Hacker Newsnew | past | comments | ask | show | jobs | submit | brandon272's commentslogin

You were right to notice the complaints. One decision remains, and it is yours, genuinely.

I appreciate humor here, but there are now a dozen of these comments on every thread about Claude. They no longer adding anything substantial and dilute the discussion.

I don't mean to pick on this comment in particular. The majority of my work day is now spent reading AI generated text, and I look at HN (too much!) because I want to read human commentary. Humans pretending to be obnoxious AI on repeat is net negative to say the least.


I agree. Hopefully Anthropic has fixed Opus' ridiculous communication style so that people - like me - no longer have any kind of weird impulse to imitate it.

It's a load bearing joke that was funny the first time but we're going to beat that dead horse until it starts getting funny again. If you beat it enough, it will get funny. Beatings will continue until morale improves.

For a lot of people involved in local LLM, ownership and control of their data matters to them a lot. It's what pulls a lot of people toward local models. Renting a 5090 off of some random infrastructure provider that is reselling some random dude's 5090 does not fit the bill.

Local models are undeniably capable of "helping with coding" today.

I so want this to be true, but for the kind of coding I do (not Flask apps), it's definitely not the case. Like I said, SOTA models just barely, barely work for me. My projects are usually 100k-1M lines of Rust or Go.

Out of curiosity, what do you find the SOTA models are simply incapable of when it comes to your Rust and Go projects?

The SOTA models now work really well in my codebases, but that's only been since Opus 4.5/4.6-ish. Prior to that, and with current local models, they simply couldn't work holistically and would just thrash around. Now I feel as if SOTA are approaching my coding levels if not surpassing it. I still need to guide on architecture, but I can see that going away within the next year or so as well.

Thanks, that makes sense. When you said they “barely, barely worked” for you I assumed that meant something different.

Oh yeah, that makes sense, sorry! I meant they just started working well and did not until relatively recently.

This. People have convinced themselves that the absolute frontier is what is needed, anything below it is an unacceptable compromise, and we seem to be speaking different languages when it comes to discussing model capability.

It's like watching a discussion about cars available to take on a 100km road trip. A new car gets released that is on par with a Toyota Corolla but it is dismissed as completely useless for a 100km trip because it doesn't have the seat massagers and air ride suspension that the new Escalades have.

The reality is that something like Sonnet 4.6 is still amazingly capable for so many programming tasks, especially if you already have some reasonable level of experience to steer it in the right direction.

And if you think Sonnet 4.6 is still worthwhile, then it seems undeniable that something like Qwen 3.8-27B is also worthwhile.


The problem is that even if you're doing CRUD apps, Sonnet level will be good enough... 95% of the time. But the 5% will kill you.

What do you mean LLM slop? OP says they built it themselves.

There are very clear signs this was mostly built by an LLM. Just an example: https://github.com/Qbix/webserver#%EF%B8%8F-architecture misaligned lines, the commit history, etc.

I use LLMs in my toolchain, same as I also use compilers. In fact, I use the LLMs to generate end-to-end test suites, and the test running is automated as well.

Yeah, and surely they also wrote the 131KB unreadable-wall-of-text README themselves.

That's like 1/9th Moby Dick, to put into perspective.


Yes, I actually iterated a lot on that README and personally directed the LLM to produce every section in there, as an artifact. Normally it would produce a much shorter README, and I should probably have split it up. That is exactly what I did on the actual website.

The README is no longer than nginx documentation, though, or many manfiles of programs. The README explains all the features in one file, with results of actual benchmarks. I asked it to put emojis in there as well.

I use LLMs in my toolchain, same as I also use compilers. In fact, I use the LLMs to generate end-to-end test suites, and the test running is automated as well.


I don't care how much you were or were not involved with the README, no one is reading through all that. If you can't be arsed to, at the very least, write the prose around your project, why should we devote any time engaging with it?

> I use LLMs in my toolchain, same as I also use compilers. In fact, I use the LLMs to generate end-to-end test suites, and the test running is automated as well.

Yeah, I could see that.


Alright, so version 1.3 has a much shorter README: https://github.com/Qbix/webserver

What do you think?


over 100K lines of code materialized over the course of 3 weeks and is full of em dashes. I think we can reasonably assume here...

Ran into the same issue with my dad the other day. Has a passkey set up on his Google account. Bear in mind that he doesn’t know what a passkey is, so Google obviously sent him through a pattern at one point to get him to create one.

He didn’t have access to it the other day and we needed access to his account. He didn’t remember his password, and we were unable to reset it because you need the passkey! No other options to authenticate for a reset were available.

Add in the fact that I was trying to help him with this by long distance call and you can imagine the frustration.


As someone with elderly parents, the Googles and the Microsofts of the world really don't seem to give a shit about the elderly anymore, if they ever did. Random UI updates and workflow changes with no announcements, thinking they're "intuitive."

One of these days the product managers who push these things thinking "oh, it's easy, you just . . ." are either going to be explaining it to confused Mom or Dad, or they're going to be elderly and irritated themselves. Until then I hope they stub their toe or step on random Legos regularly.


I'm not even elderly but I would really like an e-mail client with a stable UI. I used to like Apple Mail but it's been a buggy mess since Catalina. It doesn't even sort or search mail reliably any more. It's basically a trope at this point but a robot-rewrite-in-Rust seems inevitable at this point.

A while ago I helped an elderly relative migrate off off a copy of Eudora that they'd been using for many years. (To Thunderbird.)

Originally it was about difficulty migrating to a new laptop with a different version of Windows, but I was quite firm about it because when I realized it wasn't able to do secure connections for some reason, so the instant they took their laptop to public Wi-Fi...


P.S.: Migrating decades of local messages was a huge problem. It's been so long that many import/export tools are defunct. I was even loading up extremely old versions of Thunderbird to see if I could import to that and then upgrade.

However the GitHub edition (not Sourceforge) of Eudora2Unix [0] saved me and deserves a shout-out here.

It took some tweaks and a harness to repeatedly try the conversion, since I wanted something I could literally drop into the Thunderbird profile folder. The final result wasn't perfect, (some file attachment issues linger) but it's way better than having it all at risk of Eudora.exe just refusing to launch one day.

[0] https://github.com/jonabbey/eudora2unix


>He didn’t remember his password, and we were unable to reset it because you need the passkey! No other options to authenticate for a reset were available.

Google treats both a password and a passkey as a primary factor, and if you forget either of them you have to go through their account recovery flow: https://support.google.com/accounts/answer/7682439?hl=en

AFAIK there's nothing different about the recovery scenario for a Google account in that state regardless of whether it has a password in use as its primary cred, a passkey in use as primary credential, or both.


Thanks. Upon further investigation we could have hit “Try another way” on the account recovery process UI a few times to get to a workable account recovery method.

Would be nice if Google would lay your recovery options out for you (which I am used to it doing in regard to 2FA if you are doing a regular log in) instead of having to hammer the “Try another way” link repeatedly as it cycles through options.


WDYM -- I thought "try another way" is supposed to list all possible options instead of cycling through them!

It does when I log in. You click "Try another way" and then it brings up a menu of options you have for 2FA. For some reason they have designed the account recovery flow to be different.

So, if you are used to the "Try another way" flow on login, it can be confusing to see an entirely different "Try another way" flow on account recovery.


I set my father (late 70s) up with a physical passkey (yubikey), and a backup key. He uses it for the important accounts (google, apple, bank, etc).

It’s been fine.


Yubikeys generally have much worse recovery scenarios than passkeys do, for consumers. In enterprise if you lose your yubikey, an IT admin can help you get back into your account. If you lose a security key as a consumer, you're generally in a much tougher account recovery situation.

You need extra backup keys in a safe place. They don’t explain that well.

Not all services let you enroll multiple keys. Amazon, with all the money in the world, was guilty of this for a long time.

Practically, it is a huge challenge. I would want my day to day fob, an onsite backup, and an offsite backup. That’s a lot of hassle and potential for mistakes. To even register the offsite backup means I need access to it. Remotely copying a password database is so much reliable


If you use a security key only for the most important accounts like Apple or Google, keys are set up once and then unchanged for years.

This is why I've always been a fan of these. They are easy for laypeople to understand.

Yep, I've never had Google, PayPal, or Amazon ask me about a passkey with a yubikey.

The current ones are passkeys

Laws are threats. The difference is that we empower the government to threaten us to maintain social order, we do not empower individuals to threaten each other.

"Nice business you've got here, shame if something happened to it"

I think that issuing an ultimatum where if the person does not do what you want, that you will be sharing your concerns with likeminded people online, with the underlying implication being that the business owner will end up experiencing some kind of negative outcome from that, is absolutely a threat.

Exactly. Pure cancel culture. Just be humble to understand the world might not agree with your views and let it be. Less "this can't exist" and more of "this is not for me".

It's not cancel culture to bad-mouth a business and refuse them your patronage.

People are vocal about what they feel passionate about, and a world where they weren't or couldn't sounds hellish to me. How would society ever progress? We'd still have "whites only" restaurants in that world.

Not saying that justifies all anger targeted at business owners, but a blanket "criticism bad" stance is a bit silly. You have to make an actual argument as to why you think the anti-AI crowd is misguided in this case.


Cancel culture or not, people are allowed to tell other people they know that something is happening, and those people can decide for themselves to patronize the business or not. To call it a threat is really exaggerating.

The headline makes it sound like people are phoning in bomb threats and violence, which doesn't seem to be happening at all.

Honestly, I think the business owner should have ignored these weak "threats" altogether. It doesn't have anything to do with coffee and it's unlikely they'll get any business because they use AI, so it can only hurt them.


I get you think the word is too heavy but remember: it's a small business and bad press is a serious thing. It could be enough to shut down the place. Just see how afraid local shop owners are of Yelp. You get review bombed there and you have a serious problem. You don't know if these artists won't just unload their (just) frustration at AI on the business exactly like that. By judging the tone "change or I'll tell the artists", it's very likely the person would try to incite them instead of just "telling" them.

If the Montgomery Bus Boycott happened today most Americans would call it cancel culture.

Okay, but I (and I suspect a lot of people) really don't care if somebody is threatened in that way. If the business person doesn't want other people to know what they're doing then maybe they shouldn't do it. It's not worthy of a news article, IMO.

If somebody threatens them with violence (which is what "threat" is usually implying) then that's a different thing entirely.


If I tell someone that I am going to take a coordinated effort to undermine their business and get customers to stop shopping there, potentially resulting in the demise of the business, the loss of their livelihood and ability to provide for themselves or their family, I cannot think of any reasonable person who would not categorize that as a threat. A threat by no means is limited to violence only.

I am really disturbed by language minimizes concrete actions taken to cause harm.

"I am not hurting your business I am just talking to my friends!"


Sorry, but the business owner has to own their actions. If they have a large customer base of artists who will feel betrayed by their using AI, then it was a mistake to use it.

It's her right to use AI, but the customers also have the right to say they won't patronize businesses who use AI, and are free to tell people they know that the business is using it.

If the shop's oat milk was really cow's milk, would it be "threatening their business" to tell other customers that?


If the "business owner needs to own their actions", shouldn't the people who are seeking to take a coordinated effort to harm that business also "own their actions" and be able to say plainly what they are doing instead of misrepresenting it via "Stop being crazy, I'm not making threats, I'm just talking to my friends" gaslighting nonsense?

> If the shop's oat milk was really cow's milk, would it be "threatening their business" to tell other customers that?

To entertain this analogy, if a shop is misrepresenting their ingredients then it can make sense to spread the word, particularly if the shop owner refuses to rectify the issue, because there are health implications which represent a legitimate material harm to their customer base and the safety issue takes priority over the shop owner's bottom line.

An AI menu, assuming it offers a reasonable representation of what the shop offers, has no safety issue.


AI stealing their jobs isn't a "legitimate material harm"?

It doesn't matter anyway, because "using AI" doesn't make somebody a protected class. If people don't want to patronize businesses that use AI, there's nothing wrong with that.


> AI stealing their jobs isn't a "legitimate material harm"?

No, if you have to make several indirect causal leaps, to get from "shop owner uses ChatGPT instead of MS Word to generate a menu" to "somewhere a graphic designer loses their employment" it is not legitimate. And if that is how you interpret the world, virtually any harmful or even violent action can be rationalized in the name of stopping some probabilistic harm.

Sticking with your cow's milk analogy, serving the wrong milk to the wrong customer with a dairy allergy can cause immediate harm to their health. Generating an AI menu cannot be reasonably interpreted as causing direct harm to anyone.

> If people don't want to patronize businesses that use AI, there's nothing wrong with that.

I completely agree. But making a personal choice to not visit a business and undertaking a coordinated effort to organize a boycott against a business is in a different category altogether.


Did you not read the article? She was unaware of the controversy and had just had a baby - which tends to be quite a stressful and time consuming thing.

The milk comparison is specious and irrelevant, because consumers are there for the product. Coffee shops sell coffee, not jpegs.

But aside from that - bullying and threatening a working mother with a newborn is not the moral high ground you seem to believe it is.

She's not a witch, and your assumed right to burn her is not a good look.


"It would be a pity if I were to tarnish jlarocco's reputation within my network. I even have his GitHub from his profile. I'll just talk shit about him to all the CTOs on the block and ensure he doesn't get a job again..."

See how that works? That's why you should care. The so-called influencer-orienter culture has become a plague after social media. And these F&B businesses are the ones hit hardest - remember all those movies in the 90s and 00s about how some up and coming entrepreneur or chef had to win the approval of some hoity-toity food journalist?

Also, rest assured, I'm not serious about badmouthing you or anyone here for that matter - just trying to illustrate a point.


Oh, I get it, but it only works if the "influencer" can actually influence a large enough group of customers, and my bet is that they can't.

If it were me, I would have called their bluff and ignored them.


In that analogy, you SHOULD assume the person is an influencer.

In the case of this shop, we're talking about local retail. If you are a retail small business owner you have to assume that any customer can be an "influencer" and cause significant harm if they seek to trash your business online.


To be fair to the news, there probably wasn't much else happening in Buffalo NY

I think there are undeniable and novel risks with AI. But there are risks in everything. We could theorize all day about them. We all use tools every day that could be used for very ugly acts. An iPhone can be used to plan a terrorist attack. A 3D printer can be used to print weapons. We could sit here and fear monger all day. Time will be needed to understand, weigh and contextualize these risks.

What is a more immediate concern for me is the efforts of frontier labs to have government craft legislation that shields them from competition, cements them as monopolies, or assigns them the capability to be moral arbiters in our society.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: