About LLMs at Zig Days | Hacker News Reader

About LLMs at Zig Days

kristoff.it

82 pointskristoff_it73 comments

For those who aren’t aware, as I wasn’t, this is coming from the “VP of Community” at the Zig Foundation. So the proposal to soft ban LLMs at Zig Day meetups seems like it has a bit more weight than if it was some random community member.

(I’m not a member of the community, so not fully aware of the dynamics.)

Isn't "soft ban" still a bit harsh? I think it was reasonable take and a good reminder of the purpose of these events.

> Isn't "soft ban" still a bit harsh?

Not necessarily. The take is reasonable but I'm curious about who could be bold enough to actually talk about or disclose their use of LLMs during these events.

The whole reason for this blog post is because discussion about LLMs does happen already (to the point of being a bit suffocating).

It's funny (or sad) how I read this as an address about pro-LLM speech when "talking about LLMs" doesn't have to be just that.

I appreciate the clarification.

Only in the sense that community leadership is harsh by nature. ‘We’d rather you didn’t do that’ is the essence of a soft ban and what they’ve expressed as leader here, and even a soft ban muffling topics and therefore a subset of people, which is often seen as ‘harsh’.

This didn't sound like a "soft ban" to me.

With things happening in general, and with Bun's LLM-aided move away from Zig in particular, there is bound to be some interest in talking about LLMs and how that impacts Zig's future.

I think this was a well measured "hey, let's focus on thing we are coming together to celebrate and advance: Zig".

I had the pleasure to talk to Loris about this topic just last week and I've liked his take very much.

His point is not coming from a place of LLM demonization. He very much acknowledges their usefulness, especially in a business context, e.g. for implementing yet another standard CRUD application and for shipping all the other "average" (in quality) business features quickly.

His point is a different one entirely: Say Andrew Kelley is attending the Zig Day. Why would you ask an LLM about a Zig programming problem you're struggling with instead of learning from the man himself? There's simply no LLM as knowledgeable about Zig as Andrew and the other people working on it or with it on the daily.

In other words: Zig Days are an opportunity for people to learn from each other and to spend time together (= the "Community" in "VP of Community"), and LLM are diminishing this opportunity.

Besides, Zig itself is mainly a language for people who care not just that a problem is solved but also about how it's being solved. ("Create software you can love.") While LLMs don't prevent anyone from doing so, they make it much more appealing to just vibe-code everything and not look too closely at the implementation.

Being negative against the utility of AI is sometimes besides the point or a strawman (don’t know about this instance). Only an idiot would argue that cars don’t have utility in cities and countrysides that are already paved over. Or that they are slower. Or that the train can get you from Bob’s to Burger Establishment with less walking. But they may want other things like more walkable cities and less mass extinction in a hundred years time.

No one needs to proclaim the utility of The Car before criticizing car culture.

This piece already says that all the clanker maximalism may be correct. shrugs Then it says that this get-together is for people who like programming. Even if the whirlwind of progress comes and takes their profession. Because then it could still be a hobby.

And this is too negative-against-AI for some people in this thread? Programming as a hobby? Okay, fine. Maybe we will have sold off all our RAM in a years time and the Government will have outlawed unassisted programming as too dangerous. The piece is too optimistic.

LLMs are just pushed by everyone at us, people just can't stand hearing about it anymore even if they use it and find it helpful. It's like at PyCon2026 the keynote was about LLM, and people were just leaving in the middle, including GvR.

It's hard to avoid the topic when it literally redefines what it means to create software. If I'm using it to create some piece of software, then I turn around and say "I wrote this" am I even being truthful? But if I'm trying to avoid mentioning LLMs what other wording could I use?

Maybe just don't talk about stuff you commissioned (not created), it's entirely uninteresting to everybody else.

> It's hard to avoid the topic when it literally redefines what it means to create software.

Say that the IDE also "redefined what it meant to create software" when it entered the ecosystem as an idea and product, does that mean every conversation, community meetup and thinking needs to consider the IDEs now? Probably not, then there is no more room for the other topics anymore.

Classically, this is when a large-sized BOF conference might form an “editors” committee, delegate the entire topic to it, and ask the committee to elect one presentation each day that the committee feels would be valued by the conference as a whole. It maintains the enthusiasm for those who truly value discussing that subtopic in detail, and it’s an effective tool for keeping evangelism and/or holy wars within a given group sect from sucking all the air out of the room for everyone else. (Modernily, subreddits are an expression of that exact same model, and remarkably effective at scale for that purpose.)

You could say it's more akin to if we have all been writing in a assembly and suddenly we got access to compilers and high level languages. Would we all be complaining that every conference is "about compilers"?

This is a meetup for Zig the tool, so you would need to identify how you are interested in Zig to attend and not run into problems.

“I want to [verb] with Zig someday and want to show up and listen and learn”

“I [verb] with Zig and have formed opinions and want to swap them with others”

“I [verb] with Zig and have not yet formed opinions”

If you can’t identify a verb for such a sentence, then you probably need to gain some vague clarity on why you’re considering attending.

But if your sentences are all “I [verb] with LLM”, then there’s no point in attending a Zig meetup; attend an LLM meetup instead. “I [verb] with LLM and the LLM [verbs] with Zig” isn’t transitive to “I [verb] with Zig using LLM” in human social relations; that difference matters, even though a logical evaluation would claim that ( A & B ) & ( B & C ) = A & C. People are extremely sensitive to the difference and Zig has labeled their events as A & B, not A & C.

Specific example: “I code with LLM […] in Zig” would be offtopic, because there’s no human verb-use of Zig present; the verb “code” is bound to LLM, not to Zig, and so is not a valid basis for human connection over a shared interest in Zig.

Specific example: “I write out Zig programs on paper first” would be ontopic, but “I write Zig with pencils rather than pens” would be offtopic; even though both refer to the same activity, one is about how you perform a creative act within your self to output Zig, the other is best reserved for a stationery BOF.

(This holds true for all “I [verb] with [noun]” BOFs and is a good general principle for when to, and when not to, bring up LLMs at a Noun event. You can swap also “LLMs” for “employees” and get the same outcome: don’t go to a Noun BOF to talk about managing Noun workers; instead, go to a Managers BOF to talk about Verbing.)

You're assuming a file containing a statistical model that exists on your hard drive has agency. It doesn't. You as the user have sole agency and what you use the statistic file for.

> You as the user have sole agency and what you use the statistic file for.

Along with OS provider and any software developers that have included exfiltration and command vectors within any tools used on that drive.

Modern computers are sufficiently layered and complex that even cutting edge experts are surprised by what third parties have agency over local data.

Have you tried creating the software yourself?

>It's hard to avoid the topic when it literally redefines what it means to create software.

It redefines it because its shoved down our throats as redefining it.

BUILD 2026 is happening next week and from the agenda I already seen it is worthless to spend any time looking at its sessions, other than probably what Mark Russinovich has to tell about Azure, almost everything else is AI.

This seems like a very positive move to me. It makes it clear what Zig days are about so people know what to expect. There are 1000s of other events for people who want to build with LLMs so it's nice to have some recourse for those who still want to code by hand

The giant Umarell in the background is a nice piece of furniture.

Edit: I noticed later it was in Milan, I guess it makes perfect sense.

>My recommendation is also to not choose an extreme approach (e.g. by completely banning LLM-related discourse) unless you feel very strongly about it.

Organizers are allowed to ban the mention of certain programming topics? I could understand if it was a topic that was adjacent to violence/harassment/sensitive stuff, but come on... are anti-AI groups becoming a cult?

Yes, organizers can do whatever they want. Their event, their rules. If you put in the work to run a Python meetup, you are free to ban discussion of tomatoes if you really want to. Conversely, those with tomato psychosis are free to avoid your event if they truly can't survive a few hours without talking about tomatoes.

Given your example, how exactly would tomatoes be relevant to Python?

That's the point. They aren't.

A broken point, as LLMs on the other hand are very relevant to coding, which is done via programming languages. And it's use in other - particularly knowledge dependent - areas is rapidly exploding.

Let me introduce you to the HomeAssistant gardening community, and note that the very first Dashboard screenshot shown opens with two separate entire beds of tomato plants:

https://www.briandorey.com/post/pi-pico-lora-remote-soil-mon...

If one is leading a Python (or a Zig!) meetup and has an aversive reaction to tomatoes, and said meetup has been all up about gardening for the past three events, it’s very plausible to imagine them soft-banning people from bringing tomato plants to show and tell. Sure, they’ll be teased mercilessly about it at first, and some jerk might troll them about it and get the boot, but ultimately it’s just an outlier human foible that will be accommodated with mostly grace and humor. And then there will almost certainly also, immediately, form a Tomato Cabal who goes out of their way to farm tomatoes with Python/Zig while specifically keeping it a wink-nod secret that they’re doing so, and someday the organizer will discover this and laugh in spite of their disgust and give the cabal an honorary Tomato Day event where the cabal leads that day’s proceedings and the usual leader stays home and watches desert movies and eats chips and salsa with secret, vicious glee and never tells another living soul.

Healthy human BOFs are the best form of social ever :D

That's going pretty far out on a limb (hah!) to create some artificial relevance given the context. Meanwhile LLMs are easily relevant in almost every way that matters to programming languages, more than even IDEs.

That doesn’t limit formation of a subcommittee to concentrate their discussion for the health of the group. Any topic is eligible based on group impact regardless of its perceived breadth by those avidly discussing it.

(There is no cabal.)

Subcommittees/subgroups are perfectly fine of course. But, again, given the context, this isn't that case. This is a core person in the ecosystem making a strong suggestion about how general groups should operate. Doesn't take much imagination to see things get to a point where only groups that seriously discourage or outright ban anything re LLMs are considered "legitimate". It's their prerogative of course but... lol.

I think you misread that. He's recommending not completely banning the subject.

the part the parent appears to be commenting on is:

"[...] unless you feel very strongly about it."

i.e. complete ban is okay if the organizer feels very strongly about it.

Ahhh in yes, that makes sense. I get it now

I was once in a hackerspace that held regular "show and tell" nights for people to present interesting technology projects. We eventually hit a rash of non-regulars bringing in "projects" that were essentially sales-pitches for devices sold through multi-level marketing scams. Figuring out how to ban those without blocking someone who intended to monetize their projects was tricky.

Point being: just because a thing technically fits the genre does not mean it is something that the audience wants to listen to.

Generative AI has and their providers have become an implicitly political subject. It shouldn't come as a surprise.

> are anti-AI groups becoming a cult?

I would say that they're more becoming a vocal minority of intolerant people towards some ideas. Some have good reasons, others just love to be on the counter culture bandwagon.

I predict that we will soon have a group like PETA that is primarily anti ai. I also predict that we're going to have for some time a schism in the developer community, which we already see in hn today, but it will grow wider.

Just ban LLM. It's events for human beings, not LLMs.

I like the idea of "it's not about LLMs" better as a starting point. For example, someone could peel off from a shared coding session, ask an LLM to try a concept, then bring that LLM chat to the group to explore (and then yes close that window and go back to face-to-face). Make more efficient use of the face-to-face time.

You could be right but I can think of numerous frustrating code jams in my past when we burned a lot of precious face-to-face time on fussy setup or other fiddly stuff.

> You could be right but I can think of numerous frustrating code jams in my past when we burned a lot of precious face-to-face time on fussy setup or other fiddly stuff.

Agreed, LLMs are particularly good at this kind of task.

For example: my windowing system on Linux would intermittently freeze. Diagnosing it was a pain--so I bounced the logs off the LLM. It gave me a couple of hypotheses and the commands to enable the correct logs for when it happened again. After the third time it happened, it pinpointed that a particular USB hub was causing the issue. I removed the devices downstream from that hub and haven't had an issue since.

That doesn't mean anything in the context of a Zig Day. People will come to the event full of new experiences to talk about that relate to LLMs and also with some worries about the future of their profession, wich also will relate to LLMs a lot.

Not only these are generally reasonable things for a human to want to talk about, but what is happening in the tech industry is definitely on topic for an event like Zig Days.

The problem is when this consumes nearly 100% of the communication bandwidth detracting from the main goals of the event (applying systems thinking & making software you can love).

Agreed, it is not a AstroTurf marketing event for hype.

zig is a cool language, and worth learning about. =3

That’s correct! Since Zig is openly sponsoring these events, it is indeed by definition not astroturfing, which is only applicable with a hidden/lying sponsor.

(I know nothing about Zig, but I wanted to directly appreciate your accuracy of word usage regardless :)

"You are technically correct... the best kind of correct." =3

https://www.youtube.com/watch?v=hou0lU8WMgo

There's an irony in your comment. On one hand, it's clearly starkly anti-LLM, but on the other hand, you treat humans and LLMs as similar categories, accepting a strong pro-AI framing.

You would never say "events are for humans, not search engines" as if search engines were a similar category to humans.

"Cities are for humans, not cars."

That's like saying ban IDEs. Or ban search engines. LLM is just a tool that humans use to create things.

This is so sad. "Maybe - just maybe"

> [...] the best career move is to become proficient at buying more tokens orchestrating agents, but I would still recommend not putting all your eggs in one basket just yet because maybe – just maybe – there will still be some value in knowing how systems work, both to differentiate yourself from other developers career-wise, and as part of effective LLM steering.

This should be read sarcastically. Its an idiom in the US. You state something you view as obviously true while qualifying it with "maybe - just maybe". Its commonly said in a comedic tone.

When it's spoken .. sure.. but as I read it, I wasn't so sure anymore.

Approaching significant change with humanity asks us to have empathy for many emotions at once. With respect to LLMs & other generative models those include but aren't limited to:

* Excitement from people who are able to make things they could not,

* Fear from people who's livelihoods are threatened,

* Betrayal from artists whose work is being ripped off,

* Alarm from activists looking out for ecosystems & the climate.

To add to an already-difficult challenge: many people, corporations, & governments are pushing extreme greed, hubris, & dehumanization for various reasons.

This piece does an excellent job laying out its recommendations with sensitivity for people of different perspectives & positions. I very much appreciate that.

Let me fix your logical fallacy of composition:

" * Excitement from people who are able to [unaccountably plagerize] things they could not,

* Fear from people who's [business IP rights] are threatened,

* [overt copyright theft from] artists [and chat bot users] whose work is being ripped off,

* [well funded denigration of] activists looking out for ecosystems & the climate. "

Thankfully LLM are not real "AI", and modern hapless 'slavery with extra steps' plans will eventually end badly. Popcorn and bubble infrastructure liquidation fund standing by... =3

Why do people insist on using extreme rhetoric like this? If you don't personally like using LLMs, that's fine. The only point this comment serves is to stir the pot.

Using your first example, if it was true and universally accepted that this was plaigerism--we wouldn't use it, now would we? But that's not the universal opinion so instead you're just twisting someone else's comment to stir the pot.

Again, if you don't personally like LLMs and you personally feel like it's plagiarism cool, don't use them. Or at least make an argument for it.

But as it stands, this comment is just low-effort trolling.

Joel_Mckay is responding to a glib comment that sugarcoats and trivializes the theft. Just like you do.

People need to be reminded of reality in their newspeak bubble.

Re-stating an opinion does not somehow establish it as a fact. I'd welcome an effort to support what you're saying instead of just hand-waving this as some sort of self-evident fact of the universe.

Why would people feed a sealion?

Most of OpenAI CEO rhetoric doesn't hold up to even cursory reason.

"People talk about how much energy it takes to train an AI model - but it also takes a lot of energy to train a human. It takes about 20 years of life - and all the food you consume during that time - before you become smart." (Sam Altman)

Let us take a worst case upper favorable estimate of his flippant statement:

Adult Human peak power W: 600

Average U.S. home yearly kWh: 10500

20 Years = 7300 Days = 175200 Hours

Human_metabolic_kWh = 105120

Home_shelter_kWh = 210000

Total_20Y_Human_Life_kWh = 315120

US average electricity cost is @17.65¢/kWh, and according to sources the LLM models took “1.2 million kWh for older foundational models to upwards of 11 billion kWh for next-generation frontiers". So if true that means:

1200000 kWh = $211,800 USD

11000000000 kWh = $1,941,500,000 USD

In other metrics that means we know:

1200000 kWh trained LLM = 4 Human lives reallocated

11000000000 kWh trained LLM = 34907 Human lives reallocated

Sam claiming 34907 Human lives is a small price to pay for a 85% accurate nonsense machine is ridiculous hyperbolic nonsense, and or the data they are publishing is inaccurate.

People have a right to be displeased with the industry dismissing peoples concerns. =3

> Why do people insist on using extreme rhetoric like this?

I simply narrowed the logical specificity of why LLM may be avoided in some use cases. No one can 100% prevent theft, as some people will decide personal desperation excuses philosophical compromises. It differs from a sociopath anti-social behavior, which is a constant aspect of civilizations. People can choose to be upset, or recognize it is a facet of some in "AI" gilded blitzscaling.

LLM are good at context search, and have other tangible use-cases that does not require constantly stealing from other people.

Have a wonderful day, =3

You claimed to "fix a logical fallacy" by merely stating your opinion that this is theft. Many (most?) disagree with the premise that using an LLM is somehow tantamount to theft.

I invited you to attempt to present an argument, which you have neatly side-stepped.

That is a long conversation, but one is welcome to watch a short documentary on the subject. =3

https://www.youtube.com/watch?v=zyQwAhppWj8

https://www.youtube.com/watch?v=ERiXDhLHxmo

This rephrasing is directly unhelpful to the goal of empathy for the humans caught in the change. If we come off as insensitive we will have no hope of influencing people. Also if you see a specific fallacy, please do name which one so I can improve.

All that said, I personally unequivocally agree with each of your points. I hope you are channeling this rage not only into comments sections but also into the hard work of tearing down & replacing the many incumbent systems that plagerize, denigrate, steal, oppress, monopolize, waste, & enslave. I certainly am.

"This rephrasing is directly unhelpful" is an Informal fallacy.

One does not need to "unequivocally agree", as facts should be verifiable. =3

https://en.wikipedia.org/wiki/List_of_cognitive_biases

https://www.youtube.com/watch?v=aNSHZG9blQQ

Sorry to ask again but which informal fallacy does it seem to you that I am using in my first statement? There are several [0]. In the interest of understanding, maybe I can make my passive phrasing a bit more active: when I say "this rephrasing is directly unhelpful [to] empathy", I could also say "if I tended to use the phrasing you did I am confident I would find myself with a reputation for being unempathetic i.e. rude". If you disagree with the statement it'd be great to hear your argument, but again disagreement does not imply fallacy.

Your second statement strikes me as using two specific fallacies while also being severely out of touch. First, a moralistic fallacy [1] where you assert what facts "should" be while making a statement about what is or is not needed. Second, a false equivalence fallacy [2] where you imply without explanation that verifiability & not-needing-agreement are equivalent. I'm open to argument, but it seems to me that the two describe independently varying spaces: facts describing the space of reality with agreement describing alignment in the space of conviction. Finally, your overall statement is very strange to hear in the big 2026 when so many important & verifiable facts are so widely & disastrously disagreed upon. See the ongoing USA vaccine safety scare & measles outbreak as an example [3].

Finally, pointing to general & easily-accessible resources in response to a specific question is normally understood to be condescending & a form of insult. I'm not sure if you meant it that way but graciously declining to answer is almost always a better alternative.

[0]: https://en.wikipedia.org/wiki/Informal_fallacy, see paragraph 2.

[1]: https://en.wikipedia.org/wiki/Moralistic_fallacy#Moralistic_..., the last example specifically names should-is equivalency.

[2]: https://en.wikipedia.org/wiki/False_equivalence

[3]: https://en.wikipedia.org/wiki/Measles_resurgence_in_the_Unit...

I hope to hear from you but either way this will be my last reply here.

That is an unfortunate detractor, as I normally do hold well-reasoned arguments in high regard.

Generally, being flippant with others pAIn while ignoring their legitimate grievances is not kind. We should also not get emotional over discounted GPU hardware when the LLM bubble inevitably collapses. =3

Be interesting to see where Zig and ecosystem is in a few years with this general anti-LLM stance from it's core people. My guess is it'll just make it's way as a hobby language, left behind in the dust. Which is of course a perfectly fine thing for some.

Naw, limiting LLMs for an event that's specifically about learning, growing, and collaborating makes a lot of sense. If it ends up dying it won't be because of their stance on LLMs for a conference.

> I would still recommend not putting all your eggs in one basket just yet because [..] there will still be some value in knowing how systems work, both to differentiate yourself from other developers career-wise, and as part of effective LLM steering.

the thesis is that investing in your skills outside of LLMs pays dividends whether you decide to apply those skills to LLMs or not, plus spending time bonding with your fellow engineers is good for you too. so I'm sure Zig will be doing great in a few years

>> there will still be some value in knowing how systems work, both to differentiate yourself from other developers career-wise, and as part of effective LLM steering.

That seems like a strawman to as I can't think of anyone making a reasonable argument against. After all we still have C and even Assembly developers out there, despite the many languages-that're-more-convenient that've sprouted since.

No need to worry about manual coder hubris. This piece covers all clanker possibilties/bases.

> And even if you have full confidence that the future of commercial software is strictly hands-off agentic coding, Zig Days are still for people who enjoy the act of programming, even if that were to become just a hobby.

Maybe they’ll even get to enjoy their hobbies for a few days without worrying about getting left in the dust (perfectly fine).

My stance is that it already is like that, now TigerBeetle, Ghosty and Roc are the only major projects using it, and none of them are something that drives adoption of new programming languages.

Bun could have been that, if addons were to be written in Zig, that is now gone.

So unless something comes up where "you need to use Zig to play here" becomes a requirement, it will stay as a language where some folks have fun using it, and that's it.

People from ZSF and other maintainers have had a pretty clear stance that, while they don't necessarily like LLMs conceptually, they don't really care about if you will use them for tooling or development of your own projects. The anti-LLM stance has been on things that directly affect development of Zig (communication around issues, feature/pull requests, etc) and now, an event which is meant to be a connection place for Zig developers around the world to show off their projects and talk about other projects, and I'm sure you understand why this is a nice place to have human on human communication be the primarily encouraged method.

None of these really affect the end user of the compiler of making functioning, good tools with the language, with a LLM, if they wish to. Ghostty uses LLMs extensively, Bun was essentially vibe coded even before the Rust port. You might not wish to, ideologically speaking, develop in a language built by people who don't like your method of building things, but it's not a blocker that will turn the language into a "hobby language" (if we are judging hobbyism by lack of AI usage)

Reply on news.ycombinator.com