jumpingbud.bsky.social
Retired engineer/programmer, linguist. Childless cat person.
41 posts
132 followers
282 following
Regular Contributor
Active Commenter
comment in response to
post
And if it's not a free account, cancel your subscription and make it free again. Paying for LinkedIn is almost as bad as paying for Twitter.
comment in response to
post
Before you delete the account, verify with LinkedIn that if you delete it no one else can reuse your username/handle and then impersonate you. If it can be recreated, delete the data you don't want there and leave it dormant.
comment in response to
post
These five people: bsky.app/profile/indi...
comment in response to
post
And of course he really will say "sell us your country" because he doesn't know it's part of Denmark. UGH.
comment in response to
post
This has to be a stunt and never intended to actually fly like that. Depending on the size and shape of the lights, the added drag could be considerable. If the lights and the wiring are not held in place firmly, so many things could go so horribly wrong. Consider what ice on wings can do.
comment in response to
post
That's for sure! If she really does have children, I bet they have some stories to tell. When they're old enough maybe we'll hear some of them.
comment in response to
post
Yours is an amazingly short yet extremely accurate summary of the whole situation!
comment in response to
post
Of course you're assuming that such companies are doing this on purpose when in reality, hating that their employees are happy working from home, they're just floundering around trying to take more control back and causing themselves bigger problems. The king, the mice, and the cheese strike again.
comment in response to
post
Here is a link to the full grand jury report. For the "father" McCaa stuff, go to page 79: www.bishop-accountability.org/reports/2016...
comment in response to
post
It's not just more attention. The attention is also more positive rather than all negative. As another responder said, if this reduces the number of school shootings, that's a good thing, I guess.
comment in response to
post
Do you mean 1) "who are you telling this to" or 2) "who are you to be telling me this"? If 1) to anyone reading your thread who might find it helpful and if not no big deal. If 2) I'm just a retired, old fart engineer/linguist born too early and missing out on so much stuff developing so fast.
comment in response to
post
A few last thoughts. As a linguist and computer programmer, LLMs are truly amazing. I never thought I'd see anything like this in my lifetime. But as far as AI goes, at their present state, they're still mostly the A without much of the I.
comment in response to
post
..."simple" language manipulation devoid of real thinking. All output from these LLMs must be taken with the biggest block of salt you can find and work from there. Use them for what they're good at and then do your own post-processing. Otherwise you use them at your own peril. 8/8.
comment in response to
post
So until/unless they add sufficient post-processing that can accurately detect which sources might be the root of the result, they can't even properly cite their sources. To us humans it's truly a totally foreign method of "thinking" and writing. It's less thinking/writing and much more... 7/?
comment in response to
post
...the sources are no longer available in the strict sense. They don't have the PDFs of sources in their memory. What they have is the entire corpus, each source disembodied and pre-synthesized into one giant blob. Give that blob an input and it predictively pumps out the most likely response. 6/?
comment in response to
post
If they regurgitate too directly from a source without citation, then just as with human writers, yes, it's plagiarism. But they don't write like we do. We are generally aware of the sources we are using and then synthesize our writing through our thinking. In their case... 5/?
comment in response to
post
Since everything they generate is based entirely on what they've internalized from the training corpus (I can't call that true learning yet) and only using probability to rearrange it to an output, they really don't add anything of their own thinking to it. But does that make it plagiarism? 4/?
comment in response to
post
Both those posing questions to them looking for answers they can trust and those asking questions about their capabilities are asking the wrong questions because we assume they have the other 90%+ (a total guesstimate) thinking power and knowledge of an actual brain. 3/?
comment in response to
post
While some knowledge is embedded within language itself, the vast majority is not. These models predictively push words around and generate a lot of apparently high quality language but just as often could create high quality invented garbage and outright lies. It's all based on probabilities. 2/?
comment in response to
post
Because LLMs are capable of producing such high level language, we overestimate their actual thinking power but they don't think. They mimic one facet of the human mind (likely not in the same way), language input and output but without the majority of the real thinking a brain does in-between. 1/?
comment in response to
post
Drastic power vacuums open the door to MANY possibilities. Unfortunately, most of the possibilities SUCK as bad as (or worse than) what got removed. We can only hope the Syrians make good choices. Keep your hopes high but your expectations low to minimize your likely disappointment.
comment in response to
post
Gurl, tell that to Oliver.
comment in response to
post
I always thought her name was margarine trailerpark grease.
comment in response to
post
Apologize first to Shocking Blue:
youtu.be/8LhkyyCvUHk?...
comment in response to
post
How much worse was it compared to how "Lost" ended?
comment in response to
post
That seems too mild. Maybe more like Caligula and the madness of king George together?