The deletion discussion also takes time and consensus, and even then isn't irreversible; you can move the text to draftspace, [1] and try again when you do have proof of notability. Even if the article is deleted before you can respond you can always get a deletion review [0] get your article text back via temporary undeletion and move that to draftspace.
Right. I had an article I spent a bunch of time on last year that didn't pass muster for notability (first time I've had this happen) and shoved into draftspace where I got to edit it and retry it (unsuccessfully) a couple of times.
It was an article that really just slipped between the cracks of the rules, probably like the Odin one. All I did was take a breather for a few weeks, come back, edit it judiciously down to its core and resubmit it. It passed.
If Odin is important enough then one day there will be enough relevant citations to make it notable and it will be approved. Take a breather and come back.
Actually, I see the article has already been moved to draftspace, so nothing has been lost, except visibility, for which you will have to establish notability; you have six months to work on this before it gets deleted, and you can always go back for as many tries as you like.
The sad (or wonderful) fact is that anyone can write a new programming language now; I could for example use Claude create a wrapper on top of an existing languge (for example, a Python-like syntax wrapper for Scheme) in a few hours and start marketing it as a new language. So how can Wikipedia tell toy languages from real ones with actual useful adoption?
It's simple; get coverage in reliable sources,[0] and the article can come back. That really doesn't seem to be too high a bar to cross.
What practice? I'd like to see non-trivial uses of this outside scenarios other than this [0] which even now seems to show slow but steady improvements over time.
As for the Machiavellian scheming aspect, this seems to me to be very much like the operations of some real-world human CEOs.
This happens all the time to me. The model emits thousands of lines of high-quality code, making vast progress very quickly, but on the way does a few very, very stupid things a human being would never have been silly enough to do.
Then it takes hours or days - sometimes weeks - to find and fix the AI-induced problems. If you very, very tightly constrain the AI by using structured processes and unit tests you can work wonders with it, but you do start to wonder to what extent this is better than if you had simply coded it yourself.
Tesla's semi-autonomous FSD falls in the uncanny valley between being good enough to lull you into an inattentive state, but not reliable enough to be better than humans with all possible unforseen situations of the sort that cause real-world crashes and either deal with it safely without driver intervention or at least safely slow to a stop without disaster.
You either capture Level 4 autonomy or you don't, and currently Waymo are leading the way in this, not Tesla, who are Level 3 at best - curious, because with the ever-increasing rate of AI competence and their massive head start, you'd expect them to have cracked Level 4 by now.
The prospect of Level 4+ FSD and being #1 in an ever-expanding market without realistic competition were Tesla's USPs and they are in the process of losing both.
I suspect the market knows this and sees the recent better than expected results as a short-term reprieve, not a renewal of the dream.
> Tesla, who are Level 3 at best - curious, because with the ever-increasing rate of AI competence and their massive head start, you'd expect them to have cracked Level 4 by now.
Perhaps (1) Tesla's technology is incapable of level 4 (no LIDAR), or (2) Tesla is not as good at AI as its competitors, or (3) Tesla really truly is a droid company and the car thing was never a priority, or (4) the CEO is on drugs and should not be allowed to wave chainsaws around.
Even asking a human being why they did a certain thing is questionable. The research on choice blindness seems like a pretty definitive debunking of post-hoc rationalization:
I'm not sure what point you're trying to make. In science and engineering, being able to provide justification is a core skill. The comparison we should be making is against the human practitioners who are trained in their fields. There will always be a distribution of ability. Saying that there's evidence that people are capable of providing post-hoc rationalization doesn't say anything about the ability of experts to produce well thought out responses (in their respective fields) that don't immediately fall apart under scrutiny.
Structured thinking and deliberation are indeed important, but you can also make LLMs do structured "thinking" if you work hard enough, and generate quite plausible reasoned arguments with valid real-world results, and you can get them to write down their working as they go. But as research has shown, it's not "true" thinking, just pattern matching at a higher level, and eventually runs out of steam.[0]
But you only have to drill down a couple more layers and you are back in the void again; do you have any proof that your own thinking, no matter how structured and accurate, is anything other than pattern-matching at a sufficiently much higher level at which you are incapable of seeing it as such?
I think we will be finding some very interesting things out soon using the combination of LLMs and theorem provers, as demonstrated by Terence Tao's recent work.[1]
A cheetah is not a motorbike is not an aircraft is not a rocket.
I think a better way of putting this is that humans think they can accurately re-tell what their consciousness was doing. Whether they actually can, or even if consciousness exists at all as a thing outside the perception of consciousness is a philosophical question currently beyond answering.
It has vastly increased my hobby programming "velocity", but has improved my day job performance by perhaps a quarter at most; so much in programming is not the task of programming, but mapping the problem and dealing with the outside world.
DMCA notices are meant to be submitted "under penalty of perjury", and false notices could in theory result in civil action being taken against those who send them. In practice, neither of these occur even if the sender is a real person, like a record company lawyer lending their name to complaint that are entirely computer generated, or, in this as in so many cases, a completely fabricated identity.
Requiring verification through government ID for takedown notices should be a minimum requirement.
We got told by the 9th circuit that false claims submitted under penalty of perjury are actually just "opinions" not capable of being proven true or false.
Yes, it's "I swear under penalty of perjury that it's my opinion that this infringes my copyright."
Creator's rights need to be safeguarded but the DMCA gives legal weight to people without legal training and when they fuck it up (accidentally or intentionally) they get the no-consequence "but it's just, like, your opinion, man, and no-one expects you to be a lawyer".
In this case the statement they checked off said that there was an unlawful reproduction of a trademark. There was no reproduction at all, since the products were originally manufactured by the brand. Boggles my mind how that can be an opinion.
In this case, the perjury may apply. It seems unlikely that the person that complained is acting on behalf of whoever owns the rights to a 25 year old article.
I'd be curious if any prosecutions are a real deterrent - it seems not. YouTube has to follow the DMCA but also enforces its stricter content ID, with popular creators getting hit (and being vocal) and YouTube seeming to "fix" the issues (until next time).
Ultimately the whole system needs reform now where it's easier than ever via LLMs to send off these notices.
AFAIU, the whole deal is that the bogus claims never actually reach the DMCA stage - big platforms implement their moderation policies and copyright claim handling specifically to avoid involving the legal system. It's that intermediate layer that incentivizes automated, bogus claims, as there's effectively zero consequences to them.
[0] https://en.wikipedia.org/wiki/Wikipedia:Drafts
[1] https://en.wikipedia.org/wiki/Wikipedia:Deletion_review