In the second case, I've provided instructions on how to destroy the data, said "don't do this", and then someone has done it anyway. _They_ have destroyed their data, and it's now up to them to justify why that's my fault.

If we want to get into legal territory about it, which I'm sure we're both woefully underqualified to comment on... the CFAA is all worded around "intentionally accesses a protected computer...". How exactly do you show intent to access a protected computer here? The developer never took any sort of positive step to access _any_ specific computer. The best I could see would be negligence where a reasonable person would have to have known someone would run this on a protected computer, but that still feels like something a good lawyer would find a way out of.

That said, if you pitched me something like a Jupyter notebook style doc where tests validating the claims of the documentation were inline, I'd totally buy into that.

End of the day, though, there's a huge, obvious difference between "asking for help" and "asking for someone to do all my thinking for me".

As a _person_, I'm very sympathetic to why that might be happening. I will do everything I can to help. And sometimes it does feel like I'm bordering on practicing psychology without a license.

As someone responsible for making sure _everyone_ is getting paid this month so they can keep a roof over their head and their kids fed, I do need to be mindful of and address issues that are dragging the entire team down. Regardless of why it's happening, if we're in a situation where you are doing the more company more good by _not showing up to work_ (you're contributing negatively), we have a problem and it needs to be addressed. We can work together on addressing it, but we can't ignore it.

For whatever it's worth, every single person on the team I manage says one of the things they love most about working here is how helpful and cooperative everyone is. Everybody's always happy to hop on a call and work through stuff together and really has a mindset of a rising tide lifting all boats--people are always volunteering to pitch in and help others before being asked. I like to think I've had some part in creating that environment. I am _more_ than happy to _help_. I had to starting making a distinction between "helping" and "doing someone's work for them" because I was getting burned out from overwork. I made it "formal" because I work with the kind of people who really appreciate clear rules and guidelines for things, including communication

You need to put as much effort into the question as you expect someone to put into the answer.

It's not "fairness" or "AI" or anything else, it's that doing this any other way fundamentally fucks up the team dynamics.

You have a problem. You want someone's help. If the cost to you is effectively nil (or negative, since you're asking someone to do your job for you), but the cost to the other person is non-zero, then incentives aren't lining up here. Pretty quickly that person is going to start carrying too much load and become a bottleneck.

It can also mask that the context of the work is too concentrated in one person, and does little to nothing to help build that elsewhere in the team.

The other end of this is exactly what you're saying--put as much effort into the answer as they put into the question. You're not doing anyone a service by taking their low effort input and giving them high effort output, least of all yourself. If someone asks "how do I X", that's low effort. If you happen to know the answer off the top of your head, spare a few sentences to explain or point them where in the code they need to be. If you don't know, don't go track it down for them.

Or maybe this is just a totally different product?

I'd also call out, anecdotally, of the people in my life the non-technical people are interested in touch screens, don't care about speed as long as it runs a few Chrome tabs without feeling slow, and have literally never mentioned noise except to complain about some absolutely absurd "gaming" laptops. I've only ever heard the "nerds" talking about this stuff you're saying actually matters to the non-nerds. Maybe you're one of the nerds?

For simplicity I just use 10G LR modules everywhere. A pair of fiber transceivers is $25. Pretty sure last batch of 3M pre-terminated fibre cables I was grabbing were like $3 a piece or something. So we'll round up and call it $30.

I can get a 3M DAC for about $20.

So yeah it's cheaper, but the price isn't _that_ different. I was using DACs in quite a few places (and still am), but in general I've found it easier just using fiber. (For one, I've had a few devices that didn't get along with various DACs but worked just fine with the fibre transceivers.)

That said, are the majority of people actually even _using_ those features? For us we're essentially just using GitLab for git, merge requests, and CI pipelines. A couple places we use the static page hosting. (First thing I do whenever I create a new repository is go into the settings and just uncheck _all_ the boxes.)

All of that core functionality works really well and is more than polished enough from my point of view.

Sure, we could do it like we did radio frequencies! Most of what we use are "High Frequency" and above... Very High Frequency, Ultra High Frequency, Super High Frequency, Extremely High Frequency.

> In my own very humble opinion, it becomes "Large" when it's out of non-specialized hardware. So currently, a model which requires more than 32GB vram is large (as that's roughly where the high-end gaming GPUs cut off).

So the definition shifts over time based on the market availability of RAM? And can also go backwards? I can't really see anyone bothering to look up the state of the GPU market in order to determine correct terminology whenever they want to talk about this stuff (or interpret old comments, or...).

That also decouples the terminology from the actual capabilities which is what people are generally more interested in. GPT-3 was a "large" language model at this present time. However the the seemingly much more capable Gemma 4 was a large language model at the time GPT-3 was in use, but isn't a large language model right now.

I kinda question the arbitrary line drawn here too--32GB VRAM? Where I am that's a ~$5-6k problem. I'm not sure I'd call that a "consumer" product any more than the $20k data center cards regardless of the OEM intent, but we could argue semantics on that one too.

Fundamentally, defining it this way just seems kind of... useless? It's borderline a meaningless modifier already. This just defines it in a way that's so complex to use or interpret that it's just meaningless in a different way.

For what it's worth, I'd vote to use "large" to mean "big enough to be general purpose", more differentiating from the small, specialized models that came before.

> And btw, there is no way you can train a language model on a CPU, even with ddr5, lest you wait a whole week for a single training cycle. Give it a go! I know I did, it's a magnitude away from being feasible.

Yeah, was mostly being silly--tried to allude to that with the "intergenerational project" comment toward the end there.

Though I _did_ try doing some inference on CPU, which is how I found out that these Xeons I have don't implement AVX512. Surprisingly Gemma 4 (2B) was able to spit out a solid 13-14 tok/s! Was expecting more like... 0.13.

And it's paired with 48 processor cores! I mean, they don't even support AVX512 but they can do math!

I could totally train a LLM! Or at least my family could... might need my kid to pick up and carry on the project.

But in all seriousness... you either missed the point, are being needlessly pedantic, or are... wrong?

This is about learning concepts, and the rest of this is mostly moot.

On the pedantic or wrong notes--What is the documented cut-off for a "large" language model? Because GPT-2 was and is described as a "large" language model. It had 1.5B parameters. You can just about get a consumer GPU capable of training that for about $400 these days.

I don't get _only_ this from Mailchimp, but I definitely get quite a bit of this from Mailchimp, Sendgrid, and others. I've marked it spam, reported it to them (no response), and continued to receive the emails.

I can be kind of scatter brained and generally give the benefit of the doubt, but sometimes it's pretty clear that, e.g., I most definitely did not sign up with some accountant in a different country, in a place I've never been to, to receive reminders of tax deadlines that don't apply to me and offers of accounting services I can't use. Or if I somehow did, the signup was deceptive enough that they never received meaningful consent and I'd call it spam anyway.

(And the email they're sending this to is not some easily confused gmail address or a fat finger--it's my own name at my own domain.)

Having valid contact details or an opt out on their sign up form isn't relevant given I never signed up. It's _unsolicited_, _bulk_ email. It's spam.

Also, y'know, your lungs. Deep inside your lungs.

Running tap water in an ultrasonic humidifer's going to spike the particulate pollution (PM1/2.5/10) throughout your entire house by hundreds of ug/m^3. And it seems that children are particularly prone to inhaling this stuff and having it deposited in their lungs (~2x more particles and ~3.5x more mass).

They really shouldn't be used with anything except distilled water. The things should come with a continuity tester that disables them if the water's conductive or something.

We live in a society, etc, etc. I think it's worthwhile, or even _more_ important, to look at how these impact other people.

In some hypothetical deregulated world, I can choose to buy a car without seatbelts, air bags, ABS/TCS, reverse camera, etc and take that risk on.

My neighbour doesn't get to choose whether or not they want to take on the risk of me backing over their child. The other people on the road don't get to choose whether they take on the risk of me losing control of my vehicle and slamming into them.

The value question isn't purely economic. Regulations that force a general societal care and consideration over selfish individual choices have value in _allowing us to have a society_.

These are monitored, things do get noticed[0], and things like this can and have lead to CAs being distrusted.

It's not foolproof, and it's reactive rather than proactive... but in general, this is unlikely to be happening on major sites or at any significant scale.

I'd wholeheartedly recommend people taking some time and reading through the CA Compliance issues on Bugzilla. The entire CA program there, in my opinion, does a fantastic and largely thankless job of keeping this whole thing on the rails. It's one of the few things I can say I had _more_ trust in the more I looked into it.

[0]: https://bugzilla.mozilla.org/show_bug.cgi?id=1934361

Can we stop with this nonsense at any point?

The government can declare an emergency. Certain actions can be taken during an emergency which are outside what is typically allowed or bypass normal processes. The actions are subject to a mandatory judicial review within 60 days. The judicial review happened. The government was found to have acted out of line. It's current working its way through appeal courts.

The way you phrase this is, imo, intentionally implying "the government is ALLOWED to freeze your accounts at will". The reality is more in line with "I can murder someone at will.". Yes, yes I can. Because we don't have precogs and a pre-crime division. That doesn't mean it's allowed or accepted.

Direct your energy at this law. This is _actually_ a huge fucking problem.

Presumably opening this file I'd see some suspicious looking code and a giant bright red block in the middle of it.

I have the benefit that I'm only working in English so "flag anything that's not basic ASCII" is workable. I could see how this could get a bit messy when you _are_ working with other languages and need to differentiate the compound characters and invisible characters and things that _are_ part of your normal use versus these that aren't.

For me they were.

I still remember the first PC I put together for someone with a SSD.

I had a quite beefy machine at the time and it would take 30 seconds or more to boot Windows, and around 45s to fully load Photoshop.

Built this machine someone with entirely low-end (think like "i3" not "Celeron") components, but it was more than enough for what they wanted it for. It would hit the desktop in around 10 seconds, and photoshop was ready to go in about 2 seconds.

(Or thereabouts--I did time it, but I'm remembering numbers from like a decade and a half ago.)

For a _lot_ of operations, the SSD made an order of magnitude difference. Blew my mind at the time.

If you haven't seen it, you might enjoy Benn Jordan's video of "saving" a PNG to a bird:

https://www.youtube.com/watch?v=hCQCP-5g5bo

(Besides the hook-y title, some interesting info on the acoustics of bird songs with some cross-over into tech.)

Depending what precisely you mean by the second one, you may not even need an AM/support for that.

They won't help me use the platform, but they will still address issues with the platform. If you run into bugs, things not behaving how they're documented, or something that simply isn't exposed/available to customers they seem to be pretty good about getting it resolved regardless of your spend or support level.

(On my personal account with minimal spend, no AM, and no support... I've had engineers from the relevant teams email me directly after submitting a ticket for issues.)

So yeah, "if you know what you're doing" you probably don't even need the paid-for support.