2. They might get into trouble charging users to generate some other entity's IP, so they may revenue-share with the IP owner.
They're probably still losing money even if they charge for video generation, but recouping some of that cost, even if they revshare, is better than nothing.
You got the last paragraph wrong. They need to negotiate with rights holders on the revenue split. They’re hoping that the virality aspect will be more important to rights holders than money alone, but they will of course also give money to rights holders.
Or, in other words: here’s Sam Altman saying to Disney “you should actually be grateful if people generate tons of videos with Disney characters because it puts them front and center again.”, but then he acknowledges that OpenAI also benefits from it and therefore should pay Disney something. But this will be his argument when negotiating for a lower revenue share, and if his theory holds, then brands that don’t enter into a revenue share with OpenAI because they don’t like the deal terms may lose out on even more money and attention that they would get via Sora.
The voice is really good, especially small details like sharp air intake before a response to indicate tone. I guess the conversation topics weren't amazing but that's not really the point.
The problem I have is that if you want to engage with something like this, you need to pretend it's a human. As they say, the uncanny valley was pretty much successfully crossed. And to be honest, I don't want to pretend I'm talking to a human for a whole bunch of reasons.
The technically aspects are really impressive, but I think "pretending to be a human" in this way is a pretty scary goal. The cognitive dissonance was too strong and it was hard for me to continue a conversation very long at all. What does it even mean to have a conversation without theory of mind?
People talk a lot about how transformative AI will be but rarely from the point of view it will have on society when interactions like this are 10x more realistic and ubiquitous.
I imagine an agent with this kind of conversational capability but better. And then I imagine that coupled with a video model that presents a proper talking head that is pretty much indistinguishable from a real one
Then what if these personalities were highly configurable to your tastes, had a very long memory and were so cheap and available you could build an entire customised social circle from them.
This is a bizarre nightmare for us but will become the norm for future generations. As rigid and inflexible as it is, my kids even had a hard time understanding Alexa is not real at first, and they have come accustomed to this style of conversational interface from a young age, I think about that but in 30 years for their kids.
If I listen to robot voices for to long it gets unbearable. The fascinating part imho is that human voices apparently have other communication channels that bypass the conscious mind. Like regional accents contain a kind of collective mood but more granular. It must therefore be possible to encode a beneficial feeling into the robot voices. We could all be Irish :)
The good-faith interpretation of Musk's actions goes something like: "All of these cuts are consistent with the overall goal and eventually planned. So, Musk is just cutting things as he thinks of them."
Unfortunately, it means that the best way to avoid getting cut, for now, is just to avoid drawing Musk's attention, for any reason. The ISS thing from yesterday [1] is a clear example - something that was already planned and has just been bumped up.
To be clear, I don't really buy this interpretation, especially since over the years Musk has shown himself to be quite vindictive.
I don't want this to be a defense of elon musk. Everything he is currently doing is unequivocally wrong and he does not deserve the benefit of the doubt.
That said, there have been multiple cases where 'unconstitutional' actions had good faith interpretations. The obvious one being the Underground Railroad. Less obvious was Lincoln suspending habius corpus in 1861 (I don't know if the term executive order existed back then, but this sure felt like one). It would be another two years before Habeas Corpus Suspension Act of 1863 was passed.
Overused? Sure, beyond belief. But dropping a "The Constitution is not a suicide pact" feels particularly appropriate right about now.
It is interesting, though, that a portion of the article is dedicated to discussing how it's important to make the proof simpler and more generally available. Although machine-verifiability is surely a good goal, it can't replace humans being able to comprehend the proofs as well. In this world where proof can all be verified, we may be more confident in our mathematical knowledge, but we'll also be unable to generate new knowledge.
There are a ton of networks (think big corporate networks, schools, shared apartment wifi) that enforce too many weird port restrictions. Many of those places rarely get network or config updates. I don't think it's as bad as IPv6, but there are a lot of people for whom it isn't going to just work out of the box.
Gnome 3 has been out for 9 years now, which is longer than the time between the releases of Gnome 2 and 3. I don't think there is a lot of config churn. (It may have been worse in the early days of Gnome 3.)
> "In my personal opinion, I can deliver the same quality of education online as I could in person."
Although this statement is debatable to begin with, it also misses a huge point. Even if the quality of an online education is the same, many students cannot learn as anywhere near as well in that environment. They paid tens of thousands of dollars for in-person learning, so for that to be replaced with something entirely different in form really sucks.
As the article stated, many of these universities did give partial refunds. The students are demanding even more refunds while still collecting the education.
I’m curious: Would you consider the same argument valid for workers forced to work remotely? Would you be upset if a company decided to cut workers’ pay in half during COVID19 remote work under the assumption that some workers can’t be as efficient remotely? After all, the company paid hundreds of thousands of dollars for in-person work, but now they’re getting potentially worse remote work.
> As the article stated, many of these universities did give partial refunds. The students are demanding even more refunds while still collecting the education.
To be clear, the article doesn't state that partial refunds of tuition have been offered, but that "many did offer partial refunds of dorm and activities fees."
Developer experience is more than just technical too. I know some package authors (and myself personally) find themselves very disillusioned with the GNU approach to contribution, especially regarding licensing and copyright assignment.
> very disillusioned with the GNU approach to contribution, especially regarding licensing and copyright assignment.
TBF, given the historical FUD baggage, I don't fault them for trying to play extra safe as stewards of the project. Also, Emacs makes it so easy to get packages out-of-band (eg: MELPA, Borg, Straight, Quelpa) that I don't think the copyright assignment is a big deal unless one wants to get code merged into base Emacs.
I submitted a miniscule patch to tcl-mode (or font-lock) a couple decades ago. It was to speed up syntax highlighting for a particular style of commenting. Maybe a 3-6 line patch, tops.
I spent 20x more time back and forth on copyright assignments, including getting a release from my company, etc to get the patch in. I pushed through because I felt like I was always "just one more yak shave away from finishing", but if I knew at the start how much time it was going to take, I'd have kept the patch on our own private site-lisp. That's a problem, IMO.
The copyright release process where I work is pretty liberal, but the amount of people I have to get involved is fairly prohibitive. If your boss doesn't like doing paperwork, you're not going to get a release...
My memory is that my correspondence was directly with rms on this patch and it was required at the time, but the fact that it would no longer required is definitely beneficial, so thanks for that update!
a lot of contributors/developers are bothered by the fact that they have to play ball with copyright laws, at all.
it's a fact of life for developers with projects anywhere near the GNU-scape that if you don't GNU it, you'll catch a lot of hatred, even worse if you choose to avoid licensing all together -- and gods help you if you choose a tongue-in-cheek licensing agreement like WTFPL.
at the end of the day a lot of people just want to contribute meaningfully to a project that they use and enjoy, but the headache of licensing and catching flak by choosing the wrong one (and since all the communities have opposing thoughts, they're all the wrong one to certain folks), it just becomes easy to 'forget to contribute' -- especially when your patch or whatever is working fine locally and there is little practical incentive to catch that much heat.
I think the legalese issues turns a lot of would-be contributors into local-patcher type developers, and then they leave for greener pastures once what they needed patched is on their own machine -- especially for projects like emacs where 90 percent of development is going to be towards extensions.
...and I say all this from a position of love and admiration for GNU and the FSF.
A cheeky license, or, worse, no license at all is a potential hole for a lawsuit. That is, of time and money wasted, and of the project being at risk.
I see why you might go with Apache, or MIT, or even straight public domain licenses. But as a maintainer, I would not accept a contribution which is not properly licensed, or which is licensed in a way not compatible with the project's license. Usually such a contribution is less valuable than the rest of the project, so there's no point to introduce a real legal risk of project's closure for the sake of such a contribution.
I can relate. Used to be a hardcore GNU fan, but these days I'm probably closer to sqlite thinking (public domain, but work hard to avoid contamination).
As for contributions, presumably GNU wants ownership, but do they have a problem with assigning back what amounts to public domain rights to the author?
And I suppose, for most contributions, does it really matter? The awful truth is that I can't think of anything I've ever written that had freestanding value, as opposed to value as an enhancement to something else.
Yeah I completely agree. It's like, I agree with GNU and the FSF, but that doesn't mean I'm as strict as they are. I want to contribute to MIT, Apache, and BSD licensed projects as well, or I want to just not think about it and work on some project that has no licensing info whatsoever.
But (some definitely not all) folks are pretty all or nothing. There's a lot of jerk developers on the net though, maybe it's better to just ignore them?
> A hammer does not have [an agenda], and that does not make it a less useful tool.
Sure, but by default all software has a copyright agenda built in. We can remove copyright, but there are two approaches:
- Remove copyright so that anyone can use your code, but then they can re-add copyright to your code and sue other people for violating their copyright.
- Remove copyright so that anyone can use your code and ensure that nobody can re-add copyright to your code.
I used to be in the former camp, but I've slowly moved to the latter.
It's kinda why I like the Mozilla license (MPL). Anyone can freely use (leech even) on the project, but if they make improvements or changes to the project they have to share. It also doesn't try to control how the consumer of the project is allowed to deploy such software either.
There's a world of difference between releasing contributions under the same copyright (i.e. if you're contributing to a GPL'd project, release your code under the same GPL license) and assigning copyright to the FSF. I think that's the turn-off for many. It isn't a matter of 'what would you even do with' the contribution, it's a matter of recognition and making it much harder to change the licensing terms down the road.
For those who just want to bury their head in the sand and pretend copyright doesn't exist, they will be the first to complain when the code that they wrote is taken private and commercialized (i.e. look at the licenses this has been an issue for)... making code 'public domain' allows for that.
What happens when the woke takeover of the SFS is finally complete and emacs is re-licensed under a new social justice license where you can only use it if you don't oppress minorities?
I don't trust the FSF further than I can throw them and I don't trust them with my copyrights.
Agreed... that was my point. Specifically, it doesn't matter who takes over or what the changes are, only that direction/priorities may change and be inconsistent with the intent of the original contributor.
The existing code cannot just be "re-licensed". It's true that they might decide to make future versions available only under some weird license, but that wouldn't cause prior contributions to become unavailable. Continuing the code line under the original license would always be possible, given effort.
2. They might get into trouble charging users to generate some other entity's IP, so they may revenue-share with the IP owner.
They're probably still losing money even if they charge for video generation, but recouping some of that cost, even if they revshare, is better than nothing.