Re: A Ghastly Thought
"It just came to me in a burst of insight that AI could replace ... lawyers."
Unsuccessfully so far.
33045 publicly visible posts • joined 16 Jun 2014
The relevant issue here is that ChatGPT is providing a service which is based on what is essentially a database which is derivative of the training materials. The regurgitation is simply a means of providing proof of that. The value of ChatGPT and hence of OpenAI depends in part on their software for which they've (presumably) paid their developers but also on the training material for which they might or might not have paid but the use of which will exceed the T&Cs attached to it.
"When did people forget that the stock market is effectively gambling?"
What's more when you buy a share you're becoming part of the company. The company is called that because "company" is a collective noun, in this case the collection of people who each own a share of the company. The clues are all there in the name. So if you're a shareholder the company's funds are your collective funds. If you are a shreholder and sue the company you're suing yourself along with all the other shareholders. If you think the C-suite or directors are gypping you, sue them, not yourself. If you're prompted to sue yourself by a lawyer remember it's the guaranteed winner who's prompting you and if you want to find out who's the loser, look in the mirror.
If you look closely at it it mounts to /boot/efi but I'd expect at boot time UEFI is able to locate it directly. It has to be mounted when the system's running because it needs to be accessible for upgrades.. It contains one directory called EFI - upper case because we're in VFAT-land. Then it gets really weird. It has directories called BOOT, debian, devuan and ubuntu. Fair enough, if we're in VFAT country it looks like BOOT is going to be the one read by UEFI. Devuan, obviously because we're running Devuan, debian because devuan's largely based on on Debian. But ubuntu?!? They all have a BOOTX64.CSV file which is data, not a CSV (actually it's alternating ASCII and null bytes) and several stripped Windows executables.
Why Windows executables and file names that look as itf they come from DOS? Because whatever you run on your UEFI enabled box, Microsoft has it's sticky fingers in there. If the competition regulators really take a look at Microsoft UEFI should be on the agenda.
But the reason you shouldn't be modifying it it that you'll very likely render your box unbootable pending a re-install.
When small companies have few or one product it's quite likely that they were designed by enthusiasts who actually understand what's needed because it's something they actually want themselves. When they get bought out by big companies the products are specified by marketing who believe they know what the market wants with no interest in using the product themselves..
So they can be sold to telcos to remotely manage them.
When PlusNet decided to remote manage mine to the point where I was unable to get in to make changes to my DHCP allocations it was time to pull it and install my own. Do I want to join the manufacturer's remote management scheme? No I do not. Turn off remote management.
"If it was even a little bit important, nobody would have gone down the Intel / 80x86 branch, but would have stuck with the "cleaner" addressing modes of the 68000. Yay! let's revive the architecture battles of 40 years ago."
The early SMB-scale Unix computers I used were either Z8000 or 68000 based. That wasn't, at least directly, because of cleanliness of addressing modes, it's because, at system integrator level, it's what was available to buy.
But because IBM chose Intel for their PC and the economies of scale prevailed all that stuff became a dead end. So much for not throwing away backwards compatibiliy.
Whether you call something a directory or a folder it's still an abstraction of a place in which to store other things, some of which may also be files or folders depending on the vocabulary you choose to use. You're confusing the abstraction with its implementation.
"the cost of making the poor decisions"
The poor decisions are usually made for the sake of operational convenience rather than security. Let's not bother with the inconvenience of a separate privileged user ID with a separate password, let's not enforce complex passwords, let's not fence of different bits of the network, etc.
The hardware and languages will support better but better costs some user effort and a few seconds of time here and there and time is money...
AIRI the first Mac was little more than a toy. It required a few upgrades to become useful. Jobs had fought the developers to design it to a price. He also did his best to make it non-upgradeable, Somebody sneaked in an ability to solder in a link to enable it to take higher capacity memory without having to buy the bigger version next year or whenever it was. I wonder to what extent Jobs planned it as a bait and switch.
Over the years I've seen a few would-be memory technologies come and go (bubble memory is one that comes to mind - what was it again?). The few that have stood up to reality and their predecessors are the few we have today. Reality includes the cost of completely revising the way things are done to take advantage or having a commercially viable, if niche, use case. Mobile would appear to provide a suitable use case for PRAM. If it can't make an inroad there it doesn't seem likely to succeed elsewhere.
"there's always been an engineering trade-off between speed and capacity"
Between speed and price and it's price - including the price in terms of power consumption - that determines capacity. Can it achieve the speed of DRAM at the price of DRAM or close enough to lower the overall cost of the system?
You have a grid of icons and you can't directly see your "documents". Open an app and it picks up where it was. Turn off the fondleslab and when you turn it on it's right where it was, although some things will be slower to open at first.
That may work for a piddling little mobile. On a laptop or bigger my documents are primarily what I need to see. "Open recent" will bring up more than one document for me to pick but not many. If I've had to hunt through a few recent documents to check on something that may well have been enough to have squeezed what I was actually working on yesterday out of the list.
The advantage of the files and folders way of thinking is that the organisation can be structured so that I can navigate through them to find that document i last opened 5 years ago but need to check up right now.
OTOH while Okular will open a read-only PDF where I left it 5 years ago LO Writer will only re-open a document it was writing 2 minutes ago at the beginning. I could do with that being changed.
There's another long term problem for term If they may think they're losing the small fry they don't care about but among those are the not-yet big fish who they'd want eventually. Of course if you think in quarterly figures that's not an issue but that in itself should be enough to ensure their competitors outgrow them.
I wonder how many of those who saw it coming were already making their plans or actually migrated well before the announcement.
Again we see how businesses say one thing, then act in a completely opposite way. They always talk about how customers are so important, but then at every single opportunity, they skimp on the customer service.
It's not really saying things. It's just the standard PR process of joining together strings of words. They're not intended to have meaning. Just like generative AI. PR people will be the easiest to replace if they haven't already been.
"Fortunately for our sanity and the health of our company, she was gone (of her own volition) within a month."
On to another step up the ladder no doubt. I had one of these ambitious kids attached to my team once. Fairly recent graduate, no skills appropriate to the team. I think she'd come from somewhere where she'd been for a short time. About a month or so later she moved to yet another job having, as far as I could make out, having gained any useful experience while she was with us.
As far as I can make out if you take any single word in the middle of the output the few words either side of it are words which are more likely to be found in those positions in the training material than any other words taken at random. The run of words is very likely to look like it was written by an intelligence because it has been so written many times. But as you get a word or two further away the next lot of adjacent words also emerge a very likely from the training material but not necessarily from the same contexts as your central word and even less likely will the words a little way either side of your central word have come from the same contexts. It might all appear to make sense as a well constructed sequence of English words but not as a coherent assemblage of facts or argument being expressed by those words.
"You'd think lawyers – ostensibly a clever group of people"
Having mixed professionally with a good few barristers I'd say their cleverness followed a normal distribution with a fairly large variance. Some were sharp and picked up on what they were being told PDQ but there were one or two whose abilities didn't seem to get much further than putting the gown and wig on the right way round.
Fortunately the judges did seem to get picked from the upper end of the distribution.