A Bit Harsh
From the article: "one-man bans".
That seems a tad extreme. Who is this chap? What has he done?
No, I don't have anything useful to contribute. Thank you for asking.
Rosie
152 publicly visible posts • joined 10 Apr 2013
I'm not quite sure I understand what this comment is trying to say tbh. If you're saying that the OSA should not prevent children learning (in an age appropriate way) about gender dysphoria/incongruence or sexualities other than strictly straight they I'd completely agree with you. For a short while that wasn't the case due to 'errors' but that got corrected.
If, on the other hand, you're saying that children who may have gender dysphoria/incongruence and be gay should be blocked from accessing resources that help them understand it then I feel that would be a really, really, bad idea.
Rosie
In my head, it really is very simple. If you are running your business on anything remotely cloudy (IAAS, PAAS, SAAS, whatever), then you do not own your business, you are renting it. As soon as you find yourself unable or unwilling to meet the rent payments, you have no business.
The same is true to a lesser extent for propriety software. Once a vendor has their hooks deeply embedded in your processes they are at liberty to give a sharp yank at any point, tearing money leaking wounds all across your business.
Rosie
Also musicians. Or at least those of us that pretend to be musicians in our spare time. You just cannot record and instrument 'in the cloud' with anything like the latency you need (<=5ms from hitting the note to hearing the note) so local is the only option. Mobile isn't there by a long stretch and Mac, well, it's Apple.
Rosie
Seeing this in a Reg article makes me sad. I mean, really? El Reg's readership need to have 127.0.0.1 explained to them?
Friends, let us now gather in solemn silence and fond memory for the death of a formerly noble techn website, written by techies, for techies.
Rosie
I feel that's a tad unlikely. I mean you would really have to be properly stupid to use evidence of a crime you'd committed as evidence in a lawsuit you're persecuting and I'll give them the benefit of the doubt. I've not checked where this case is in the protracted and byzantine legal process but aren't there steps where one party is allowed to have a rummage around the other party's documents?
rosie
That thing about moving from information gathering to information verification. How? If you've asked the AI to fetch some information for you then that's likely to mean that you don't have the information and to verify the information the AI has given you, you have to go back to the source, i.e. find the information.
I may be missing something but I'm not getting how this helps.
Rosie
I get where you're coming from but that's not what that that text means. The HAN interface is the radio interface within the home as opposed to the keypad on the meter. A supplier can turn of the energy supply by sending a SR7.2 service request to the device.
In reality that functionality is locked behind serious RBAC and ahs never been used. Outside of an emergency it's unlikely it would ever be used.
Rosie
I'm not sure why noise is being raised. Using the Joby for reference, the figures (certified by NASA if Joby's website is to be believed) are 45.2DBa for flight and 65DBa for take-off and landing. Which is pretty quiet. For comparison the same numbers for a helicopter are 82-97 for flight and 86-95 for take-off (source: https://www.epd.gov.hk/eia/register/report/eiareport/eia_1132005/EIA/html/App%203.3.pdf). Given DB is a logarithmic scales that's a pretty massive difference, enough that I don't think you can transfer an experience of being near a helicopter to being near on of these things.
That's not that I necessarily think these are a fabulous idea, certainly not until after a few rich people have selflessly helped work out the kinks (flaming death optional).
Rosie
You'll find it here: https://assets.publishing.service.gov.uk/media/5a80f0fced915d74e33fd44d/Ofcom_Framework_Document_23_June_2016_signed.pdf. Section 5 would seem to be the bit that applies and it (to say the least) a bit spartan on detail.
Rosie
As I was reading this I was thinking "that looks a lot like the simple debugging you do when you haven't got a better IDE". Which, in my mind, is a good thing. It's level two fault finding and fixing, just after "throw some stuff at it and try to work out what's going wrong from how it fails". So now I stuck with a mental image of what a full-on debugging environment for ML/LLM would look like and hoping it's a navigable, multi-coloured, 3D lattice with the ability to prod the odd esynapse to see what happens. Utterly pointless musing but it helps pass a Friday afternoon. Hopefully it'll quicken the arrival of something that actually deserves to be called AI.
Also, that's a bit of dead clever boffinry, isn't it?
Rosie
This seems to be, sort of, the start of a useful think that got abandoned before it was thought through properly. Given the problem is 'how do I tell what has been mass produced by an AI disinformation farm?' what's being proposed doesn't look like much of a solution to that. Mulling it over there are 3 broad categories I can think of that apply here:
1. Information that is generated by AI and the producer/relevant authority wants it to be labelled as such.
2. Stuff that just anyone wants to post without being particularly bothered about whether anyone pays any attention to it or not.
3. Information that the producer wants to unambiguously tag as produced by them.
For the first point, "relevant authority" was put in there to cover scenarios where local law requires that AI content be labelled as such. It would work for people that are interested in following local law but obviously fall flat on its face otherwise. A producer may want to label something to prove that they have ridden the prompt dragon skilfully enough to get it to cough up something worthwhile, mostly like a visual artist signing a painting.
The second really, really doesn't need nor should require any kind of crypto-authentication shenanigans. If I want to post utter dross like this on the internet that's between me and my own foolishness and people should probably treat it with all the respect that deserves. Not a lot, for those not used to British sarcasm.
For the third, that might actually be useful. If I want to be sure that something I'm reading really has been produced by the organisation that it's claimed has produced it and hasn't been altered in any way having a handy way of doing so would be helpful. I know that's been possible for decades, it just hasn't spread beyond the niches where the techies find it useful to just be a thing that everyone uses.
Anywho, random ramblings - feel free to point out any and all silliness in the above.
Rosie
Hmmm...maybe I'm misunderstanding what it is you're trying to say but isn't the whole point of an election campaign to convince people that one particular set of problems is more important than another and that you are the best person to solve those problems? Which, at least in my head, falls into the fairly broad definition of what could be called propaganda.
Not that I'm defending the self-promoting, fantasy definitions of "problem" and "solution" that are getting thrown around far too much at the moment.
Rosie
AI is unlikely to be useful for reviewing the sort of documents I review on a regular basis. It'll be OK as a turbo spell and grammar checking doodad but that's only a tiny part of a proper review. I don't see how it's going to cross reference half a dozen or so documents, handwritten notes in a day book, conversations had away from the computer, memories of an what happened years ago and all the other plethora of nonsense that I need to use when reviewing a document.
Rosie
Tsk, tsk, tsk. Silly downvoters.
The thank you is for making me aware of something that I wasn't previously aware of, more knowledge is always a good thing. So far as wanting more of it, it's more of people who feel they are negatively affected by something getting up off their behinds and actually doing something about it. Much, much better than caterwauling on unsocial media about how it's NOT FAAAAIIIIRRRRR!!!
Whether blowing up property or not is an appropriate course of action is debatable. In the same way that whether someone is a terrorist or a freedom fighter is debatable.
Rosie
I was over in London yesterday and work colleagues were complaining about the rain, about the level you'd call "a nice day for a stroll" in Wales. I messaged the same colleague today when I'm back in Wales - it was raining hard enough to set off car alarms.
Welsh rain is different. It remembers when it was part of the sea and could wreck our stupid ships and drown our puny bodies; it deeply resents having that taken away from it.
Rosie
Normally I'd agree with you but the article has the quote "...driving click-through rates upward compared to manually generated content...". I feel that's the metric that counts here. Even if it were quicker and cheaper to generate the content, if it's not causing a positive effect then it's just money down the drain.
Whilst there's the usual caveats of it being the CEO saying it and the article is a bit thin on detail so it's difficult to make a proper assessment it would seem there's at least the possibility Grab have worked out how to use GenAI effectively for their use case. Which is at least interesting enough to want more information on whether it's true, if it is why is it true and whether there's anything that's been done which could be made more generally applicable.
Rosie
So long as I remain in control of who gets paid, and how much. Given that, things I'd want to see are:
1. An opt-in - similar to the exceptions in NoScript. If you're name isn't on the list you can't have any money.
2. Graded levels of permission. If you're asking 0.01p go ahead and take the money, if you're asking for 1p then I need to confirm I'm OK with that, it you're asking for £1 you're almost definitely not worth it and can sod off.
3. I get to set the levels in 2.
4. Proper, audited, maintained security controls with someone in the firing line should they fail attached to the payment mechanism.
That said, if Google/Amazon/Microsoft/$insert_web_megacompany_here are getting anything from it the answer is no.
Rosie
I've used Ardour with Ubuntu Studio for years and hadn't paid penny until I felt I wanted to; Mostly because people need to eat and partly because once I worked out the hourly cost for the fun I was having it was pennies. Granted I've run Linux on my own machines exclusively since forever so have never run acropper of the costs (in either time to compile or cold, hard, cash).
It's a great, big, capable program which takes a bit of effort to get your head around but also an absolute doddle compared to a 'real' recording studio. It takes time and effort to build something like that and, personally, I don't feel at all hard done by to contribute a little to the effort.
Rosie
That's about the most jarring bit of US English I've seen on the site so far. In the UK grafting is working your whatnots off, whereas in the US it's skimming a little something off for yourself. Same word but completely different meanings.
In a related point, does anyone know of any sites that resemble El Reg as it was? Sort of mid-2000s to mid-2010s heyday era. I'm liking the new site less and less with each passing week.
Rosie
I thought that would get me some down votes, which is a bit sad really - what's wrong with learning the language of the country you're visiting? I learned a bit of Welsh when I moved here (because I didn't want to be one of 'those' people) and it's a right giggle, more people should definitely give it a try. It's very, very different from English in all kinds of fascinating ways.
Rosie
Hmmm...I'm not sure how many legs using TLS as an example would have TBH. The sane implementations that I've seen use something like n F5 as a termination point, everything inside the termination point is plain text, only outside is encrypted. Which is sort of essential if you#ve got tp do anything based on packet payload a a Good Idea(tm) if you don't want to have key handling hell to deal with and servers wasting clock cycles decrypting when they could be doing something more useful.
Other than that, agreed. This is all going to end up under the umbrella of "aligned with Good Industry Practice" and largely ignored.
Rosie
I thought that as well but if you go a bit further down the article you get to "And of 42 deaths in people with Delta variant infections, 23 were unvaccinated and seven had received only one dose. The other 12 had received two doses more than two weeks before."
So 28.5%. A little less than 30% but not a lot.
Rosie
"know who the bastards* are and can easily avoid them"
Isn't that what Apple are doing?
*Leaving aside any judgement on whether the guy is a bastard, has only acted like a bastard, or just said a things for the lolz in circumstances where for the lolz isn't appropriate.
Rosie
Way back in the early 90s when I was developing code for a software house that specialised in recruitment software text matching was certainly a thing. There were OCR and matching modules to do exactly that - scan a CV and let the machine do the CV assessment for you. I've long since left that area but I'd be surprised if it wasn't still being used, certainly it was on lots of people 'must have' list.
Rosie
I'll admit I've not dug into the low/no code stuff so am slightly commenting from a position of ignorance but from the blurb it mostly reminds me of the MDD/round trip engineering voodoo that had a bit of a hypebubble in the late 90s. IIRC Visual Cafe did a not completely horrible implementation of it; draw your UML and it generated the code for you.
Obviously that fell to pieces when you had to do anything even moderately complicated. As soon as the logic gets a bit involved it turns out it's easier to understand code than it is to follow a diagram. I thought that had killed it off a fair while back.
Anyone out there who knows a bit more than me (low bar to clear) care to point out why this is so very different?
Rosie
I know not an iPad but I've a Galaxy Tab 6 and whilst it wasn't useful for much more than Netflix and games without anything extra installed on it a couple of apps have helped it a lot.
Nebo is great as a replacement for a (paper) note book. I can happily scrawl away but end up with something I can search through easily. Which dead handy when someone asks me about something that came up in a meeting 4 months ago and had a like span of about weeks before fading into obscurity. Until it became urgent again.
Ibis paint is also a lot of fun. Plus it has the advantge of infinite undo for the utterly hopeless artist (raises hand and pleads guilty).
Both Nebo and Ibis come in Apple flavour as well so I think the comparison stands, provided you get an iPad with a stylus.
So far as plain consumption goes. The screen is nicer to look at and the sound quality much richer than my phone. So from that point of view, not really a must have but a nice addition if you're prepared to trawl around eBay looking for a decent second hand one.
Buy new? Don't be ridiculous!
Rosie