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

I like this take. Especially because one of the sibling comments framed Anthropic's stance as "paternalism." Trying to be ethical and to minimize harm, even at great expense to one's finances and reputation, is paternalistic apparently.

I mean, if you take HN commenters to have the thoughtfulness and foresight of children, then the word kind of works.

No — we’ve just taken Ethics 102 as well, so we understand good intentions don’t entail positive outcomes, therefore you may need to criticize or oppose people who state good intentions to bring about good outcomes.

Insulting and demeaning people for that, rather than engaging their arguments in good faith, is a breach of ethics.


To make the discussion constructive, can you give specific reasons (ideally with examples) about why it is so useless for you? How exactly are you using it that you think any output from it can easily be replaced with a Wikipedia search?

The cybersecurity and bioweapons filters reach so far that they set in as soon as the model even glazes anything STEM-related. It might give a good impression of ones ex or write a decent fanfiction but anything that could bring humanity forward is strictly off-limits.

OpenAI is the only real competition. Chinese models are 6-8 months behind Opus 4.8/GPT 5.5, and at least a year or more behind Mythos.

And it doesn't look like OpenAI will have a good answer to Mythos anytime soon. Based on what their chief scientist wrote to staff recently (https://archive.is/fN2pg), GPT 5.6 is a "meaningful improvement" over 5.5 - in other words, just a normal version bump. And no news or even rumors regarding GPT 6.


If the guardrails were so useless, people wouldn't be complaining about them.

People are generally complaining about false positives. Now if you really wanna know what a real criminal organization would do... They'd just buy data center hardware even if it costs 200k because a successful targeted hit could yield far in excess of that. So yes it's speed bump at best.

> it's speed bump at best

To be fair, speed bumps work. If it's actually speed bumping nefarious activity, that gives authorities more time to react.

The correct place to police rogue nucleotides is at the labs. Not the compute layer.


> speed bumps work

Yea. To slow you down. They don't prevent you from getting somewhere.


> To slow you down. They don't prevent you from getting somewhere

Again, yeah. That's how fences work, too. And alarm systems. Pretty much anything that isn't foolproof. Pointing out that a defence is surmountable isn't a rejection of it per se.


what does this mean

Well you see when a daddy H100 and a mommy H100 meet....

They should have designed a guardrail that doesn't make a probabilistic system less reliable. That's hard though. I'm afraid the only way to prevent accessing certain knowledge in a model is not to train it on those materials that enable them.

If we learned anything in the past years of LLM-s is that these guardrails will be jailbroken in no time. I've had some fun time too circumventing them.

Anyone cares about a fable about my grandmother's dream she had in morse code about an alien species signaling her a DNA sequence?


It's entirely reasonable for them to be really annoying to legitimate users while still being useless at their intended purpose. Just look at DRM.

Murder is very (100%!) effective at preventing cancer. And yet, it is a useless method of preventing cancer.

The complain because they get wrongfully triggered

> if you ask it to write secure code, it assumes it is cybersecurity related work instead of software engineering best practices, and you get downgraded.

Will code created this way more or less secure?

And I bet malware developers will find ways to circumvent them.

It’s like those "you wouldn’t steal a car" anti piracy ads that DVD buyers were forced to watch while users of the pirated version could simply watch the film without such useless annoyance


What does that mean? Have you never worked on extremely difficult problems as a side project?

I guess my comment got lost in translation. The project OP linked in his comment is a toy project, not a difficult problem as he led others to believe.

So you could have done it in your sleep, with your hands tied behind your back. Got it.

(You may not realize it but simonw is one of the cofounders of Django, Python's web framework. If they find a Python problem difficult, it probably is.)


Read the log he posted. If this is very difficult, then what would you consider AI, kernel development, computer graphics, etc.?

Web development is not a domain I would consider noteworthy of making a framework given how much development there has been in that area.


Unnecessary based on what exactly? Your vibes?

That’s odd, I used it on a pretty complex refactoring task and it worked for 22 mins and used only 15% of my 5-hour limit. I’m on the $200 Max plan though.

Well the $200 Max plan is 4x the usage quotas of the $100 so it's "within reason"?

Yeah, but 25 days holiday plus bank holidays means you're working like half the year at most. ;)

And don't you knock of at lunch on Fridays anyways? So that's like a 4 day work week, because let's face it, you're not really doing anything on the day you're knocking off early anyways. See you at the pub!

Read-Only-Fridays, and having a pub lunch so you're not doing much all afternoon anyway!

>> It's so you don't have to ask anybody for permission. That's it.

This doesn't make sense because there's one party whose permission you always must ask, and that's the government. They are the ones who get to decide whether you can launch your rockets.

A more accurate version of your claim would be: datacenters in space allow you to deal with one party (i.e. the government) instead of many. So long as your relationship with that one party is good, your business plan is safe.


Fair (at least for now, maybe in a decade we will be manufacturing stuff up there, but for now, yeah).

Totally fair - and with Star Shield and basically SpaceX being the only reasonable launch provider, and a Musk-Friendly government currently in the executive… then I think my thesis holds. The only people who can tell SpaceX no at that point are like 3 nation states with ASAT capabilities?

Regardless, he won’t have to ask the “city” counsel of Asslick Indiana if he can “please build here pretty please!”

Mark my words, they’re going to build “up.l


I dig into problems way, way deeper with AI than without. I can also add a lot more polish to features, add more test coverage, write more documentation, explore multiple approaches rather than go with gut-feel, and so on.

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

Search: