So no IP in Watson.ai was in Watson Health, which IBM sold? Completely, 100% distinct products, no common code? Despite both being machine learning,'AI' platforms?
Can Watson.ai be trained on medical data and recreate Watson Health?
1608 publicly visible posts • joined 29 Sep 2016
I didn't really mean performing, but I get your point, having been on the pointy end of the point some time ago. I now work for local Govt, and one of our web sites got scraped, and somehow a draft proposal that was never published, got into the hands of the UK PRS (Performing Rights Society). The proposal was to license busking pitches around town as part of urban development, so shoppers got live music and they were spread out so didn't interfere with each other. It was never enacted, but that didn't stop the PRS assuming the performers would be singing covers of copyright material and demand a licensing fee, from us. Now, at least one of the regular acts that busks (and we do not license, or charge, it's just an allowed thing in certain places) perform traditional, non copyright music, but the PRS wanted a slice of that, so Ed Sheeran could have more money. I'm not knocking Ed, but I doubt the PRS reimburse small artists pro-rata. But then this reminds me of the CDR levy, and the music industry wanting a slice as they assumed they were losing revenue.
I just asked chatGPT to read me Harry Potter and the Philosophers Stone in it's entirety. and it said it can't do that as it would take several hours, and exceed the capacity of the platform. It then suggested I purchase a physical or digital copy, borrow a copy from the library, or buy an audiobook.
Maybe you can get it in chunks, but I think there are probably simpler ways of finding the text, like a second hand bookshop.
I am not convinced that reading can violate copyright. Copying and reproduction, passing it off as the work of another, yes. But if I buy copyrighted material, I can lend it to a friend if it's a book say, or they can listen to copyright music at my house, I don't see how a computer reading is any different, as long as they do not copy, or distribute. I guess these are interesting times and we need to iron out the detail.
I used to work for IBM UK, and I worked from home for about a decade, before taking VR in 2015. So I, and may of my colleagues were WFH long before the pandemic saw widespread adoption through necessity; we were encouraged to WFH by big blue so they could sell off offices and save money. My base offices shrank considerably during this time, from two entire wings of a building, to eventually just one floor at one location.
So what's left, 'hot desking'? It's one way to make your staff feel unwelcome. Oh, wait, they'll want to shake the tree a bit so they can make people leave and try to replace them with AI.
Heck thinking was drummed out of people.
I used to get quite frustrated trying to fix broken processes. I worked in security, and the templates were created in the USA, then automated by a colleague in the UK, he annotated the doc, and it got sent to Poland, where a 'gap' document was created. Problem was, he ticked off stuff he had automated, so when the team in Poland (not technical, not English as a first language) would see clarifying statements in the doc (that of course had not been automated) they added them to the 'gap' document we were supposed to answer manually and add 'is this the case'. I had to complete these and they were scored by a team in India. Explaining to the Indian team that these were not supposed to be questions was really, really hard. They'd been told to follow procedure, so there was a 'question', it had to be answered. My problem was that if an auditor saw the 'question' they'd realise it was complete b*llocks and start digging. So I had to get the documents amended, but that wasn't part of the process, so it was swimming against the tide.
Well, that did happen in 2018, in the UK: https://www.bbc.co.uk/news/uk-england-beds-bucks-herts-43934504
But this of course adds fuel to the argument made against Tesla, that many people thought the system was more capable than it actually was, and then we have to ask how they formed that opinion.
Does the AI have a bank account to pay the sweet licensing royalties into? Of was Thaler planning on banking all of DABUS' earnings himself,... just how much autonomy do we think he was granting here,.... surely, if AI becomes considered sentient, you have to pay it, or it becomes slavery,....
Scraping for material,.... how many acts cite 'The Beatles' or 'The Rolling Stones' or Jimi Hendrix as an influence? New music borrows from what has gone before, and record companies can't monetise inspiration, yet when it comes to AI, they try and stifle it, instead of signing it. If there's one thing you can guarantee, when new technology comes along, the music industry will do the wrong thing.
It's hard to see any energy saving when most folks are going to access their cloud PC via another PC / laptop, so the cloud side is an additional consumption. Our VDI solution runs on a Hyperconverged cluster of servers,... it's on 24x7, and while it might not use as much juice idling as it does when it's processing peak demand, it uses a lot more than a PC that gets switched off at the end of the day.
I keep meaning to pick up one of those little frames that hold zip loc bags open, for portioning out. If I cook something in my slow cooker, I make sure it's full and portion out the rest for lazy days, and Tupp, sorry, generic branded plastic containers use up a bit too much room in the freezer.
I think it works the other way around, nation state level subversion of technology company employees allows them full access to source code, and to embed dubious code and exploits into commercial products. We get regular briefings from a chap from the NCSC, but NCSC is kind of like the word 'love' tattooed on the knuckles of GCHQ. They are trying to help, while their comrades actively exploit, and do not report vulnerabilities. I suspect some GCHQ staff are also on other payrolls.
I've not tried Bard yet, I'm still on the waitlist. ChatGPT produces some useful results, and Bing AI less so, although I just tried it again and it did OK, even though the answers aren't constructed as well, they do seem more like search results that have been wrapped into a sentence, unlike ChatGPT essay format. Can't wait to try Bard though.
I'd like to think it was #1. Looking at UK suicide rates, things were a lot worse over here in the 80s, we had a steady downward trend until 2007, and it's been creeping back up, although not to the same levels yet. I doubt all of the increase is down to social media although it's probably contributed some, I would also imagine mounting student debt, job prospects, inflation, the economy, worrying about climate change etc also have a effect. If today's youth are seeking out counselling more, that's got to be a good thing.
Yeah, I'm not convinced I'd try to make explosives by following a chatGPT recipe, without checking them against some reputable source, just in case it was a little cavalier.
I've generally had good results from chatGPT, but recently asked it for a picture of the Mona Lisa, it said it could only generate text, not images, so I asked it for an ascii art Mona Lisa.... it tried,... but the result looked more like a motorcycle helmet. So I'd take it's answers with a pinch of salt(peter).
I guess the issue would be Stable Diffusion wanting credit / royalties for art their product created on behalf of the requestor, should the art prove to be lucrative. Although AI art might be poised to kill the poster market, why have something on your wall that's a copy, when you can have your own bespoke creation? I'm kind of surprised Stable Diffusion doesn't have a click through to order a physical copy.
This happened to someone I work with. A local councillor had created some presentation about urban regeneration, and used some images they got from a stock image place. The presentation got uploaded as meeting minutes, the site got crawled, and then they were looking at a copyright infringement and had to cough up some cash.
A slot for a keyless key might have saved a former colleague of mine a right PITA. She'd spent the weekend camping, packed up, driven off, then some time later stopped for coffee. Returning to her car, she realised she didn't have her car key. It must have fallen out of her bag near her car, allowing her to start the car and drive off. She called the campsite but it had been a popular event and lots of cars had exited the car park since and the keys were not visible. So she needed recovery, and a replacement key,... all because you can start the car without knowing exactly where your key is. Not a fan of that idea myself.
" I somehow doubt that datacentre staff will be allowed to poke the reactor."
Nor would we want to : -)
I used to be located at a datacentre, and we all took on various responsibilities for the infrastructure, one guy used to play with tanks at the weekend, so looked after the diesel generator, I used to inspect the aircon / chiller room,... neither of us were nuclear engineers, and you don't want a tecchie saying 'what does this button do?'
no idea why your post got a download, I LOLd at 'Retch Media'.
I presume the fast food articles are advertorials really, and sponsored, as they certainly aren't news.
On spelling and grammar,... this article said they'd proof read AI created content,... well, plenty of spelling and grammatical errors slip through the meatsack generated content, so I think few articles get checked as it is, so I doubt many AI articles would in the end, they'll just publish and be damned.
My local rag, the Lester Mercroh is a Reach publication. The 'news' articles are often 'journalists' telling us some people on Mumsnet had an opinion about something, where to get free fast food, or that people watching some TV program were upset. I only know this because I use an Ad blocker, otherwise the site is unreadable.
AI written articles can't be worse, surely.
"Now we have AI, and there is half a chance you are arguing with a confrontational computer."
Ha! A colleague and I were musing on this the other day, with the news that Windows 11 will incorporate Bing AI in it's search function,... I can't wait for Bingy (the nickname we gave it already, as I presume it will pop up like 'Clippy' and try to be helpful) will interject, and say "you appear to be in an online spat, would you like me to cherry pick some data to support your argument?' Only for the person at the other end, to be wielding Bingy as well. Maybe just like WOPR in Wargames, it will soon realise it can't win the silly game, but then there was that story about Bing AI declaring victory in a game of Tic Tac Toe it lost,.... so maybe we're all doomed.
... and ChatGPT likes telling stories. Ask it to tell you a story, give it the setting, and the characters and it will tell a story, with a beginning, middle, and an end. So, I guess in this case, i felt compelled to write and end to the story of Alexander Hanff.
Some years ago, I was checking in for a return flight from Glasgow to East Midlands. The security dork at the gate took umbrage to my wallet chain. Said "It could be used as a weapon". I pointed out I'd flown up with it, and nobody stopped me. That the lady in the queue in front of me had a bag with rope handles, that the guy she was with was wearing a tie, that we were all wearing belts, had shoelaces,.... but nope, i had to go check my chain in. So I did, went through, and what was on sale in one of the shops on the other side? Yeah, wallets with chains. Security dork didn't even know his own airport.
I work in local Govt, and we have the same issue, the workforce is getting older, people leave and we aren't attracting young hires, as they can get better money elsewhere. We have one team where everybody left, so we've lost all the historic knowledge they had. Many of us are in our fifties and trying to work out how to retire as soon as possible.