Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror

Comment Re:Very cool... [Super autopen!] (Score 1) 53

Mod parent funny. But so far the joke seems to have sailed right over the heads of the moderators.

My new "tech joke" attempt involves a different kind of computerization. It's a conspiracy joke about the bruised hand. The YOB can't sign his own name now, but he can't admit that he's using Joe's autopen, so they made a robot arm that clamps over his arm and fingers and guides them in signing documents. It clamps on a bit too firmly, but they had to focus on loyalty and secrecy over competence (as usual) as they were building the device in a secret closet somewhere.

The signing ceremonies? Faked. The documents are signed before he even opens them. If they want to pretend they are showing a camera angle with the pen in action, then that's being handled with AI video (AKA AI slop) now.

Comment Re:Wrong. No soup or money for you, Mozilla. (Score 1) 74

Pretty sure I tried that but failed. I actually suspect it is some kind of manual problem related to minimizing FP abuse?

In my list of dimensions of evaluating generating AIs, I forgot three of the most negative ones. Criminal profitability is pretty obvious, though the crooks don't show their profits so you'd have to steer backwards by assessing their efforts, presumably correlated with the most profitable scams. Another negative dimension is related to "best for generating AI slop". The third that comes to mind is "best support for self harm". And I'll probably think of a couple more as soon as I click the Submit button...

Comment Re:Another part of the story. (Score 2) 272

I think you mean "effective opposition", but that's along story.

In solution terms, perhaps each YOB supporter should be obliged to talk to a couple of "nice" foreigners long enough to learn how much disgust the YOB is producing everywhere in the world. I think three minutes should suffice.

Comment Wow, a troll implosion (Score 1) 118

And in the process of feeding a troll and propagating a vacuous Subject, too.

Funny story time. I knew one of the seminal trolls many years ago. Perhaps he thought he was a flame warrior? When I first encountered him he was actually a master's student in the same department where I got my CS degree. He just loved being a nasty offensive and trollish person, apparently for the sake of being nasty.

Later on I heard that he was working in a fast food place, perhaps flipping burgers. Hard to imagine a person with his personality at the counter. Later I heard he had died. At that time I actually wanted to discuss his history in terms of "lessons to be learned", but the reaction was "We mustn't speak ill of the dead." (Which may remind you of one of the YOB's latest gaffes.)

Comment Wrong. No soup or money for you, Mozilla. (Score 1) 74

Yeah, I'm going for Funny again. And failing. As usual.

But if I had been asked which feature I WANTED to pay for, then disabling the AI is much more likely the feature I would donate for than the feature to add "AI" to Firefox in the first place. If I WANT AI, then I know where to go looking for it. I've "played" with lots of the so-called generative AIs--and so far I hate them all and I am currently unable to imagine the "project pitch" that would persuade me to donate for that feature.

But maybe they could ask the AI to write such a persuasive proposal that I would donate to pay for adding AI. However my counterproposal is that you don't hold your breath while waiting for my donation.

I'm still looking for a useful comparison of the AIs. For example, on the dimension of best apologies, I think Copilot is first with Google's Gemini running second. On the dimension of most sycophantic, I think DeepSeek is the leader but it's a close race for second among several candidates. Then there's the dimension of programming support, where I actually like ChapGPT. Has anyone tried to assess the various AIs on their hallucinations? I'm inclined to give that prize to Gemini, but perhaps only because it's always in my face when I visit the formerly nice Google search. But on the dimension of utility so far it seems to be an aggressive race for last place.

[And what's with the graying out on Slashdot? When I first saw the story it had only three comments but was unclickable. I'd think I should know after so many years, but...]

Comment The ACM needs a viable business model (Score 2) 20

I don't see how this is sustainable, and I was a dues-paying member of the IEEE Computer Society for about 20 years and of the ACM for about 10 years. I even attended meetings and donated quite a bit of time to the CS. Minor writing and enough refereeing to become a "senior referee" at the end. I think they are doing some important stuff, but there are costs... And eventually I stopped paying dues. (But now I'm also remembering a database conference that may have been paid for by my employer.)

By the way, I used to read the magazines I received cover to cover. That time actually became a significant negative factor. Lots of good stuff, but too much time required. On that front I think the main effect of paperless publishing will be to significantly reduce the incentive to read all of an entire issue... Why not just ask an AI to summarize the parts that are most relevant to my work?

So if you're going to push me for an overall assessment, I think it's a net negative and will make the ACM less relevant. Perhaps even imperil it's survival.

But I also have a solution approach to ignore: What if the ACM supported books with special webpages to address the time problem? Each computer-related book would have some QR codes pointing to the errata, a bibliography, a searchable and dynamic index, and even forward links to later work on related topics. Kind of a post-publishing future bibliography? In this fantasy, at least the publishers would be providing some funding to sustain the relevance of the books they are selling.

Comment Re:The spammers LOVE money (Score 1) 20

I think you have lost track of the current priorities of the DoJ. But perhaps time for the old joke? Unfortunately I can't remember the exact form of the quote that underlies the joke. A Russian guy? Something about "Behind every great fortune there is a crime"?

I actually think there are two ways to become excessively rich, and crime is only one of them. The second way is to be lucky, though the luck can take various forms. Most common form is inheriting the profits from a parent's crimes.

So the joke is that the old saying is obsolete because these days you use some of the loot to bribe the cheapest politicians to fix the law so that it's no longer a crime. Investing in politicians has had an obscene RoI for a long time now.

Comment Re:No surprise[s in today's SF?] (Score 1) 131

Now I think you're mostly referring to the publishers. Funny business that, even before Amazon tipped over the table. Most of the books they publish are failures that don't even recover the cost of the first print run. I've heard numbers from 80% to above 90%, but that was a while ago, and largely from a delivery driver who delivered fresh books while picking up the unsold ones. The profits were entirely from the bestsellers, and of course the publishers love opium-like books for that reason.

I acknowledge that things have probably changed now. I've already mentioned the Amazon problem, but I also think AI is offering new ways to assess which books might become bestsellers (resulting in fewer books making the first cut).

Comment Re:Friction free engine accelerating to infinity (Score 1) 106

I'm trusting Michael Lewis on this, but I don't think that applies to the members of the exchanges, and the high speed traders are members. The fees for trades are limited to the little suckers like you and me.

Well, also I've read some of the proposals for transaction fees, and none of them seem to make sense unless the current transactions are without fees.

Comment But is this Korean company more evil than Amazon? (Score 1) 32

Fishing for Funny in the dark. Pretty sure I didn't get there, but also expressing my disappointment than no one else got there first.

Getting away from funny, but 'modern capitalism' is supposed to be based on a kind of adversarial model. The companies want to sell us as much stuff as possible with the highest profits, while we are supposed to be trying to find the best values to force the companies to offer better products at lower prices. But the powers are not balanced in this 'game'. Individuals are acting alone and mostly in ignorance, while the companies continue to become larger and increasingly powerful. From this perspective, collecting customer information is like ammunition for tomorrow's attacks on the customers' credit cards.

Slashdot Top Deals

He's like a function -- he returns a value, in the form of his opinion. It's up to you to cast it into a void or not. -- Phil Lapsley

Working...