I did just literally say that this has never ever happened to me. Likely that's a regional thing, perhaps panel interviews are popular in USA - but my practice has always been 1-3 rounds of <1 hour interviews, at various sized places including a multinational bank (where I also did many interviews from the hiring side both for technical and nontechnical roles, but it always was ~3 rounds of <1hour interviews with no take-home tasks; if more people were needed, they were included at one of the interview rounds, not as extra time), a major government institution, and a matured tech ex-startup - and while your question was reasonable, IMHO framing your prejudicial assumptions as "how tiny are these places" was an unprovoked insult that's not entirely appropriate.
I've never had such a panel interview in my quite long career and possibly never will. I'm not allocating a whole day for an interview, that job offer isn't likely to be that much better than the current job or other alternatives to justify jumping through such hoops.
If you want me to do work for you and judge me on that, I'm available as a contractor.
How would I know that the task actually takes 30 minutes before scheduling as much time they say the process would take? I presume that when they say "come in for an hour of talk and a three-hour practical task" (which the point where I'll refuse) I won't even know the task and can't already assume that I can actually be done much faster.
A three-hour test is above what I'd consider acceptable for most potential employers. I can talk with you a bit, but if you want me to come in and do work for half a day, I'm going to refuse, I've got actual paid work to do.
But for the question of anti-trust action it doesn't matter whether it's "a lot" of people, it matters whether it's an insignificant number of people - if there were 10 smartwatch users in the US, the argument "who cares about smartwatch users" could be valid, but it makes no difference whether there's 219 million people in the US or 219 thousand people in US, since even 219 thousand users is definitely much, much more than sufficient to justify intervention.
We aren't so starved so as to have to prioritize eating efficiently, people will eagerly prioritize aspects of taste and eat inefficiently, and it's great that they can do that - it would be horrific and a symptom of unacceptable overpopulation if people couldn't afford to eat "inefficiently" and would have to resort at eating whatever as long as it has sufficient nutrients.
One man's "pour your limited resource on the ground" is another man's "valuable use of that resource" - if they're choosing to do something with the energy they purchase, apparently they consider that this is a worthwhile use of that resource. If I'm in the desert and want to pour some water on the ground for growing orchids, that's my choice on how to use it - if it's really important to conserve water, then it will be costly enough so that I won't waste much of it on luxuries, but if it's cheap enough to do so, then people are voting with their wallets that it's NOT really (yet?) so important to conserve it so much as to refuse spending that on luxuries.
But as we don't have mass long-term energy storage, future energy is not exchangeable with current energy, thus the future energy cost shouldn't matter for the current training of LLMs, and whether tomorrow it will be a good use of energy to put funny hats on pictures of cats, it's something that people will be able to choose based on the availability of energy - it's up to them to decide what's essential or not, it's up to them to decide whether they are willing to sacrifice some entertainment just to save energy, and that's what they should be doing with money as the mechanism of measure - if a person wants to do something silly, it would be inappropriate for someone else to decide "no, that's random bullshit" and prohibit them to spend their resources (energy or otherwise) on it. The fact that this decision will affect the price of e.g. energy for others is NOT an externality, it's the core economic mechanism of allocating scarce resources - if some people want to spend the same resources on cat picture and they are fully paying the costs of that, then that's directly reflected in the price they're paying, not an unpriced externality.
It's not a replacement for the cost of operating existing gas and coal plants; building up new solar capacity enables you to burn less gas or coal at the times when the sun is shining and thus cut its fuel costs, but it does not enable you to shut down the plant and cut any its other operating costs, because you still need the same capacity for e.g. winter evenings when at peak daily consumption there is zero sunlight.
There are unaccounted externalities in many areas, but energy use of LLMs isn't one of them - people are paying full market price for the energy of training or using LLMs, that energy cost is large enough so that people care a LOT about it, and make intentional tradeoffs about what is worth it.
And if people vote with their wallets that they are willing to pay for a years worth of energy on training LLMs on cat pictures, we should respect their choice that apparently this is what they consider a valuable use of the limited resources they have.
If your target audience is businesses, not individuals, then you can go a very long way with fully manual onboarding, invoicing, etc. It's different for things like consumer services or e.g. forum users, but why couldn't you manually vet every business your business trades with?
The key aspect is that such memory generally works on a "block" level so making any smaller-than-block write on a SSD requires reading a whole block (which can be quite large), erasing that whole block and then writing back the whole modified block; as you physically can't toggle a bit without erasing the whole block first.
So if large sequential writes mean that you only write full whole blocks, that can be done much faster than writing the same data in random order.
The resulting model doesn't have access to the information about what follows what, so it can recreate paragraphs but can't recreate their proper order for a chapter or book. Well, it can try to guess..
One thing that we did with distributing certain copyright-protected textual material was to scramble them at the paragraph level.
If you take every paragraph in the Harry Potter saga and sort the paragraphs in alphabetical order, it's just as good for training short-context-window models, but not a "harm to the market" leading to a lost sale for anyone who wants to read the books.
Yes, the difference is in what all our tools and infrastructure presume to be more or less sensitive.
Sending a GET request to a site for the password-input screen and POST'ing the password will get very different treatement than sending the same amount of "authorization bits" in the URL; in the first case, your browser won't store the secret in the history, the webserver and reverse proxy won't include it in their logs, various tools won't consider it appropriate to cache, etc, etc.
Our software infrastructure is built on an assumption that URLs aren't really sensitive, not like form content, and so they get far more sloppy treatment in many places.
If the secret URL is short-lived or preferably single-use-only (as e.g. many password reset links) then that's not an issue, but if you want to keep something secret long-term, then using it in an URL means it's very likely to get placed in various places which don't really try to keep things secret.
There definitely is a practical organization need for a way to call out nonsense - but this article is about the widely used workaround process (namely, honest talks during after-work drinking) that does enable this valuable feedback to happen despite the "harmonious interaction culture" in the default environment.
In cultures where direct confrontation isn't accepted, intention does not matter , you simply can't do it - if someone believes that intention is all that matters, speaks their mind directly with e.g. some not-properly-veiled criticism (no matter how justified or needed), and justifies it with "I didn't aim to upset", even if they genuinely didn't aim that and everybody understands that they didn't aim to upset, they have still "been a dick" and behaved socially unacceptably. In the professional world in such cultures it can cause customers to break cont(r)acts, request that your company never ever send you to the customer, or if you speak your mind in public to (or about) the boss, firing you - it's behaving in an unacceptably insulting way that can't and won't be tolerated in the society.
How do you perform repair on something inside a solid steel, titanium or aluminum block, which is what this structure replaces?
These are structural components that don't have anything "inside" but are used as material for carrying/distributing loads, and their strength and weight is the thing that matters. If their integrity is compromised, you generally don't repair structural components if the repaired seam wouldn't have the same strength as the original material, but replace them.
You could have a much simpler parsing than ordinary parsing if you know/assume that you definitely have a valid number from -99.99 to 99.99.
For example, you could find whether it starts with a '-' and where the delimiter is to know the length of the number string representation (that's the "simpler parsing" part), and then don't do any work at all to decode the number, simply use these 1-5 bytes of that string (without sign and separator) directly as an index into a very large very sparse memory region in which all the valid values are pre-populated with the proper result.
You'd need to allocate 4 or 8 terabytes of virtual memory address space for that lookup table, but you'll touch only a tiny fraction of the memory pages, so it doesn't require an unacceptable amount of physical memory.
The wine (and also e.g. scotch) industry runs its core business on not-that-old product, on which it makes almost all of its money, and may keep a relatively tiny niche of really old stuff for sentimental (for the owner), PR or branding purposes. It doesn't work that way for commodities - a forestry or lumberyard selling a few really old oak logs doesn't help them charge a premium for their 20-year old lumber in a way that works for a winery or distillery.
A key reason for this is that we don't live a feudal society divided in serfs and lords by birthright, so most workers who do have the courage to stand up to management can (and in most industries do) simply become managers. So the population of people in abusive jobs isn't representative of the whole society, it's a self-selected subpopulation of people who are willing to tolerate that.
It's article 22 (https://gdpr-info.eu/art-22-gdpr/) "The data subject shall have the right not to be subject to a decision based solely on automated processing, including profiling, which produces legal effects concerning him or her or similarly significantly affects him or her." - with the emphasis on solely, and it does have certain caveats which perhaps maybe might make it permissible.
Like, this restricts automatic refusal of service due to automated profiling, but doesn't restrict automatic acceptance of service due to automated profiling, and it doesn't restrict automatic recommendation to refuse service which then is rapidly 'reviewed' by a human.
You can probably claim they are not answering their phones, but what would that do in small claims court? Do they have some legal duty to answer phones at all, ever? Can't they just say - "oh, if we see that Lio is calling, we would never answer them as a matter of policy, and it's fully within our rights to do so!" ?
They might be required to have a legal address and be reachable by certified mail, but IMHO there's nothing wrong with a business (no matter how large) simply having no phones at all.
In general, the government wants to hear the opinions of various industries or interest groups when discussing things that affect them. Furthermore, if the politicians have decided that they want to implement a certain change, the details of that change need input from that industry or interest group.
For example, some representatives may have a general political position that they want to protect local industry from overseas competition with worse standards. There is a valuable role for a lobbyist to come and say "hey, you have taken measures for the steel industry and honey imports, but what about us, the widget industry as represented by ACME inc. also is relevant your goals!" - and if the representative agrees that this should be done, it makes sense for them to ask the lobbyist / the industry to make an exact proposal to how different types of widgets should be defined/distinguished in that new law, instead of having it be written by some staff paralegal who doesn't understand what the industry is about and what those terms mean.
On the other hand, I feel it has been an enabler - the ability to track their phone if they aren't back home when they're supposed to probably allowed me to let my kids go unsupervised to various activities at least a year or two earlier than I would otherwise have; the phone tracking ability meant that they could go to sports practice at the other side of the city on their own with bikes or public transport (depending on season) instead of being "delivered" and supervised by me or the spouse.
If we have to relearn hunting and gathering, that's not even closely compatible with feeding 10% of current human population, that's consistent with <1%, probably 0.1% or less of current population surviving. Feeding even 0.8 billion people still requires intensive mass agriculture.
The endurance running adaptations seem to start 1.9 million years ago for our common hominid ancestors that significantly predate the split between us and neanderthals which happened more than a million years later.