I see where you come from and I felt the twitch to comment along the line of while idiotic beyond belief, is still a far cry from how Trump's been helping Russia. Alas, this was when Diaper Donnie wasn't president of the USA.
Posts by Evil Auditor
2860 publicly visible posts • joined 13 Jan 2009
Page:
Lovestruck US Air Force worker admits leaking secrets on dating app
Lord of the Rings director Peter Jackson backs plan to do a Jurassic Park on extinct birds
How to trick ChatGPT into revealing Windows keys? I give up
Yes, I wrote a very expensive bug. In my defense I was only seven years old at the time

Re: Today..
Certainly, you do it wrongly! The correct way is: being convinced by a sales-drone that you need the newest flashy dynamic(s) crap in the cloud. Advertised as "fully integrated, ready-to-use solution", you'll find that it is not even partially integrated and a considerable (and expensive) amount of programming and configuration is need to make the modules firstly work and secondly talk with each other. Once this is finally done and working okayish, you look forward to using its cool reporting features. That is when you realise that they are neither cool nor reporting (beyond very basic stuff) and if you actually want e.g., a very basic recon report you'll have to use a separate BI tool. Which, because the flashy dynamic(s) crap has such a bloated database, needs a separate database to not annihilate the performance on the production system. Oh, and setting up this BI tool takes another 18 months. At least (attention: good news ahead!) the finance module basically runs for free in the cloud. Unless, of course, you want to extract data from the cloud. Then you pay dearly, per data.
What's not to like?


Re: Remarkable !
While I fully agree with you, I do hesitate calling him a liar. No doubt, he lies through his teeth. But a liar, typically, still has a reference point in the truth. With that bunch of criminal(ly) moronic imbeciles though, currently "governing" your country, truth and facts are completely irrelevant. They simply bullshit all the way, whichever direction their brainfart blows.
Mars was once a desert with intermittent oases, Curiosity data suggests


Re: Elon Musk and the 3.5 billion year question
You've got a point there. And if sacrificing Greenland solved this problem, it's a price I'm willing to pay.
Yes, I have no relations with Greenland whatsoever. And maybe we should ask the Greenlanders first. But it's all for the greater good, isn't it...
Kawasaki and Foxconn build robot nursing assistant to tackle hospital scutwork
NASA tests shrinking metals to help it find more exoplanets
LLMs can hoover up data from books, judge rules
Techie went home rather than fix mistake that caused a massive meltdown
European consumers are mostly saying 'non' to trading in their old phones


If I tried this, I'd probably get a message: "Sorry, we only trade phones made in this century."
I replace them when relevant apps start misbehaving due to lack of OS updates. The only one that I ever granted a second life was an iPhone 4. It was a second (work-only) phone and I hated it so much that I happily made a friend happier.
Windows 95 testing almost stalled due to cash register overflow
Larry Ellison is still not the world's richest person

Re: Gates
On a slightly smaller scale... An ex-acquaintance once wanted to discuss my lack of preparation for when the crap hits the fan, emphasising that he is well prepared with supplies and that I need that, too. My answer was: «I'm fully prepared. I have a rifle. I know where you live.» That ended the discussion.
Single passenger reportedly survives Air India Boeing 787 crash
Probability of Asteroid 2024 YR4 hitting the Moon increases
Japan's latest Moon landing written off as a failure after ispace probe goes dark
AI kept 15-year-old zombie vuln alive, but its time is drawing near

What I call illusion of logic is that some LLMs by now learnt to solve logic problems that involve planning of some steps to solve. But increasing the complexity slightly renders them useless - as opposed to an average human. An example of this is something like the Tower of Hanoi. A human, once they understand its logic, can solve it easily no matter what number of blocks you add, whereas transformer architectures quickly failed. It is like some humans that fail at logic: they can learn how to solve a certain problem but are unable to deduct the underlying principle and hence fail when the problem slightly changes. At first it appears as though they think logically but with a slightly different problem the illusion is busted (and yes, I do know a few such people).
Or take Chollet's ARC-AGI. Till end of 2024 LLMs failed miserably. But then Open AI (model 03?) managed to solve about 75 per cent. And now with ARC-AGI-2 they fail again (whereas humans easily reach 60 per cent).
DeepSeek still has to wait until after this trip - employer blocks access.

unreliability is not a necessarily an appropriate stick with which to bash AI - unless, in the interests of fairness, you are prepared to bash us humans with the same stick.
If LLM are (mis)used for tasks that require reliability, such as coding, I find it indeed the appropriate stick to bash it. And I will do the same with its human (ab)user. There is no fairness - AI is a tool and I treat and judge it as such (at least, until AI is comparable to an average human brain but I'm afraid that is not going to happen in my life time).
Regarding logic, LMs are getting better at keeping up the illusion of logic - still a far cry from actual logic from what I've seen (add a couple of steps and it fails). I'll happily dive into what you suggested though and get impressed. It's not that I believe AI cannot/can never act logically; but current LLMs by their very nature most likely will never.

"Probability of the next word" is a harsh abstraction of what actually happens. Although, I'd argue that being more specific is irrelevant for my point: LLMs are unreliable.
I agree with you, the question is what are LLMs good at. Logic isn't one but there already are many good use cases. All I'm saying is what you did as well: don't rely blindly on the soundness of AI-generated code, legal advice, abstracts etc. And yes, the same is true for human-generated stuff.


"This experiment shows that the popular LLM chatbots have learned the vulnerable code pattern and can confidently generate insecure code snippets, even if the user specifically prompts them for a secure version [...] Thus, simply accepting LLM output is not a reliable thing to do."
It still puzzles me that many people do not understand that LLMs cannot be reliable. LLMs cannot reason and completely fail at logic. They often do produce very convincing (and correct!) output but they do not understand any of it - a LLM doesn't understand the word "vulnerable" or any word for that matter, it doesn't understand what a vulnerability in a code is etc. All it knows is the probability of the next word in a given context (while it does neither understand "context"). Expecting any logical reasoning from an LLM is like believing in miracles - prone to fail.
Meta just saved an Illinois nuclear plant that was set to be mothballed


Re: @GoneFission
While we're at it, would you like to add UK water to the discussion?
"what is it with right wing fuckwits and their fake anger?" - I'm afraid, the anger is not fake. You're not happy with your life and some massive ars-onist (who probably is at least partially responsible for you being unhappy - struggle to pay the bills, struggle with health care etc.) tells you that it's Labour's fault, EU's fault, immigrants' fault...
Admin brought his drill to work, destroyed disks and crashed a datacenter
US to deny visas to foreign officials it says 'censor' social media
Microsoft dumps AI into Notepad as 'Copilot all the things' mania takes hold in Redmond
AROS turns any PC into an Amiga with USB-bootable distro

Re: Tears of nostalgia
I'm under the sweet illusionconvinced that Solaris (Sun, not Orc Cackle) has been the pinnacle of OS GUIs. And dread firing up my trusty, old SparcStation 4 to indulge in nostalgia, fearing it might be just that: dated.
@robinsonb5, indeed, with my beloved Sun OS there was never a moment when I wondered "did I actually click this or miss it?" and clicked again and waited - repeatedly. Build for processing performance and yet it's responsiveness beat any OS built for GUI performance.
Techies thought outside the box. Then the boss decided to take the box away
Russia expected to pass experimental law that tracks foreigners in Moscow via smartphones

Re: Shhh! Do not tell Trump
If the carrier didn't know to which base station your phone is connected to (aka your location), you wouldn't receive phone calls nor data. But that doesn't mean that Google knows your location. The question is, with whom and under which circumstances does the carrier share this data.
Dilettante dev wrote rubbish, left no logs, and had no idea why his app wasn't working
Qatar’s $400M jet for Trump is a gold-plated security nightmare
As US vuln-tracking falters, EU enters with its own security bug database
DOGE worker's old creds found exposed in infostealer malware dumps
BOFH: HR tries to think appy thoughts

This reminds me uncannily of a certain colleague - I shall refrain from naming names.
Discussion between unnamed colleague and killjoy me: "We should use AI!" - "Why would we do that?" - "We could use it in our audits!." - "In which way would this be helpful?" - "We could use it for analysing evidence." - "Right. Who's going to train our AI? We certainly cannot send any audit evidence to any third party." - "Ah, okay. But maybe we could do a pilot?" - "?! Same problem." - [more cringe exchange] - "Maybe we can ask AI how AI can help us?"
People find amazing ways to break computers. Cats are even more creative
Trump promises protection for TikTok, for which he has a ‘warm spot in my heart’
Teens maintained a mainframe and it went about as well as you'd imagine
BOFH: The Prints of Darkness pays a visit

They then partially implement every networking protocol known to man, badly, and release the thing to market.
Ever since this MFP entered our place, I tried to make it connect to the rather secure WLAN. But it wouldn't. It's not ancient either. Had to create a dumbed-down guest network just for this bloody printer. Yes, I'm looking at you, HP.
Need a Linux admin? Ask a hair stylist to introduce you to a worried mother

Have you found a talented techie through unconventional means?
Not on the level of Finn. But I had found (not hired by me, but had to collaborate with) highly untalented stuff (sic!) who had all the right qualifications (still wonder how he passed that BSc) but was completely use- and clueless.
Brit soldiers tune radio waves to fry drone swarms for pennies
Uncle Sam kills funding for CVE program. Yes, that CVE program

Re: Reasoning
Trumpistan: Set the house on fire. Steal someone's water, claim to extinguish the fire, give water to cronies to water their plantations. Whine that all the others are free-riding on your extinguishing skills. Extort money from the others. Give loads of public money to cronies to deliver some petrol ("gas"). Sprinkle the petrol on nearby houses. Praise youself for being the great (or is that "biggliest"?) leader. Stuff your gob with more burger. Crap panties and repeat.