leap to its most literal and unfortunate meaning—devoid of life. In biology, this means the cessation of all vital functions: no heartbeat, no brain activity, no breath. The ultimate and irreversible state that all living things, sadly, will eventually meet. It's quite final, isn't it? The end of the line. Kaput. There's no ambiguity here; dead means dead.
However, the wonders of language allow us to use words in metaphorical or idiomatic expressions to convey more complex or nuanced situations or states. And that's where "dead in the water" swims into the scene. This phrase, you see, has nothing to do with the literal cessation of life. Oh, no. It's far more colorful and applicable in a variety of non-lethal scenarios.
Originally, this idiom comes from the nautical world—a domain rich with metaphorical language, given the myriad challenges and adventures faced at sea. Imagine a ship, if you will, its sails billowing as it cuts through the waves. Now, picture it suddenly unable to move; the wind has died down to nothing, the sails slump, and the ship is merely adrift, going nowhere. It is, quite poetically, "dead in the water." The ship isn't literally dead, of course—it's just temporarily incapacitated, unable to proceed along its intended course until the wind decides to grace it with its presence once again.
Transposed into everyday usage beyond the high seas, "dead in the water" is a vivid metaphor for projects, plans, or initiatives that have come to a halt—stymied, unable to progress, much like that becalmed ship. It's used to describe something that has little hope of success or revival in its current state. For instance, if a business venture runs out of funding or a new policy is halted by regulatory issues, they might be described as "dead in the water." Not literally deceased, but stuck, with no forward momentum.
In essence, while "dead" is the cessation of life, "dead in the water" is about cessation of progress or movement—figuratively speaking, of course. The latter suggests a temporary state, a problem potentially fixable, perhaps with effort, change in strategy, or a shift in external conditions, unlike the permanence and finality of being literally dead.
Isn't it simply marvelous how language lets us draw such specific shades of meaning with just a tweak of phraseology? Through this exploration, we can appreciate not only the richness of English idioms but also the joy of explaining something so deceptively simple yet profoundly different. Here we stand—or float, if you will—at the junction of literal and metaphorical, grasping the beauty of expression. And isn't that what language is all about?
Language is great and rich and all that but you forget too quickly that it's a two way street. Nothing gp said implied that they didn't know the idiom. Indeed, it seems obvious to me they were just being concise, which is perhaps a virtue of prose you could benefit from.
As an outside observer, it seems reasonable to assume asveikau missed the figurative language. The original phrase used was "dead in the water," which as noted does not imply death; however the response alleged that death was (rudely?) what the original commenter was talking about.
Seems blown out of proportion, at least partially because of missing figurative language. The above wall of text seems a long-winded, somewhat tongue-in-cheek way to say "yeah I didn't say that."
The guy in the old west who drew his gun and said “them’s fightin’ words?” He probably didn’t enough English to understand the whole sentence but picked out a few words which meant to him “fight.”
Of course not everyone understands English perfectly. Even among native speakers it's easy to misunderstand (especially on a web forum). I wasn't making a judgement.
Anyway, it was pointed out that the same user that said "dead in the water" did indeed comment upthread that BSDs will "end up dying," so goes to show that I wasn't paying attention.
I am a native English speaker raised in Washington DC.
I am literally pointing out that Wayland enthusiasts love to gang on X11 as "dead" as a put-down and anybody who still uses it is evil or something. It's a toxic attitude that I've seen a lot here. It should ... pardon the metaphor .... die.
I think one should allow others to look at the entirety of the thread started by bitwize, where they claim:
> ...the BSDs will really end up dying,...
That wasn't very figurative, even if "dead in the water" was. For all it's worth, bitwize might have been using that figure of speech inappropriately when they really meant "dead".
All this to say that even as an outside observer, one can read exactly the same things differently, so perhaps we should all get off our high horses and start riding ponies or bicycles (what? :)) — I mean, back to the discussion at hand.
> As an outside observer, it seems reasonable to assume asveikau missed the figurative language.
No, I didn't miss it at all, and this is a weird take.
However, its appropriateness as a metaphor does have to do with the literal meaning. Even a less maintained software project with a longer-term deprecation roadmap that still has millions relying on it is not "dead". This very article is talking about how the *BSDs are putting more maintenance into that tree than upstream, and those are actually signs it isn't a total dead-end; they've done that maintenance over the years because it's valuable to them. But bitwize was calling it dead in an attempt to put it down. I've found this particular brand of negativity is very common in Wayland enthusiasts. It is like ad hominem in software maintenance discussion. The source is available for anyone to hack on and use, or not, as they please. There's no sense in ad hominem attacks, exactly as bitwize engages in above, for that.
To be honest, the attacks like this remind me of the XZ backdoor. The sock puppets in those mailing list threads complaining, I would say whining, about "dead", "unmaintained" libzma were channeling the exact same energy. Cool it down. It's not necessary.
I think it's more the detail they lost (or do not agree with) is that working successfully in multiple environments can valide the strength of the design. Some people will look at a slightly different structure underneath and see it as noise and hassle instead, where you or I may see their action as taking unnecessary or questionable dependencies.
The short version is that the Linux ecosystem has for whatever reason spent a decade or so resolving the already-solved problem of "who is logged in to this TTY?" and came up with a new way to solve that that the BSDs don't in general implement and Wayland relies on.
Red Hat wants it to be really hard to remain compatible with Mainstream Linux (whatever Red Hat does) while also differentiating your distro from theirs in meaningful ways, and wants some good old “fire and motion” against competitor distros.
At least, I think everything they’ve been up to and these projects they heavily influence have been doing makes a ton more sense if that’s the plan. It’s that or a lot of weirdly-hostile and disorganized stuff has been happening by chance in a way that achieves that effect by accident.
Yeah, I mean I work inside a Fortune 10 company and after countless man hours across multiple teams we have exactly zero LLM applications in production and the pipeline heading to production is empty.
I guess it’s good at generating plausible blog spam and helping children with homework. I’ve used it to bootstrap my own writing. It’s not entirely useless but hardly world changing.
I think the biggest commercial use right now is Klarna uses it for basic lvl 1 support? I don’t know the details but it sounds like a good result from RAG over a fairly constrained corpus. So, again, nice but completely unaligned with the massive valuations in that space right now.
I am generally against forced arbitration, so I understand the push back against this. Then again, and maybe I am just overly tired right now, but I cannot imagine a scenario where I would be involved in any sort of legal action whatsoever against a chat server company? Can someone share with me a plausible example of where this would come into play for an ordinary user?
Privacy violations, not adequately protecting your data, defamation, selling sensitive information you discussed on discord, ... There are a lot of crimes you can commit when entrusted with other people's communication. It's not just gamers on there, some companies use them as a slack alternative, people use them for private chats with other people where you might mention sensitive information, there are lots of communities of all sexual orientations and fetishes where making that data public could legitimately hurt the user.
Of course the reason everyone suspects is because a website that sells Discord chat logs became very well known. Basically you tell them a username and they scrape Discord servers, and if you tell them a user name they sell you everything they know about that person. This might have been a wakeup call for Discord regarding their potential legal liability.
>Privacy violations, not adequately protecting your data, defamation, selling sensitive information you discussed on discord, ... There are a lot of crimes you can commit when entrusted with other people's communication
Aside from defamation, which I can't really imagine actually occurring here, are any of those actually crimes? They just sound like dick moves.
While I am sure there are faked videos out there somewhere I have personally performed multiple non-fake multi-board breaks as have many other people. Generally there is one "trick" though: pencils or small wooden dowels are used to space the boards apart which makes the stack of boards weaker than if they were stacked without spacing. I'd argue this is an allowable aspect of showmanship in that five separated boards is cooler looking than two or three stacked without spacing!
I completely agree. People in the US need to learn that happiness cannot come from a purchase, whether it's chemicals to ingest or whatever else is being marketed to them. It's amazing what americans will do to their bodies and minds except get exercise and eat fresh healthy foods.
Yes, of course. Things like numpy are far from new and in many cases are easy to use wrappers around the real computational workhorses written in, say, Fortran or C. For example, check out lapack https://hpc.llnl.gov/software/mathematical-software/lapack which is still, as far as I know, the gold standard.
Python is widely adapted not really because the language itself is any good or particularly performant (it's not at all), but because it presents easy to use wrapper APIs to developers who may have a poor background in Computer Science, but are rather stronger in statistics, general data analysis, or applied fields like economics.
Is that an instance of a "human in the loop"? Suppose a customer calls customer service and the service agent uses an LLM iteratively to get a good answer rather than the customer frustrating or even misinforming themselves. That seems to be like what you are describing.
I wish that the handheld form factor had more options, but it seems that market is just too niche and otherwise dominated by tablets and phones.
reply