ailef

They are just a bunch of floats, and they are usually stored in a binary non-human-readable format.

> Part of the job especially when there are either tight deadlines or bad bugs

It's fine if this happens occasionally on deadlines or unexpected problems.

But it definitely shouldn't be part of any (good) job unless your contract explicitly provides for this possibility (with extra compensation).

Hi HN!

What I'm showing here is my latest open-source project, which might be of interest to Java/Spring Boot developers. I know they are a minority here, so I don't expect to get much traction with this post. However, even if you don't fall in this category and want to give me some feedback just about the website landing page it would be awesome.

Cheers, Ailef

Java + Spring Boot. Old school server-side rendered templates with Thymeleaf if the site is static, Vuejs frontend instead if something more dynamic is needed.

Alternatively, Flask/Jinja/Vuejs if I'm working in Python.

Have you tried setting a very low temperature, possibly zero? I've had good success with it.

Or the ones canceled by the US, like the ABM and the INF treaties with Bush and Trump.

How do you create reusable styles this way?

Indeed. I believe it's even stricter than that; you actually have to request a quota increase that has to be approved in order to spend more than 120$, the default spend hard limit.

I'm really baffled by this. For example in the US, it's enough to just associate something with Russia for everyone to immediately drop any kind of critical thinking. You don't even have to provide compelling evidence for it. The association alone (real or made up) is already enough to completely corrupt the subject's image. This trick has been played dozens of times but somehow people still fall for it...

Can somebody explain what are the practical implications of this?

This is indeed the supported way to do it.

I would but I didn't use an actual prompt. I fine-tuned a model on n-gate's content, so my prompts just contain information like the submission title and url.

I just have a few blog posts about for fun side projects which maybe you might consider interesting. I've been lucky to make HN home page a few times, actually at rank #1 once!

https://ailef.tech

Some time ago I was banned as well, even though I don't remember exactly why but it was a similarly petty reason.

What solved it for me was telling that my account was hacked in the support ticket. If this was an automated ban, maybe shifting the blame to the "hackers" might get your account back if you're lucky.

This is really interesting. How would this compare in terms of performance/resources necessary for training/inference w.r.t. neural networks?

Would this be leaner and run on less or would it reach the same complexity eventually?

My blog is at https://ailef.tech

I write mostly about my for fun personal side projects. I don't have that many posts but I'm proud of having hit the HN front page a couple of times!

That's true but humans are, or at least should be, accountable. When the AIs start making/contributing to impactful decision it will be hard to determine who to consider accountable for them (if something bad happens as a result).

I've shown you that your own source is highlighting these problems, but you chose to deliberately ignore them. It's really hard to continue this conversation assuming you are replying in good faith, so I'm not going to.

All I see is very vague wording and metaphors which are very far from a very specific *military* objective that is conquering the entirety of Ukraine.

They might mean that or might not, but it's not explicitly stated. You are choosing one possible interpretation and sticking to it like it's the only one.

The problem with all of this is you have to correctly assess their "expected bias". And this assessment is, in turn, affected by your own bias which you have not considered in your analysis. Given this, it seems easy to fall into the trap of just accepting whatever already fits your narrative and discarding the rest.

I tried to read the article using Google Translate but I found no evidence of this statement (capturing all of Ukraine). Could you point me to the relevant passage?

Please note that this is different from statements like "bring Ukraine in the Russian sphere of influence" or similar things. Unless you wanna conflate military objectives with geopolitical ones.

Again, this is not relevant. It seems extremely biased to reject his opinion just on the basis on who quoted him. Noam Chomsky is frequently quoted in Russian propaganda outlets. Should we disregard everything he says?

By they way, you conveniently glossed over all the other points, where even the same person *you* quoted admitted the huge problems with the current narrative.

So Russian media is now reliable when what they say fits your own purpose?

> If you want some rigorous examination of the events, ones that aren't described by academics as' "not academic", "chaotic" and "politically driven"'

Why not quote the full sentence?

> Katchanovski's original paper was criticised by David R. Marples. Marples called it "not academic", "chaotic" and "politically driven", but also acknowledged that Katchanovski "has raised some new evidence that suggests new investigations into the sniper massacres are much needed," that "the official version of events is indeed deeply troublesome and his gathering of new material is commendable," and that "his paper does provide evidence that there were several separate groups of snipers, including anti-government ones."

I'd like to particularly emphasize:

> the official version of events is indeed deeply troublesome

> there were several separate groups of snipers, including anti-government ones."

The fact that he is quoted by Russian media is irrelevant. If that is enough for you to automatically discredit it, that says more about you and your unwillingness to examine alternative versions of this story rather than the story itself.

BTW, keep in my mind that I never said this alternative version is the truth. I'm just noticing that it exists and that the official version has too many holes to just drop sentences like this:

> He fails to mention that the protestors were killed by sniper fire by Yanukovych

> thats exactly what he's trying to do right now.

And how could you possibly know that? The article highlights how this is most likely not the final goal of the Russian aggression against Ukraine and provides logical and rational reasons as to why.

Look at the reactions in the US to the story that came up just a few weeks ago about a possible Chinese "spy base" in Cuba. Cuba being a sovereign state and all, I'll let you draw the parallelism yourself...

> He fails to mention that the protestors were killed by sniper fire by Yanukovych

There's no conclusive proof that this is what happened. No official investigation has taken place (or has been concluded) and Ukrainian authorities have shown time and again that there was no political will to really discover how the events unfolded. (Ukrainian authorities have also been "reprimanded" for this lack of investigation by the EU or UNHCR, can't remember now and I don't have time to look for a source).

There are several alternative reconstructions of the events, for example the one by Ivan Katchanovski[1] which has also published several hours of original footage on his YouTube channel[2]. I would suggest anyone with an open mind to take a look at those.

I would be really wary of making unsubstantiated claims like those on events that are this controversial.

[1] https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2658245

[2] https://www.youtube.com/@IvanKatchanovskiPhD

ChatGPT is really and at chess. I'd just paste it in one of the online chess engines.

> FWIW, I think everyone should stop worrying about nuclear war. The US would win hands down and everyone knows it.

This line of reasoning is really dangerous and I'm flabbergasted that there are people who take this stance. Really unfortunate that not just randoms on HN but likely also people in Washington might believe this BS.

There doesn't need to be any conspiracy. If it's good for PR reasons to show they are fighting, companies will do it anyway without the need to coordinate.