There are two layers to that.

The first is how to develop skills. And you do that the exact same way everyone before you did it: you actually do the work. Calculators are awesome but you still learn how to do long division and the like because it gives you insight into how to approximate things. Same with sims/solvers versus actually solving PDEs.

The other is… if your boss wants you to feed everything into an LLM then you won’t have a job much longer. So you can either look for a new one or work toward more advanced tickets/tasks. Make it clear that LLMs have limitations and that some stuff will need a proper coder and that YOU are that proper coder.

But also AI cannot currently do everything, so you need someone to fill those areas.

And who is going to be able to fill those gaps? Probably not the person who “knows what I want to achieve but (…) don’t know how to actually implement it”.

Which ties in to

their capability to learn, their personality, will they mesh well with the existing team, have they got drive to make things better, do they have soft skills to position themselves to become better, is the person adaptable

is the bar for what is considered fundamental shifting?

If the bar is “I know how to ask a magic box to do my job for me” then there is genuinely no need for previous training and experience and a company won’t be hiring engineers or spreadsheet gandalfs or marketing experts. They’ll hire the cheapest “prompt engineer” they can, underpay them, and then replace them the moment they ask for a cost of living increase.

And… the companies considering that really aren’t the ones with any longevity. Yes, yes, any port in a storm. But they will RAPIDLY run into that wall and have no way to move past it. Whether that is getting the senior engineer in cargo shorts to do it or curating training data to improve the model.

but as time went on we got new levels of coding and so knowing how to write low level code is no longer a required skill.

And that is another barrier that MANY companies have run into.

The average coder? Yeah, they don’t need to understand how to optimize a loop. But when there are forty tools on the market that all just call pytorch? The one company that knows how to optimize a critical path function suddenly looks REALLY good with their 10% performance (and thus power) savings.

Again, these tools are incredibly powerful and I regularly use chatgpt et al to generate a first draft of a utility script. And I’ve been using editor plugins for… sweet Eothas over two decades now, to generate docstring stubs and even a lot of unit tests. And people SHOULD know how and when to use these tools.

But you also have to consider what you can get out of it. “AI” generated documentation is pretty much worthless outside of checking off a box that you have documented every function in the code. Your LLM won’t understand what that function was trying to achieve or why “it is wrong but that is because this library is wrong” and so forth. Any documentation that is actually meant to be referenced still needs a proper pass from whoever drew the short straw in Engineering.

Same with testing. AI can generate tautologies. AI won’t stress test your code because it doesn’t know what you think that code might do in the future. By all means, generate the boilerplate, but you are still going to be the one who has to go in and add that really weird corner case that TOTALLY didn’t break prod lats month.

And… you know who historically did those tasks? Interns and junior engineers. The same ones who are adamant that their entire job can be done by chatgpt and lamenting that they don’t know how to move from idea to implementation. And guess how you learn how to do that?

I have a rough idea of what I want to achieve and some steps on the way there, but don’t know how to actually implement it.

That is literally what the job is. If you can’t do that then you aren’t an engineer.

I’m concerned that there are skills I am missing out on developing, but at the same time if AI is being pushed so heavily is it not something I should lean into to be better equipped in working with it?

I’ll tell you what I told my nephew: Yes, everyone is going to use AI to one degree or another. So why would I hire you over anyone else? Or, more pointedly, why would I hire someone at all?

Getting to that interview gets harder and harder every year (every month, really). But engineers (and even many managers) can immediately tell someone who knows their shit versus someone who “vibe codes” all the “hard parts”.

These days it more or less explicitly refers to asking an LLM to write your code for you based on prompts.

But on a broader spectrum it is just the idea of (I forget the buzz word) Ticket Driven Development. A manager defines software based on a series of (jira, gitlab, kanban, whatever) tickets/issues and someone below them (in this case, an LLM) implements it.

Done properly? It is incredibly effective as it allows designers and “idea people” to work to their strengths and junior developers to work to theirs. The problem being that, much like when it is a junior dev under them, the person making the tickets likely has no idea what they are doing.

Which is the big problem. Someone who has been writing scripts for decades? Using chatgpt to get the syntax of a function or even to write a utility script is great. They can focus their brainpower on the harder/more fun stuff. Someone who has been writing code for, at most, a year or two? They never learn those foundations and never have a way to do anything the LLM can’t (or verify if the LLM is correct).

You uh… don’t need to tell people stuff like that.

I mean… I have seen some REALLY REALLY stupid drivers so I could totally see multiple people thinking they found a short cut or not realizing the road they are supposed to be on is 20 feet to the left and there is a reason their phone is losing its shit all while their suspension is getting destroyed.

But yeah. It is the standard tesla corp MO. They detect a dangerous situation and disable all the “self driving”. Obviously because it is up to the driver to handle it and not because they want the legal protection to say it wasn’t their fault.

As if I weren’t already on enough watch lists:

Consumer drones aren’t conducive to “a 9-11”. They just can’t carry enough explosives. A terrorist attack using drones will be isolated attacks on public gatherings and likely not even put a dent in the M-4 tax we pay every fucking day.

Also: Defense against drones actually isn’t all that hard from a military and hardpoint standpoint and mostly uses existing tech. The same mic arrays that installations use to triangulate gunfire? Picking up drone motors is of comparable difficulty and gives a pretty fast angle to point the baby phalanx (or net guns) at. Would be HORRIBLE in an area dense with civilians but would protect anything that “we” actually care about (military resources and whichever politicians trump likes this week). Chaining the net gun/tiny gatling gun to the sensors with a human in the loop so it doesn’t violate any treaties covers that.

The reason they are so ridiculously effective in Ukraine are because Russians have horrifically bad opsec and tend to leave armored vehicle hatches open (even at bases) and have sprawling unfortified trench lines that are only designed to keep infantry out.

I think the basic issue is that you seem to have the maximalist goal of wanting PeerTube to be at similar numbers to YouTube, and missing that, you don’t see anyone finding worth in it - so I will argue against that point, hoping I did not misread you there.

No, my point is that there is no reason for a Content Creator to actually put effort into Peertube (or even the youtube alternatives that they aren’t co-owners of). And there is no path toward that.

For the same reason many thousands of people can download 4k movies and shows concurrently via torrents with just a few dedicated seedboxes

People run seedboxes because they get something out of it: Private tracker access.

You apparently are happy to donate your bandwidth. Good for you. I suspect you would think otherwise when a video “goes viral” and you suddenly get a call from your ISP telling you they have decided you are hosting a business and that you need to pay for a different internet plan. PLENTY of people learned this the fun way back in the 90s/00s.

The real costly part is storage, bandwidth is surprisingly affordable considering the project we are talking about

In the context of file hosting, “bandwidth” is usually used as a catch all for both the raw bandwidth of a single node AND the requirements of a CDN. Otherwise that server gets hugged to death, everyone is angry, blah blah blah. Again, this is a lesson we learned in the 00s.

he influx in audience a “big creator” could mean to an instance, as well as with it potential support in donations both directly and indirecty, could very well outweigh the addional costs

Ah, so now content creators are working specifically to support Peertube. Which means their effective operating costs have just skyrocketed because now they are paying for their own hosting AND paying for all the time and materials to make the video in the first place.

Which is WHY Youtube became so massive. Paying for your own hosting is REALLY expensive and tears into already thin margins for the vast majority of Content Creators.

I’m assuming I missed a valid scandal here that led to closing of an instance, but - Lemmy (and PieFed, and mBin) is very much alive and we are discussing on i

This board is on the dot world instance. Back when Luigi allegedly popped that guy, dot world was leading the charge in terms of complete nonsense CYAs to protect the instance from getting a knock from the FBI equivalent. Similarly, I have my account on the dot zip instance and we have very weird rules regarding the UK because of their data privacy laws.

Or, to be an old again: Every even semi-public FTP server goes the same way. Everything is great. Then you suddenly realize that porn has appeared out of nowhere. And, best case scenario, it is the kind of porn that gets people arrested (because it can get SO much worse…). The same issue came up with file sharing back in the day. And it is why a lot of lemmy/mastodon instances have very specific rules regarding image hosting and NSFW content.

It is great you have had a good experience. Others won’t and will rapidly realize just what they are doinating their bandwidth to.

And audience-wise, it already is a fitting niche for people you disregarded

Great. I didn’t “disregard” anyone. But you have to understand that You Don’t Matter. Because all those things that you (and I) hate about modern video content? That is done to make money.

Google are constantly running analytics to figure out what length of video is most profitable for them in terms of storage, bandwidth, and monetizability. Content Creators are constantly figuring out what Google wants to get a video promoted to the front page AND what will make people not only click that video on the front page but also keep watching so that they can get the metrics that get them those sponsorships.

And its great that you and the people who like Peertube don’t care about that. But that gets back to the same exact point I have been making the entire time: if creating content for a platform can’t even meaningfully offset the cost of creating that content in the first place, the VAST majority of people won’t and you are basically left with the independently wealthy people.

Already mentioned it elsewhere in the thread but this is a story as old as history itself. Creating art costs money. Canvas and paint costs money. Having the time to stay another day to look at that landscape for just a bit longer costs money.

Having a good camera and mic costs money. Having the newest fanciest cell phone with the auto stabilizer so you don’t need a rig costs money. Having the time to learn and use that video editing software costs money. Having the time to keep your head clear enough to really do a solid edit costs money.

And same with the Maker Youtube style content or even the DIY content. Having that second CO2 canister to do another run? Or even just having a piece of wood and a clamp so you can show how that dishwasher airgap actually works without needing to try to angle a camera into a tight corner.

Hell… being able to Go To The Zoo on a less popular day? Guess what that costs?

And taht is the fundamental issue. Yes, there will be people who make content and some of it will be genuinely awesome. But there is also a pretty massive ceiling that will basically mean only independently wealthy people have the time and resources to do a “good video”. And… we kind of actually saw that in the early days of youtube where a LOT of the old hats are from rich families or have cash from their startup being bought and so forth.

What peertube (and basically all the youtube alternatives) lack is any way to move on from that.

To reiterate and expand: The life cycle of a successful youtuber (or twitch streamer or whatever) is:

1. Create content at severe personal cost
2. Qualify for ad revenue. Offset some of that cost but still require a day job. Gauge popularity based on ad revenue
3. Qualify for a referral link for online retailers. Actually make genuinely good money that can potentially lead to this becoming a full time job. Well, not so much after the honey nonsense is likely going to make all this go away but…
4. Qualify for sponsorships at decent rates. THIS is where things can reasonably become a full time job and you can start making true Art rather than fitting a build in between your normal job
5. Get popular enough that you can get enough of a following that people actually WILL “just put some money in the tip jar”

Peertube et al only really exist starting on step 4 (because you can bet most instance owners would strip or hijack those referral links…). And… no company people will want to deal with is going to be sponsoring content that goes to a fraction of the audience that is also predisposed to consider any form of marketing or capitalism a personal insult.

Which basically leaves 5. Which… we are already seeing. linus media group LOVES to pretend Floatplane (which might actually use Peertube under the hood, I forget) is some mega successful business. But he still does youtube content front and center and Floatplane mostly exists for him to profit off a few other youtubers and to encourage his rabid fanbase to give him more money. And there will likely be other creators who decide to “support Peertube” while still primarily making Youtube content. And… they will just crash the ecosystem while getting nerd/FOSS cred.

Yes. Discoverability is the real key. Which also is not at all addressed on peertube and, as I mentioned above, mostly still comes from youtube for the creators who have branched out to other platforms.

A Michael Reeves can get away with just having a kofi and making massive bank because of how big he has gotten… from youtube. Whereas even a Not An Engineer has a channel that lives or dies by collabs and shoutouts from other youtubers (I do suspect he has an independent source of income though).

Also… you put even Not An Engineer on your peertube instance and he is going to consume a disproportionate amount of bandwidth. Let alone a Michael Reeves who would crash the entire fediverse during his annual video.

What you are describing is “if you build it, they will come”. Which is patently false. Ad revenue gets worse and worse every year but it usually is essential to even offsetting parts and labor for a video for smaller creators. I think it was Gamers Nexus that discussed the different tiers of monetization in the context of the honey scandal, but the basic idea is that ads are what let you know if a channel has any legs and referral links are what keep you alive until you are big enough for a sponsor to care.

Which… is also the issue. The kind of sponsors who would fund a peertube video (and just look around at how fediverse folk view ANY form of monetization of the content they consume…) are going to be more bluechew than not, if you catch my drift. And they aren’t going to pay much.

Which gets back to: Peertube as a concept is great for official tutorials and MAYBE blog posts by “nobodies”. Why would anyone go out of their way to join in decentralized hosting of that? And while it is conceptually a great way to “can’t stop the signal” an important video… it either rapidly becomes liveleaks or we see the same thing that happened with Lemmy where the instance owners get a phone call from their local FBI equivalent and rapidly say “I don’t want that smoke”.

But Peertube as something people would even want to browse or create Content for? I have yet to see any path toward that that isn’t “Well, people really love the ideologies of FOSS so they’ll do it out of the goodness of their heart”

And good for you.

But it is the fundamental issue with art. Art needs funding whether that is a patron or a tip jar or whatever. Canvas and paint costs money. Being able to spend an extra afternoon looking at a landscape costs money. Having time to edit and revise your script costs money. And so forth.

And same for youtube. It is the difference between being able to “get it right” versus just using the footage you have and “making it work”. Hell, I love videos that explain how to do basic repairs that I should have learned long ago. And having the money to set up a free standing pipe or cut a bathtub spout in half works a lot better than trying to hold a camera to record yourself replacing a dishwasher airgap and not actually capturing where the clips were.

The issue with “Algorithms” is that you need a lot of data to generate recommendations… which tends to mean centralized (or a LOT of data scraping).

Which is why stuff like Nebula and Floatplane and Ian McCollum’s latest “apolitical” side hustle all are still 100% dependent on Youtube. Hell… I actually still subscribe to Legal Eagle and NileRed on Youtube so that I know when they have a new Nebula video.

And just to be clear: None of the above (including Peertube in general) is competing with Youtube. Well, Floatplane says it is but that is because linus sebastien is a dipshit con man. But it still speaks to the fundamental issue of where Content comes from and what it takes to have the time and resources to do a “good” video.

The problem is that it is just a fundamentally un-profitable platform for creators. Ads don’t work (… period but also) because of the decentralized nature of it. Any instance/frontend that added ads would be shunned in favor of those who don’t. And any video hosted only on a single “instance” would rapidly cost way too much if it ever became moderately popular.

Which means there is no reason for Content Creators to… care. So the best it can ever get is “early youtube”. And people SAY they want early youtube videos but everyone is deeply spoiled by the difference between a video that was made in a week of after work tweaking versus weeks of full time planning and editing.

Which is why peertube in general is one of the “fediverse” products that… I feel really weird about. I forget if Floatplane/Weapons of Guntube/whatever use it or something they rolled themselves, but this really feels like the kind of software project that has the end state of getting “adopted” by a corporation and the major devs hired on as consultants.

Like, twitter (mastodon), reddit (lemmy), and even instagram (??) make sense to me and are very conducive to self hosting since… they are message boards and that is how we used to roll. But video is expensive and hard AND needs incentives to create “good” content for it.

And if it sticks, good. But it still has the fundamental problem of needing to re-train all your existing employees AND train new staff who haven’t been brought up in that system.

Its on a completely different scale, but plenty of tech youtubers have done the “Let’s get rid of all the Adobe in my life”. Some succeed. Most tend to come down on some variation of “I can do about 99% of what I used to do in these two or three tools. And these ten things are actually genuinely easier and more performant. But we can’t take a month off making videos to get all of our editors up to speed. And this also removes our ability to contract out an edit to someone with the industry standard workflow”. And from my professional experience in different fields, that is true. Hiring someone and then spending a week or a month so they can use YOUR tools becomes a huge burden in not too long of a time.

I really hope Germany pulls it off this time and more governments follow. But I also remember all the other times I have read this story.

I want to say various cities/regions in Germany make statements like this every few years? And they usually end up rolling back when it becomes clear the cost to retrain both existing staff and new staff isn’t worth it.

That said: This gets the national security bump so maybe it will stick. Also nobody on the planet likes to use Teams.

Which… is basically worthless because of just how many cameras there are out there.

A “fun” exercise a couple buddies and I did a few years (… decade?) back was to just use an afternoon of plugging python packages together and scraping county traffic cam feeds to track someone, with their consent, over a few days. And it was ridiculously easy to get their schedule down basically day one and even get a LOT of data on who they were seeing or where they went after parking just based on when and where the car “disappeared”.

And that is just publicly available traffic cameras. Not the giant mess of speed and red light cameras and all the other crap we have in a modern surveillance state.

So even if people are climbing traffic poles and midlining over to the actual boxes to smash them? Those are even less of an issue than normal outages from rain on a windy day.

Bingo. Just make sure you are masked up and know WHERE you are masking up.

Which is not an “ai poison pill”

While they’re at it, why not just hack the government to reverse last year’s election, amirite?

I know most of us loved Mr Robot and watching dinozzo and abby double team a keyboard and Wolverine getting a blowy and all that fun stuff, but that really isn’t how things work.

These aren’t off the shelf pre-trained models. The model is a big part of the company’s product and, increasingly, the cost of training is being partially offloaded to customers under the guise of “tune the model to your data”.

And IF we have a Bones situation where someone has inscribed a virus onto human remains to destroy a one of a kind machine or whatever: That is what version control is for. “Hmm. The May 2025 model isn’t working. Okay, switch back to April”

Also, these “models” are a lot closer to just running OCR on a feed and logging which traffic camera saw one of the flagged license plates.

deleted by creator

nVidia cares less and less about the consumer market every year. We basically only exist to buy the factory fourths so that the overall yield of any given wafer can be maximized.

2400 for a single component is still rather insane even by server room standards. But 12 or even 18 load balanced? That starts to “make sense” for higher end data centers or even on-prem server rooms at the more tech oriented companies.