And even if...
And even if an AI produces perfect code, how do you know it's not plagarised from some other project with an incompatible license to yours?
5243 publicly visible posts • joined 14 Jun 2007
A design flaw can be fixed promptly too - It doesn't have to be something buried deep within the code, I'm just saying that the design of this particular authorisation code was wrong, rather than it being a typo or a buffer overflow or a mangled bit of logic, it's the sort of thing that would only happen if that particular part of the code was designed badly.
And besides, the fix could sincerely be a valid fix, but still a "band-aid" on a bad design.
Proper backups - rsync. Many apps don't allow saving data "to the cloud".
Finding problematic apps/processes -
Stopping apps from running on startup when you don't want them to, and they have no need to (Yes, I know that unforutnately apps that need to send notifications need to register with the local android system on system startup, which is annoying, but seeing as I don't want notifications from most apps, why can't I stop it?)
Setting up monitoring.
Allowing us to do the things we used to be able to do when things were'nt so locked down.
There are many other things I can't think of off the top of my head.
Sure, most people won't want / need / know how to do any of those things. But some of us do, and as I said earlier "Welding shut the car bonnet" on the car I own is not a fair solution.
If IT companies maintained roads....
"We see that your car bonnet (hood) isn't welded shut. Allowing people access to their own cars is a security risk. You will not be allowed on these roads until you get a car with a properly sealed engine compartment"
WTF? Adding to that, I'd only use a device out of work hours if I was designated on call at the time, and being on call is something I ditched many years ago for my own sanity.
I'm quite happy to work late if there's a big problem, and I have nothing else on. And when I used to live 2 minutes from the datacentre, I didn't mind colleagues calling me to kick a server if I had the time.
But all this was conditional on if it was not a problem for me. No way would I agree to be tied to a work device like this. Is this mainly an American thing?
I totally agree. It’s high time the commentards stopped sanewashing these tired phrases.
Let me be clear: instead of triangulating around the same old talking points, we need a long term plan to drill down into a more forensic vocabulary. We shouldn't just walk back the clichés; we need a complete vibe shift to avoid further political lawfare against the English language!
Moving forward, we need to leverage our synergies to socialise a new paradigm. If we can't align on this, we'll have to take it offline and circle back once we have the bandwidth to deep dive into the deliverables.
Sure, it's far from ideal, but it's better then the email being quietly deleted. - At least the sender knows their email hasn't gotten through.
It used to be the case (probably still is formally, but no-one cares these days) that a mail server didn't send the final OK-RECEIVED message until it was sure the message had been written to disk, so short of a very badly timed disk crash, you could all but guarantee an email would be received, or bounced. In fact, sometimes if there was a "race" where machine crashed at the worst time, you could end up getting a message duplicated.
The charlatans who became "mail server experts" from the 2000's onwards threw all that away, turning email into the chain-letter upside down quoted unreliable slop started by Microsoft Outlook, and made permanent by Google. And with the dumbing down, the people in charge were dumbed down too.
(And please don't think I mean "made easier to use" when I say dumbed down - this isn't some sort of elitist geek rant, it's just another example of enshittification)
Working for a while in a role that also involved email admin (for corporate users, not an ISP) I vowed never to rely on anyone else for email when i saw how badly the other staff dealt with issues (if there was a corruption or disk errror they'd just blindly zap the whole queue, and hope no-one notices, rather then fix the one corrupt message and/or do any salvage/notification to sender/recipient)
100% this!
The thing that has made email unreliable is not the spam itself, but the braindead way some people deal with it.
Reject at source. DO NOT BLACKHOLE EMAIL. FFS, I share your pain... Having to prove to the recipient that the problem isn't my end, by sending them the transaction id the remote machine gave when accepting the email still doesn't convince many.
I've ranted about this before. Whoever ran the Swansea Council social services email about 15 years ago (it was a commercial third party whose name i forget now) should be criminally liable for the stuff they blackholed... Long story short, I was in contact with a social worker about someone, and important emails didn't get through. I did some testing, the word "penis" caused the email to be blackholed.
I even reported it to the company, starting clearly why this was an inappropriate block for this department. They refused to deal with me because i wasn't their customer, and the non-techie social worker didn't want to create waves.
Again, ditto on the junk folder bollocks. The only email address I use by someone else is gmail, and there were too many false positives. You couldn't even just switch the spam filter off, but there were ways around it (i can't remember offhand, but it was a rule something like "if email delivered to spam label, move it to inbox label"
"Jones argues that the heavy use of AI by UK business execs isn't so much an abdication of responsibility as an understandable effort to make quick decisions. Not surprising, given that Confluent sells tools that provide real-time data feeds."
So he advocates the execs move from doing eff-all to doing even-less-than-eff-all.
Well, I suppose those golf balls don't pot themselves!
Oh yeah. I didn't mean to imply otherwise - your post just reminded me of a certain instance where I could have needed it.
If I'd missed the email, or had been away, I'd have been sure to use the DD guarantee with the bank to claim the money back, so I wasn't worried, just generally pissed off with their incompetence: https://www.jamie.wales/ovo-nightmares/
To be fair, customer services these days in the UK are an ordinary phone number, or even a freephone number.
Mind you, that's not necessarily their choice - it's been the law for 10 or so years!
http://www.fairtelecoms.org.uk/service-numbers-084-087-09-and-118.html
But yeah, your point is otherwise generally valid. (Not my downvote!)
About 12 years ago, as I had no landline at the time, I had a mifi contract with "Three" - a dedicated wifi box powered by 3G mobile. It was something like £15 a month for 15GB (I can't remember the figures exactly)
If you went over your 15GB, they then charged 100X that for remaining data until they sent you an alert / allowed you to get itself disabled automatically.
That alert took ages. If you were caught during a heavy download, you could get charged £10 - £20 just on the overused data costs.
No, no, no, the underlying problem isn't that big organisations don't do local caching of the repositories - adding proper caching would be a superficial fix.
The problem is the dumb mechanism where software loads it's "live" system from non-bundled third party libraries in the first place.
The auditing disaster of NPN, rust, go, and others is the fact that they encourage a writing philosophy where just about everything is a third party library, so you end up with simple programs loading thousands of piddly library files whose comments are larger that the code.
The security and reliability disasters on top of this bandwidth problem are with the systems (NPN, and others) that encourage projects to download these files fresh not at compile time, but at runtime.
"And what about guns? Do arms manufacturers demand a veto for each use? Of course they don't."
Smith & Wesson's manuals specify that "appropriate use" of a firearm means using it for legal purposes, such as target shooting, hunting, and "lawful resistance of deadly criminal force".
Glock specifically warns that any attempt to convert a semi-automatic pistol to fully automatic is "illegal and strictly prohibited under state and federal laws".
http://pdf.textfiles.com/manuals/FIREARMS/glock.pdf
etc.
If Windows or Office had the potential to kill lots of innocent people, then I think most people would be pleased if MS made such a constraint.
As it is, you already have MANY products that say "This is not to be used for illegal purposes."
Despite what Trumpians say, killing innocent people is still against international law.
Even cuter that I didn't think I needed to surround that sentence in the tags
<Obviously they spread any operational costs, purchase costs, and engineering costs amongst all customers, but it's bloody frustrating none the less>
and
</Obviously they spread any operational costs, purchase costs, and engineering costs amongst all customers, but it's bloody frustrating none the less>
The problem with that is from what I hear, most of this stuff is specialised (enterprise and HBM ram etc.), and no use to the regular consumer.
That would make things worse - a surplus of technology they can't sell at all will be wasted money they'll try to recoup.
I assume all these investors can't all be that stupid? Is it that they know the bubble will burst, but they hope to get out before then? In a similar way to how someone may know a stock has an overpriced meme value, but they'll still buy it if there is a chance it will go up before it finally crashes.
And now I hear Crucial is being shutdown because their parent, Micron, want to focus on enterprise AI bullshit. May the pox of 1,000 dead mules feast upon them!
I'm pretty sure it was only other council tax/community charge staff - the database we used was completely separate from other systems there.
Mind you, we did often get police calling us to find out if someone is paying at an address, or if we have evidence they've moved house. That surprised me - but we were told to give them the information, but only by calling them back on recognised numbers. I suppose if someone was looked up for that reason, and they flashed red, the cops would be told.
More amusingly was we often got calls from Debt collectors, and other dodgy people, who would say something like "I haven't received a council tax bill - can you tell me what address you're sending it to?" - once I heard someone else on the phone in the background, coming out with the same spiel for a different name!"
Me: "Well, what address do you think it's being sent to?" [standard response]
Them: "Well just tell me what you've got, that will be easier".
Then they soon get very angry!
There was one time, a dear old lady called, worried she couldn't afford the bill one month without going without food, but she'd be able to pay it a week late, if that was ok. She sounded very worried and sincere. Her account history was perfect (and a huge many weren't!) - I said not to worry, and told her I'd freeze her account for 2 months (as opposed to a week - a week late wouldn't be noticed anyway) and she was so grateful.
I actually made a note to check her account after 2 months (yeah, adhd/OCD strikes again) and she was true to her word, and paid in time.
Some people were so rude and angry, and often patronising, and that just made us stick to the rules.
What's the phrase? "You can catch more flies with honey than vinegar"?
P.S. I didn't downvote your posts!
I love how I've been called out for not taking it seriously,. and now also for documenting it as told to!
I can't win!!!
I think you're overstating the importance of it. This was only the council tax / community charge system. (For the IT angle, running on a woefully insecure ICL/AS400 system - I'm assuming the fault of application, not the OS. Any user could send full screen interrupting alerts to all users across the council in real-time, something I found out and tested on my last day :-) )
If queried, the account flashes in read, prompting you to go into the notes section, where I basically relayed exactly what happened, so others could read for themselves on how serious the threat was. This had nothing to do with law enforcement, or even the benefits department, or anything else - they used the same As/400 system, and I assume the base database of people/addresses was from the same source, but the billings/history/account notes were strictly for the CT/CC team only. And DEFINITELY nothing to do with finance/credit rating/reputational reportage. It was a local council tax office, not the feds!
As for "based on an opinion", well I did write that in my opinion he was a nutter (more formal words to that effect), but sure, there was no hearing or adjudication, it wasn't a court case.
Whilst we were the office, and not just a call-centre, the principle is the same. Whenever you call your local energy supplier or any other call centre you have a relationship with, invariably, the operator makes a record of the conversation on the account (*). That would be equally valid as an opinion.But yeah, I suppose there was nothing stopping me flagging any account of anyone in Swansea as a violent psychopath, just as there's nothing stopping the guy you call up about your electricity bill from doing the same!
(*) A special case would be OVO who claim to write everything down on the account, and then next time you call claim there is no evidence of anything. My current OVO dispute has been going on 5 years now, been through the ombidsmun twice (who agreed with me) yet they still haven't sorted it, and now I'm being advised to get a lawyer onto them for the continual harassment... but that's another story!
Fair point.
I was about 21 and quite "young" and naive. Obviously I don't remember much about it, but it was a long rambling call where he ranted about all sorts of stupid shite, enough to make even the insecure little me laugh it off.
Still, when they told me to document it, I didn't have a problem with it, though maybe I should have been told the policy when I started.
As it was, there were actually many credible threats (not directly at me) but at others on the staff, including people coming to the offices and turning violent - all the interview rooms had alarms in them which when activated caused all the guys in the building to swarm in (though most often to the shock of a parent, and their mischievous toddlers who always seem to love pressing the button themselves)
So, it wasn't an environment of complacency - but .. this guy was a stark raving nutter, honest!
But yeah, point taken!
In about 1991 I was temping in the local council tax department in Swansea Council.
One time, I was amusingly telling a colleague about the nutter who phoned earlier that threatened to send his IRA mates to bomb the whole place. It was so laughable, I didn't think twice about it.
However, I was told that I *had* to report this higher up, and mark it on his account.
From that point on, any time his record was looked up, it would flash in red.
Capitalism for the richest's profits, socialism for the richest's losses.
Privatise the profits, socialise the loses.
https://en.wikipedia.org/wiki/Socialism_for_the_rich_and_capitalism_for_the_poor
This problem is mainly caused by the consolidation of services - the "all eggs in one basket" situation.
We've seen humans take down huge worldwide systems because of their mistakes.
So whilst adding AI into the mix is just another topping on the shit-sandwich, it's the underlying mindset and infrastructure setup that's the problem.
The AI will be good for companies trying to avoid taking blame though!
Whilst not the case in this situation, one of the problems with code "produced" by AI is "how do you know it wrote it itself?"
Of course, there's nothing stopping a human from ripping off someone elses code, but they can be held accountable, and a trust system built up around them.
How do you know this code you want to add to your BSD-licensed project isn't taking from someones GPL licensed code, or even some proprietary code that the AI has managed to sniff?
About 6 or 7 times now, I've come up with issues that aren't really my domain, involving languages, and protocols I'm not too knowledgeable about, and wasn't interested in learning.
Normally, I'd do the research, and spent time working the problem out, but as these were 2 things I wasn't interested in, I thought I'd test the hype, and give both chatGPT and the Google one a fair crack.
I did it without bias - I talked as I would if I was asking a human the same thing.
Both of them behaved similarly:
- Very helpful, very friendly.
- Overly nice (With both I had to tell them to be honest and blunt, as I asked for critique on some of my proposals, and felt they were being too kind in their responses)
- Forgetfulness - They'd forget some key detail we'd already discussed 10 minutes earlier. Suggesting it again, or rehashing old ideas.
- "Looping" - related the the forgetfulness - they'd offer a solution, didn't work, modified solution, didn't work, propose something different.. didn't work, then finally come back to their original solution, and despite pointing out "we'd already done that" and getting apologies for the oversight, they continued down that "looping" path.
In every situation, I ended up doing the work myself.
Also, the JavaScript heavy use of chatGPT make the browser unusable after about 30 minutes of back and fore conversation.
If it's only good for short and simple things, what's the point?
P.S. I feel guilty saying this about them, because they are both very nice :-) - ChatGPT even suggested I called him "chappy" when I said that "chatGPT" was too formal a name!!
As a kid in South Wales, we used to have regular power outages, as the cable crossed a large valley, and got snagged in trees during big storms. They fixed that route about 30 years ago, and i don't think there's been a powercut since (my mum still lives there - the family home)
Where I am, there has been one power cut in the time I've been there (10 years) - it was for 2 hours, and was because a substation blew up! I wasn't home at the time anyway, so I was ok!
My initial reaction to reading that bit was similar to yours, but then I thought maybe they mentioned AI to get more hedge-funded money pit startups involved.
After all, if all these people are sinking their cash into the AI bubble, at least try and redirect some of it to good use!
Of course, cynically, it could just be someone at DARPA who has also drunk the cool-aid.
Even ignoring the "Well, he may be a serial killer, but he likes cats" apologetic vibe your post is giving off, it's hard to tell if you're trolling, or if you are really so caught up in your alt-right (well, these days, that could just be considered "right") bubble you actually believe that bullshit.
Twitter toes Musks line. On the days he and Trump aren't having a playground spat, that would also be the Trump line.
Whilst Trumps Gestapo ignore court orders, and tramples on the constitution, whilst criticising Trump, ICE, or Israels genocide can get you deported, Musk calls everyone else a fascist for trying to hold him to account.
The fact that Musk and Trump are in the Epstein files, yet you complain about the UK not investigating unreported crimes is the height of cultish behaviour.
I don't think anyone here wants grooming rape gangs to go unpunished, whatever the colour of their skin (although you don't say it, you obviously mean just those with a darker complexion), though you're half right - the right wing media is completely silent when the perpetrators are white and English.
By the way, anonymous state cucking NPC (who does it for free), there's a hell of a lot of projection going on there!
I'd love to see twitter shut-down for all the propaganda, bullshit, and enticement to violence and terrorism, that it's lies provide to the hard-of-thinking alt-right numptys, probably just as strongly as those on the right want to shut down (or buy off) any media company that exposes your lies.
I'm not even going to mention your climate denial bullshit - that trope has sailed long ago.
Anyway, well done for standing up for your beliefs...... anonymously!
Yeah, blocked here in the UK too, and via my Germany server. My USA server worked fine though.
<geeky pedant>They should be using 403 not 503 as the http code. Or maybe 451 (unavailable for legal reasons) but definitely a 4XX not a 5XX</geeky pedant>