Nobody in this space gives a fuck about anyone outside of the people paying for their top-tier services, and even then, they only care about them when their bill is due. They don't care about their regular users, don't care about the environment, don't care about the people that actually made the "data" they're re-selling... nobody.
It's not. The discipline of statistics knows a lot about population sizes, sample sizes, margins of error, and that sort of thing. If you're going to seriously question the fundamental ideas in statistics, there's a whole lot of empirical evidence you need to refute.
> Should an advertiser be able to claim the same immunity?
Have you ever purchased anything that was marketed using statistics? Lots of things, I imagine. How many polls have you answered about those statistics? I'm not really sure how you'd get the impression that anything else was ever the case.
That goes without saying in the software business today. I was in software for decades and I’ve never seen it so cynical. Shameless profiteering seems to be the gold standard strategy. It’s like Gordon Gecko style greed.
It’s somewhat subjective, but I disagree it’s easier to fleece photographers than audiophiles. There are professional art photographers that use Leica cameras because they’re great, and $250 is pocket change for a lot of serious optical equipment. Look at the Canon L lenses and the like. Lots of people that buy that stuff don’t need it, but it’s not expensive solely for the sake of being expensive.
I have yet to find a professional sound engineer, producer, or artist that calls themself an audiophile or uses the insanely overpriced gear marketed to them. Lots of that stuff is demonstrably bullshit and only valuable because it’s expensive.
I’ve had too many bad sales experiences to deal with that. The second someone tries to force me into a sales call for a non-customized or self-configurable service or product, I assume they’re just shamelessly setting me up to extract as much money from me as they possibly can. I just can’t assume good faith on the part of a company that only distributes product information through someone making a commission. It feels like they’re inviting me into a mouse trap.
I don’t think qubes is targeted at mainstream users— even mainstream developers— and I don’t think something has to be targeted at the mainstream to be interesting, especially here. There are probably a lot of people here that won’t ever use it that will still find the idea and ethos interesting.
Nope. There’s a lot— a lot— written on artistic process and intention. What art is and what it means is a deep philosophical topic that has been pulled apart for millennia. The sheer arrogance of nearly the entire tech world reducing the value of the artistic process to a few simple logical business friendly ideas using a few a priori thought experiments while disregarding the existing philosophy on the matter is the sort of thing making the tech industry so deeply unpopular right now.
Well, in web dev, people vacuuming up all of your code to feed a NN creates something that eliminates the tedious annoying parts and leaves the more interesting work. Conversely, even in the technical end of art, people vacuuming up all of your output to feed a NN eliminates the interesting, fulfilling parts and for professional tasks, leaves the existing tedious annoying parts, and even adds a few more, while devaluing the entire skillset. And then everyone in tech pretends they know more about your job than you do, and calls you a jerk for not being happy that SV companies killed your job market by offering C- results for F prices without the inconvenience of any of the people that created the initial “data” getting paychecks. So, considering the pros and cons, I think you made the right choice.
It seems to me that for the most capable and useful models, openness almost exclusively benefits businesses, or maybe academic organizations with money for serious hardware. I know what I can run on my 4090 at home but the results pale in comparison to the commercial services. I see why people consider these matters important from a theoretical standpoint but from a practical standpoint it doesn’t seem particularly consequential. I self-host a few FOSS server applications that are primarily sold as SaaS subscriptions, and folks are often very critical of those businesses benefitting from the “open source” label because they’re often seemingly deliberately difficult to self-host. This seems to be an order of magnitude less open than that. Is there some use case for people with reasonable hardware that I’m just not aware of?
And at that level of expense, having no insurance, bad regular insurance, and catastrophic insurance is not much different. However, if you get into a car accident, have diabetes, eat a burger with E. coli oh157 which destroys your kidneys and you need dialysis to stay alive, get appendicitis, get HIV, get cancer, or possibly get exposed to rabies and need human rabies immune globulin, having no insurance is very very very very very different. If you have bad insurance you haven’t used that year, it will cost you $500 or whatever your deductible is. More than that with catastrophic insurance. When my wife and I were exposed to rabies during a fantastic stay at a super clean and well-run air bnb and needed the shots, the insurance company portion of our bill was $38,000.
Blanket black-and-white statements about what will or won’t ‘solve’ gun violence can only exist in oversimplified worlds that conflate causes, catalysts, prerequisites, cultural influences, economics, etc. If anybody is expecting one sphere of concern— e.g socioeconomic equality, bans, mental health treatment, draconian sentences for violent crimes, etc.— to essentially resolve the issue to the exclusion of all others, then nothing will ever look like it’s going to work because it’s not one problem. It’s not even one symptom. Street gang violence, criminal cartel violence, and mass killings by extremists are serious problems with very different root causes that express themselves in very different ways. It’s also totally disingenuous to say their having a common tool is inconsequential, or that the same approach to addressing that makes sense everywhere, or that taking away the rights of people with different values is an inconsequential concern. You can’t solve all problems at once, and dismissing one approach out-of-hand because you can cite problems it won’t solve and assert all sorts of assumed negative effects without any evidence is nothing more than avoidance.
Nope. EMTALA means hospitals have to evaluate, and if necessary, stabilize you, and that’s it. If you’ve got a growing cancer or some expensive chronic disease that’s causing serious problems, they’ll do what they can to stop those problems from killing you in any given visit, but they absolutely will not treat cancer with the intent of curing it or help you control a chronic disease knowing they’ll never recoup payment. They’ll give you a prescription for insulin but they’re certainly not going to fill it for you, no matter how big of an emergency it is or how much the hospital visit cost. Even then, they still bill you for all of the services they do render and medical bills still bankrupt people. Needing to make sure someone isn’t going to die in the next few hours is very different from taking the financial responsibility for treating a very expensive medical problem.
I have found copilot autocomplete to be somewhat useful for small blocks of code.
Coca-Cola and Toys R Us have found them useful for making terrible commercials cheaper than making terrible commercials by hand and way cheaper than making good commercials that actually improve their brand image. Seems weird they’d do that for immensely expensive holiday television spots rather than throwaway 5 second YouTube spots or something but hey — I’m clearly not a corporate genius.
This the end game that Silicon Valley created. An automation arms race between two competing groups that were initially trying to save a little time or cut down on staffing but escalated it to the point where the default approach would be considered unforgivably assholish 15 years ago, people that don’t buy into it at least somewhat are drowning in bullshit, and nobody’s happy— but on paper everybody’s got record productivity!
With LLMs, this same exact scenario is playing out in other realms. Look at writing and publishing. Sure you’re on top of the world before everyone else catches up, but when they do, there’s now just a boilerplate of exponentially expanding bullshit and counter-bullshit that everyone has to circumvent to do anything.
It’s not even like a language — it’s several languages spread across half a dozen purposes -built environments. I’m definitely interested to see where apex goes, but it’s going to need a lot of work in both the workflow and how they communicate its value before it gets any real adoption outside of some tech artists working on super complex skeletons and rigs. As of now, everything I’ve seen has been essentially a coding demo with visuals. Working with Advanced Skeleton in Maya is just so elegant and capable— between that and industry inertia, I have a hard time imagining it’s going to be the standard anytime soon. I think the real sleeper here is UE. You can model, make skeletons, rig, paint skin weights all right in there with a pretty smooth UI. I think they’re doing to lap everyone in like a decade.
Oh and I forgot to mention that it’s all glued together with python, which replaced their previous non-compiled scripting language based on csh. (No shit. Yes, that csh. I do believe the first version of Houdini’s precursor came out in the late 80s. Wasn’t tcsh the application that spawned the turn-of-phrase ‘considered harmful?’) You can do pretty much anything with Python— not just glue work and making procedural parameters. But from what I gather, it’s not nearly as performant as vex or VOPs for doing per-particle or per-voxel operations when you’ve got more than a few million. I’ve never pushed Houdini’s Python that far, though, so that might be old news. Vex and VOPs are stricter on type, but are still garbage collected, etc so I don’t know where the efficiency lies.
Houdini really is a different thing. Its procedural-only focus is all-encompassing, resource intensive, and difficult to combine with other paradigms. People that talk about Blender overtaking Houdini as a professional tool have no clue what they’re talking about. It’s not that it can’t, it shouldn’t because they’ve got totally different focuses that make for a pretty awkward fit. Blenders sims are making great strides but to shoot for feature parity would make blender a worse program. So much unnecessary complexity if you don’t need that niche toolkit.
I know more people that love Maya then love Max, which is funny because IMO Max is much better for modeling. Maya, however, really is great for spline-based animation, generally, but specifically character animation. Blender has been making big jumps there though. The reason I’m glad that I learned the big propriety clients in school— Houdini, Maya, Max, Zbrush, Nuke, Mari, etc.— is because it’s a much more marketable skill for big studios, and much more difficult to get that experience yourself. Our program’s tooling was entirely focused on getting students into the big studio career pipeline be it in vfx, animation, tech art, game design, etc. I guarantee that students would have come out of that program better artists, fundamentally, had they learned how to do all of that stuff in Blender. Given my career ambitions, I’m not sad I got what I got though.
I’m setting aside UE because Blender isn’t and was never trying to be a game engine. When it comes to high-end physics sims of any kind— particles, pyro, rbd, liquid, cloth, hair, etc– blender isn’t close to Houdini. The blender toolkit will get ‘something’, but the Houdini toolkit will get you ‘anything.’ More broadly, it’s so easy (well, “easy” like you don’t have to modify the program source. This stuff is going to be complex and fussy no matter what) to infinitely control absolutely anything using everything from node-based no-code stuff (VOPs)that compiles down to performant machine code, to their built-in c-like for performant calculations in complex networks (vex), to popping in nodes where you can write your own openCL if you really wanted to, all animatable using anything from key frames to their built-in non-time-bound math system (CHOPs… I know… the names…) originally built to process wave-level audio. If you’re clever enough, you can even use all of that to drive your geometry, audio and shaders to some extent simultaneously, without (textual) code. The geometry nodes in blender are cool, and they seem to be developing them the right way – ensuring quality rather than racing for feature parity with bifrost or Houdini or wherever — but last I checked you couldn’t even animate their parameters. Huge progress, a great tool, and many people’s preferred tool for modeling. Frankly you wouldn’t catch me dead sculpting in Houdini, and Houdini has a way to go before their compositor is on par with Blender’s, even if the procedural bent makes it more useful for some unusual use cases now. But for the really deep procedural stuff and simulations, anything involving volumes, and also fitting into complex pipelines, we’re talking etch-a-sketch vs oil paints. Imagine if blender started as a purely procedural tool and only focused on that use case? The differences would go so much deeper than the feature list would let on.
I think the differences are a lot less significant with other modeling-focused DCCs like Maya, Max, and C4D. All great programs in their own right. Maya really is an incredible tool for character animation, C4D is so killer for motion graphics type stuff… but they’re all much closer to blender for their intended use cases.