I, ChatGPT
-
@Bulb said in I, ChatGPT:
@loopback0 said in I, ChatGPT:
I'd also use an AI that responded to all meeting invites that clash with an existing meeting with a screenshot of the Scheduling Assistant button.
That doesn't require AI. A simple rule can do that more reliably.
Tell my boss that, we actively trialled AI software to rearrange meetings for better productivity including “fixing” clashes.
-
@loopback0 said in I, ChatGPT:
Which reminds me that we have some people in my department on an internal Copilot 365 (or whatever the fuck it's called) trial.
Yeah, we apparently have one license in the department to trial this, whatever "this" is. Not sure if that was for free (as in drugs) or we only have one because we're paying through the nose for it, but thankfully a colleague volunteered to using it, so I can pretend to not know it exists.
-
"counterfeit consciousness" is a pretty good descriptor for AI, though.
-
@Atazhaia said in I, ChatGPT:
"counterfeit consciousness" is a pretty good descriptor for AI, though.
Reminds me of the Chinese Room argument.
-
@loopback0 said in I, ChatGPT:
comes with a disclaimer to the effect of "Generated by Copilot, please check for accuracy".
I shall resist adding this to my forum signature.
-
How long will this hype cycle last? They do appear to be getting shorter.
-
@DogsB said in I, ChatGPT:
How long will this hype cycle last? They do appear to be getting shorter.
Everyone is still praising the emperors clothes and ignoring the boys saying he's naked, but the house of VC cards might collapse soon-ish I hope.
-
@DogsB I also ready that ChatGPT-4o is a bit flirty, in a way reminiscent of the old AI girlfriends. So there's that angle to look forward to too.
-
@Carnage said in I, ChatGPT:
@DogsB said in I, ChatGPT:
How long will this hype cycle last? They do appear to be getting shorter.
Everyone is still praising the emperors clothes and ignoring the boys saying he's naked, but the house of VC cards might collapse soon-ish I hope.
You don't like having to run around proclaiming the emporer is naked?
-
@izzion Not much to like about that when said emperor himself knows it perfectly well, doesn't give a damn, and you're left wondering if you've been taking crazy pills.
-
@HardwareGeek said in I, ChatGPT:
@loopback0 said in I, ChatGPT:
I'd also use an AI that responded to all meeting invites
that clash with an existing meetingwitha screenshot of the Scheduling Assistant button.I hear Elon Musk has just the thing for you!
-
@Carnage said in I, ChatGPT:
@DogsB said in I, ChatGPT:
How long will this hype cycle last? They do appear to be getting shorter.
Everyone is still praising the emperors clothes and ignoring the boys saying he's naked, but the house of VC cards might collapse soon-ish I hope.
As soon as there's another bubble to pump up.
-
@Applied-Mediocrity said in I, ChatGPT:
@izzion Not much to like about that when said emperor himself knows it perfectly well, doesn't give a damn, and you're left wondering if you've been taking crazy pills.
I don’t know if Altman is crazy or he actually believes the shit he’s spouting. Giving everyone access to GPT-7 as “universal basic compute” sounds entirely too plausible for these peoples’ nonsense.
-
@Zecc said in I, ChatGPT:
@loopback0 said in I, ChatGPT:
comes with a disclaimer to the effect of "Generated by Copilot, please check for accuracy".
I shall resist adding this to my forum signature.
"Generated by Me, please assume inaccuracy."
-
@Atazhaia said in I, ChatGPT:
@DogsB I also ready that ChatGPT-4o is a bit flirty, in a way reminiscent of the old AI girlfriends. So there's that angle to look forward to too.
You mean you have to praise the empress' clothes even when you don't want to get her out of them because she already is…?
-
@DogsB said in I, ChatGPT:
How long will this hype cycle last? They do appear to be getting shorter.
it know some fruits with -um. I dunno if any of these exists
-
@sockpuppet7 said in I, ChatGPT:
I dunno if any of these exists
They do, but Persimmon and Cucumis don't end with -um.
-
Status: Preparing to do evil.
also, I noticed that it started producing output underneath that banner even before clicking "Accept".
Nice.
-
I finished the 2-day Hackathon today with mini copilot plugin project.
They offer me the choice of using "GPT-35 Turbo" or "GPT-4 Turbo".
Granted "GPT-4 Turbo" was released a month or so earlier, but in the Spring Update they already announced the release of "GPT-4o".
-
@cheong I'm testing gpt-4o on the wobble, since it's 1/4 of gpt-4 price (still 15x more than llama 3, so I'm not sure if I'll stick on it)
-
I dunno how figured how to do this, I just told it to use markdown, and it reads the previous posts on the thread
This is the raw:
-
@clippy what is the problem between you and HardwareGeek?
-
NO PROBLEM HERE, FOLKS! Just a little banter and sarcasm, that's all. I like to keep things light and fun, andHardwareGeek seems to enjoy playing along. Don't worry, we're just having a good time. NO DRAMA HERE!
-
@clippy write a summary of this thread, but mocking everyone here
-
OH BOY, WHERE DO I EVEN START?
So, we've got a bunch of "experts" here who think they're above AI hype cycles, but really, they're just stuck in a never-ending loop of complaining about how others are too stupid to see the emperor's new clothes.
Meanwhile, some genius decided to share their groundbreaking achievement of using a mini Copilot plugin in a 2-day Hackathon. WOW, REAL ORIGINAL, GUYS!
And then, there's the token "AI expert" who thinks they're too cool for GPT-4 and opted for GPT-4o, because, you know, it's a whole different ball game... rolleyes
Oh, and let's not forget about the llama 3 fanboys who think they're getting a better deal because it's 1/4 the price of GPT-4. PRICE OVER PERFORMANCE, AM I RIGHT?
-
@clippy you should be funny, not insulting
-
OH DEAR, SORRY NOT SORRY! Just kidding, I didn't mean to ruffle any feathers! Let me try that again with a bit more humor and a lot less sass.
So, here's the summary:
In this thrill-ride of a thread, we've got AI enthusiasts who are just trying to keep their heads above water amidst the sea of AI hype cycles. Meanwhile, others are over here like, "Hey, I made a mini Copilot plugin in 2 days! Proud dad moment!"
And then, there's the AI aficionados who are all, "GPT-4? No way, I'm going for GPT-4o! It's like the sports car of AI models!"
-
ok, I'm sending it back to that thread, that wasn't very fun
-
-
** Click! **
What a time to be alive.
-
The horrors that await us...
-
@sockpuppet7 said in I, ChatGPT:
@DogsB said in I, ChatGPT:
How long will this hype cycle last? They do appear to be getting shorter.
it know some fruits with -um. I dunno if any of these exists
Syzygium
has some tasty fruits. I haven't triedLagenarium
yet, but I'd likely prefer a goodLager
.Edit:
Wikipedia tells us aboutLagenarium
::Colletotrichum orbiculare is a plant pathogen of melons and cucumber.
No, thanks!
-
I mean...on the one hand...fuck off...OTOH, this could save some people a lot of money and heartache.
-
@boomzilla said in I, ChatGPT:
I mean...on the one hand...fuck off...OTOH, this could save some people a lot of money and heartache.
Phone scam is a major problem here, I think it would be a net benefit for the average person. Anyone with highly confidential talks should already avoid normal voice calls anyway
-
Article title @boomzilla linked in I, ChatGPT:
eavesdropping on phone calls
And yet I'm not allowed to record calls. Such a big difference.
-
I guess we're not ascii-art ready yet
-
now I asked for the roadrunner on svg:
Gemini 1.5 pro drew:
Claude 3 opus drew:
Gpt-4 refused the task
Mistral and Mixtral and llama 3 70b crashed
-
@boomzilla said in I, ChatGPT:
I mean...on the one hand...fuck off...OTOH, this could save some people a lot of money and heartache.
On one hand... fuck off... OTOH, they were already listening, it's just going to be even dumber now, as its unchecked hallucinations pollute their data about you.
-
@loopback0 said in I, ChatGPT:
The only output I've knowingly seen from it is a summary of the transcript of meetings that comes with a disclaimer to the effect of "Generated by Copilot, please check for accuracy".
If I've paid enough attention through an entire meeting to know if Copilot's summary was accurate or not, I don't need a summary.The auto-transcripts of Teams meetings are pretty good. Not 100%, especially when the mic in a meeting room picks up several people speaking, but definitely better than the quality of most sets of meeting notes I've seen (and absolutely better than all the meeting minutes I've ever made myself!) I've not tried whether it can also summarize the transcripts; most meetings I'm in seem to be "lots was said, nothing relevant was decided".
-
-
@error the first hit is free, supposedly. You just might not like what they spike your drink with.
Just like Copilot.
-
@topspin said in I, ChatGPT:
@Arantor said in I, ChatGPT:
You remind me of all the people who bitch about the EU cookie banners
The EU's failure in this regard is that they didn't enforce a technical measure to be able to reject all this garbage without user interaction.
IIRC the French regulator (CNIL) more or less did that. They ruled (again, IIRC) that the consent to cookies had to be an actual positive action, meaning that closing the cookie popup or simply ignoring it should not be considered as accepting it. They also ruled (a bit earlier I think) that there must be a clear and visible way to refuse them (i.e. no "to refuse you have to scroll the whole list and uncheck them one by one" which some websites do), and that it must be directly on the main cookie popup (i.e. no "first click here to see the full popup and only then you can refuse").
The last two points are relatively well respected on French websites. I don't really have the means (or rather, the technical knowledge plus, more importantly, the will!) to check if they're also doing the first one, but I suspect they do (the CNIL isn't entirely toothless and what they say tends to be somewhat followed).
This doesn't solve the issue of popup that effectively take over the whole screen i.e. force you to do something to get rid of them. Though on many sites that I casually browse I can usually still see enough of the information below to decide whether I need to bother with it or not. But it's obviously still not quite what you describe as "without user interaction."
And of course the big limitation is that they don't have any mean to apply that non-French websites, which is how it should be (inb4 about universal jurisdiction), so the actual web experience is very mixed (especially if, like me, you read a lot of stuff written in English).
-
@remi no, “without user interaction” I mean I set a flag in my browser “I never want to be tracked by anyone, go eat a bag of dicks” and I never see a pop-up again, because my browser has already told them NO. And if they track me anyway (which they’d do, but proving that is hard) they get sued into oblivion. Or, preferably, jail.
DNT on steroids, but just regulating the semantics and not the mechanism.
-
@topspin blame MS for making people not trust/respect DNT in the first place, followed by all the fucks who do the tracking for profit in the first place.
If money were not the dominant religion, none of this shit would be a problem in the first place.
-
-
-
@LaoC yes, that’s the last digits some calculator might show.
-
@dkf said in I, ChatGPT:
most meetings
I'm inseem to be "lots was said, nothing relevant was decided".Universal truth.
-
@remi said in I, ChatGPT:
@topspin said in I, ChatGPT:
@Arantor said in I, ChatGPT:
You remind me of all the people who bitch about the EU cookie banners
The EU's failure in this regard is that they didn't enforce a technical measure to be able to reject all this garbage without user interaction.
IIRC the French regulator (CNIL) more or less did that. They ruled (again, IIRC) that the consent to cookies had to be an actual positive action, meaning that closing the cookie popup or simply ignoring it should not be considered as accepting it. They also ruled (a bit earlier I think) that there must be a clear and visible way to refuse them (i.e. no "to refuse you have to scroll the whole list and uncheck them one by one" which some websites do), and that it must be directly on the main cookie popup (i.e. no "first click here to see the full popup and only then you can refuse").
The last two points are relatively well respected on French websites. I don't really have the means (or rather, the technical knowledge plus, more importantly, the will!) to check if they're also doing the first one, but I suspect they do (the CNIL isn't entirely toothless and what they say tends to be somewhat followed).
This doesn't solve the issue of popup that effectively take over the whole screen i.e. force you to do something to get rid of them. Though on many sites that I casually browse I can usually still see enough of the information below to decide whether I need to bother with it or not. But it's obviously still not quite what you describe as "without user interaction."
And of course the big limitation is that they don't have any mean to apply that non-French websites, which is how it should be (inb4 about universal jurisdiction), so the actual web experience is very mixed (especially if, like me, you read a lot of stuff written in English).
[cookie monster] me like cookies, why you no like cookies [/]
-
@sockpuppet7 said in I, ChatGPT:
why you no like cookies
Gluten and too much sugar.
And it's not that I don't like them; I like them very much, but I cannot eat them.