We had no idea that 20% of our revenue for the year came from illegal sales to China! And neither did the president!

Same here, somewhat. No token limits, but we haven’t outsourced review or QA yet so at least the code is being verified. But there’s talk of an end to end AI pipeline, so we’ll see how that goes. If that becomes the standard, probably time for me to jump ship.

I am in software and I don’t think it’s that clear cut tbh. There’s definitely a right and wrong way to use AI. I think a lot of companies are moving fast and breaking things right now, but a lot of them are going to find out that the thing they broke is the product and they won’t be able to slop their way out of a problem they slopped their way into.

At a micro scale, AI is pretty good at writing workable code quickly. This obviously boosts productivity at the individual level, especially in the short term. At a macro scale, good maintainable software requires higher order thinking that transformer based models simply cannot do. So a lot of the stochastic micro decisions an AI makes lead to technical debt that is not trivial to remove, and can’t really be resolved by the AI itself, at least not in a way that won’t contribute to the problem.

These issues can be curbed by the programer at the point of origin, before it gets out of hand, but obviously that requires more judicious use which cuts into the productivity gains. Because writing code has never been the hard part of software.

TL;DR: fully embracing the slop machine will eventually poison your product in a way that is not really easily recoverable. It is useful, but not “too good to not use.”

Also not the person you responded to but in my experience, yes. I don’t tell it to “save it in its memory files” because I don’t think that’s a thing, but including “do not actually edit any files, only propose changes” somewhere in the initial prompt of a conversation has worked well for me.

That’s cool but some guy who read a lot of Harry Potter fan fiction about it says we’re all gonna die. /s in case it’s needed lol.

The hardware requirements are substantial. If companies or individuals want to run models locally, especially at the speed you’d expect from the big name models, then everyone needs a pretty serious GPU on their work station. That or the company needs to build and maintain a server capable of running whatever loads they expect to see.

None of this is impossible of course. But you pay for the tokens one way or another. Either you pay the upfront cost of the hardware and continuing cost of the electricity to run it, or you pay with the inconvenience of running a model very slowly on lower end hardware, or you pay a subscription for someone else to worry about all that.

This is why local models are still niche. Only hobbyists nerds are willing to go through the trouble of setting up and running a local model. Most companies still are not sure if AI is actually worth it for them. So it’s much easier to pay for tokens than the giant upfront cost required to build you own infrastructure.

If the numbers the AI companies are putting up as far as infrastructure cost vs revenue are anything to go by, it probably isn’t worth it to most companies.

Many cars have the modem / LTE as a module that can be disabled. On my car, a 2018 Chevy, the LTE antenna plugs into a networking module in the footwell. I bought an antenna cable and dummy load resistor on Amazon for ~$10 and simply unplugged the LTE antenna in the car and plugged in the dummy antenna. That disables all cell network connectivity but doesn’t effect things like Bluetooth. It’s also easily reversible if someone wanted to do that for some reason.

You can probably Google something like “year make model disable LTE” or “disable telemetry” to get an idea of how hard it would be on your car. Some make it very cheap / easy and some make it very difficult.

I don’t think the person you’re responding to is making any claims about AGI. But the fact is, the models are far, far more capable then they were even 3 years ago. I think that has more to do with the harnesses we have now that it does with the models. The models are a bit better (and way bigger) but the harness is where the capability comes from. Software dev has changed significantly in most places because the fact is, this shit is useful for those tasks. People, appropriately or not, are trying to apply that to every office job basically.

I don’t think that’s gonna pan out for anthropic or open AI financially, especially in the long term. But there has been a lot of progress in capability in the last 2-3 years. I think they’re right that if the timing works out badly, many employers could overcommit to AI and then collapse before they can correct once the bubble pops.

I’m running a 4070 super on my setup with cachyos with zero issues. I never had to do any special setup or anything. I think when I first installed, I was running off the displayport on my motherboard to be safe, but if I remember correctly, the OS detected my Nvidia card and installed the driver automatically. (That or I selected that I wanted Nvidia drivers during the install, I can’t remember which.) Switched the cable over to the card, no issues since.

Everyone complains about Nvidia support on Linux but it’s come a long long way the last few years and it’s honestly just as easy or easier than windows now.

Coming back around, this is your original claim.

that the whole 3D printers are dangerous weapon manufacturing sources is BS. 3D printers, at least those affordable to hobbyists are a damn poor choice for creating fire weapons. Yes, you can print some non-critical parts but that’s about it.

And now we’re here.

it is a fairly easy method and can enable cost effective low number production, yes it can. So yes, if you will, then it has made it more accessible for those things.

So I think I’ve made my point.

The functional core components of a Glock are not made from organic polymers, for a reason.

Neither are the core functional parts on a 3D printed glock. They use the exact same pressure bearing components, because those components are easily available. What’s less available is a frame, and 3D printing solves that problem.

It is really like with all those other tools that can produce such a frame just as well.

What other tools? What other tools are available that can produce a glock frame of similar quality and with similar ease as a decent quality 3D printer for a similar price? (~$300)

But then, if you say most people print it with PLA you might have a strong point that most people 3d printing that don’t know what they are doing or don’t care about PLA’s serious limitations for that application, and do it anyway.

They know perfectly well what they doing and that PLA is a perfectly viable material for the application. Glock frames are not pressure bearing parts, and PLA is plenty strong. I’m saying this with confidence because it is not a hypothetical. People have built these guns and shot thousands of rounds through them without issue.

I wouldn’t consider using a 3D printed jig … and ECM as a part being produced by 3d printing. A helper part is for producing it by other means is 3d printed. Again, plenty of alternative methods to do so.

You don’t have to consider it, it’s been done and it works. Making rifled barrels is not trivial. This method is certainly possible without 3D printing but is made much easier with 3D printing. What is another method of producing a rifled 9x19 barrel at home that’s easier than 3D printing a jig, and running some chemicals through hydraulic tubing with a fishtank pump in a bucket? Please link an example.

Are they all going to be restricted?

I don’t know. I’m not advocating for restrictions. People have been making guns at home in the United States since before the founding of the country and they will continue to do so whether it’s legal or not, regardless of the availability of 3D printers. Even trying to regulate the printers themselves is assinine, there’s too many open source designs and software. Anyone sufficient motivated will get one. Trying to regulate printers in the name of gun violence is a none solution to problem that barely exists.

All I’m saying is that 3D printing has absolutely made the home manufacturing of firearms more accessible than it has ever been, and those firearms are perfectly viable weapons.

Well, you asked for it lol.

you could easily show me a source with the number of victims of 3d printed fire weapons, legally owned commercial fire weapons and illegally owned commercial fire weapons, in California or the US as a whole.

For starters, no one in the US is doing this. There is no source for the number of victims of 3D printed firearms anywhere, the best you’re going to get is from the limited places where authorities are tracking the number of crimes involving “ghost guns”, which could be a 3D printed firearm, a home manufactured firearm by other means, or (most commonly) a legally purchased firearm which has had the serial number removed. Everytown (puke) has some statistics claiming 208 murders or shootings linked to a ghost gun and 134 “recoveries” of ghost guns, nationally since 2013. As you can see, that’s not a lot. My point is, you’re right that ghost guns aren’t really a big problem, but that’s not because they aren’t viable as firearms. I think this is a moot point anyway, there are over 400 million guns in the US and millions more every year. The cat is so far out of the bag that it returned with a litter of kittens which are also out of the bag. Trying to control gun crime in the US by limiting supply is futile, but it looks good on a political platform so it’s gonna keep happening.

But that’s not even really what I was commenting on. The idea that FDM printing is not viable for home manufacturing firearms is at least 10 years out of date. First, it’s important to understand that in the US, the “critical” part is the one that’s legally considered the firearm. Every design has a serialized part which is legally “the gun” and which you (generally) must pass a background check to obtain. Everything else to construct a firearm is available online shipped directly to your door or over the counter without any kind of background check, at least in the vast majority of states. Even in the places where that’s not legal (like California has recently done), it’s very easy to for anyone trying to dodge that law to do so.

As an example, one of the most commonly 3D printed firearms is a Glock. With glocks, the frame is the serialized part. It’s already made of polymer. So most people 3D print a frame, generally from PLA, and buy the rest of the parts to assemble the firearm. This is what Luigi (allegedly) popped a CEO with. In his case, he also 3D printed a suppressor which rendered his gun no longer semi auto, which would happen with a factory glock too, but generally these 3D printed glocks are nearly as viable as a “real” one for at least a few thousand rounds. I’m not gonna bother linking directly to anything here because again, most commonly 3D printed gun. Put “3D printed glock” into youtube.

For a more “real” scenario, you can take a look at Myanmar, where rebel forces have been 3D printing firearms to fight the Junta directly, in combat. They’ve mostly used the FGC-9, which, while not 100% 3D printed, does not require any “real” firearm parts to produce. Everything you need can either be 3D printed or acquired at a hardware store. You can even produce a rifled barrel at home (PDF warning) using a 3D printed jig and ECM.These 3D printed guns have mostly been a means to acquire actual military firearms from the Junta itself, but it worked, so these guns have been successfully employed by teenagers against a state military force.

For some other examples, here is the Orca (3D printed AR-15), the Plastikov (3D printed AKM), the mac n cheese (3D printed mac 10) This one is extra fun because you can get a set of “wind chimes” to complete your build. There’s also the NylAUG, a surplus AUG parts kit which can be purchased over the counter with a receiver 3D printed in Nylon. Hell, the US army 3D printed a 40MM grenade launcher nearly a decade ago, and there are DYI versions too. There are many, many more designs. I think I’ve made my point.

Maybe spend 5 minutes looking into it before commenting on something you know nothing about.

My understanding is that the lithium itself isn’t the issue, it’s that lithium batteries require other rare earths like cobalt where as sodium itself is not only more common than lithium, but it uses more common material like iron or tin in its battery chemistry that are also less problematic.

Did you find a good alternative to fusion?