Down to 4.9 Stars for No Reason?

[nickkold]

3 minutes ago, domenikbrenner said:

AI looking over chats and analysing them for our seller score is a bad idea.. Even if a human looks over it, it doesn't really tell how the client was picked up and made understood or how they felt about it. Sometimes a bit of tough love is necessary and even expected, at least I can imagine if someone hires Vickie for her literary criticism, which is literally that: My text is bad and I want it to become better, so you need to tell me what's wrong and fix it for me.

Wait what is happening? Is Fiverr using AI to scan our chats with the client and use this information for what? Determine our score/rating or? Can someone link me to this information or help me understand? 

[domenikbrenner]

Just now, vickieito said:

The only inkling that I have that AI might be doing this is in Fiverr's "Success score, explained" article where it keeps referring to "tone" when advising sellers how to communicate with buyers (e.g., polite and friendly). But let's hope that Fiverr isn't that reliant on AI...

We can only hope, but I am afraid, we will never know for sure. I remember we have been told many things, that just turned out to be wrong. For instance, how the customer support can negate the effects of an unjust cancellation so that it doesn't affect our profiles, and that turned out to be wrong and it did affect us. Or how private negative reviews don't affect your profile or visibility (impressions). 

I hope Fiverr will work on the obvious issues and give us the transparency we asked for and thought we finally received. But now I have even more questions than before the newest changes.

[domenikbrenner]

Just now, nickkold said:

Wait what is happening? Is Fiverr using AI to scan our chats with the client and use this information for what? Determine our score/rating or? Can someone link me to this information or help me understand? 

It is unconfirmed, but it is a possibility that was very lightly hinted at. Vickieito mentioned it. Please don't take it as a fact yet, it's pure speculation. But if it were true, it would be a bad idea.

[emmaki]

This is Google's NLP demo. I've posted it here before:

https://cloud.google.com/natural-language?hl=en

About halfway down there's a test (Natural Language API demo) thing you can use to analyze text for various things. Anyway, when you run the test, some tabs will show up and one says "sentiment" (green = positive yellow = neutral red = negative). 

Now, I don't know what AI Fiverr is using to judge tone of course, but it will likely be similar to this. Try running some of your old lit crit convos through here and see what happens. 

EDIT:Also check out the category tab. That's basically how the AI interprets whatever you wrote. Although tbh I don't really understand that tab. The scores below are for the rather bland demo that Google kicks off with...

Edited March 9 by emmaki

[domenikbrenner]

10 hours ago, graveyardshiftx said:

Welcome to Fiverr 2.0, where 5 means 3, 3 means 5, yes means no and "I love my new logo" means "You suck at your job".

Basically, this new rating system is shrouded in mystery and it's gonna stay that way because it's in Fiverr's interest to do so. Now they can just blurt out "Because algorythm!" whenever something like this happens and since we're not privvy to that information, all we can do is... well nothing. We can do nothing.

I'm beginning to suspect that this "algorythm" is in fact 1 part AI, 1 part private feedback and 3 parts unicorn queefs because it makes absolutely no sense.

I'll print this out and frame it.

[maitasun]

9 minutes ago, vickieito said:

We can see a little more data, but we can't see our scores in all six areas (only 2-3 key areas). For the others, we don't know how those key areas are rated by the AI (i.e., strong negative impact, negative impact, room for growth, positive impact, or strong positive impact). And even further, we don't know how AI breaks down each of those ratings.

I insist Fiverr is throwing things at the wall to see what sticks. There's no consistency whatsoever.

Just as an example: The gig I posted above showing "room for growth" in the communication area, and "positive impact" in the order cancellation area, used to show "strong positive impact" in the order cancellation area and nothing else. So there, "used to show": past tense of the verb "not anymore". 

It's admirable to see how Fiverr pulls things out of its hat. Pure magic, and we all love magic. Don't we? 🥰

[emmaki]

3 minutes ago, maitasun said:

It's admirable to see how Fiverr pulls things out of its hat. Pure magic, and we all love magic. Don't we? 🥰

Do you mean this kind of magic? 

(It helps if you think of a different hole)

 

[domenikbrenner]

Just now, maitasun said:

I insist Fiverr is throwing things at the wall to see what sticks. There's no consistency whatsoever.

Just as an example: The gig I posted above showing "room for growth" in the communication area, and "positive impact" in the order cancellation area, used to show "strong positive impact" in the order cancellation area and nothing else. So there, "used to show": past tense of the verb "not anymore". 

It's admirable to see how Fiverr pulls things out of its hat. Pure magic, and we all love magic. Don't we? 🥰

I go with @graveyardshiftxand say it's on purpose to keep us in the dark and this is just a pseudo update to stop revolts.

I am sorry, but obviously there is some system behind, why is it so difficult to give us insight? It's more work to show us "wrong" and "buggy" output than just highlight how it works.

All we want as sellers is understand how we are rated based on what and which things affect us positively and negatively. It doesn't mean we are suddenly going to break the system, because we still have to make clients happy, right? But yet we are kept in the dark here.

Not happy with this. Everyone has got something about this new system that they just simply don't understand because it is inconsistent with what's being said AND what happens to others in the same situation.

I personally think they shouldn't publish the new system when it doesn't work properly yet.

[emmaki]

All Fiverr's gonna do is change about 3 minor things (tbf they already did that by changing 3 words) and then direct everyone to scream at the CS templating machine. 

Maybe a little pep talk about transparency in the mix somewhere. 

[domenikbrenner]

Just now, emmaki said:

All Fiverr's gonna do is change about 3 minor things (tbf they already did that by changing 3 words) and then direct everyone to scream at the CS templating machine. 

Maybe a little pep talk about transparency in the mix somewhere. 

CS Templating machine is spot on.

[mandyzines]

10 hours ago, graveyardshiftx said:

Now they can just blurt out "Because algorythm!"

...which enables the dodging of accountability even when a human makes the final call, although I suspect that's not the case here.

It's happening outside of Fiverr with far more dire consequences, like, oh people not being allowed to live anymore.

But, AI's so....sexy. It's such a cool tool that will make our lives easier. We simply adapt.

[domenikbrenner]

2 minutes ago, mandyzines said:

...which enables the dodging of accountability even when a human makes the final call, although I suspect that's not the case here.

It's happening outside of Fiverr with far more dire consequences, like, oh people not being allowed to live anymore.

But, AI's so....sexy. It's such a cool tool that will make our lives easier. We simply adapt.

Yeah people get lazy and think AI will do the job just fine. The reality is: AI is not as smart as people think it is yet. That is a dangerous mentality, really. 

[maitasun]

4 minutes ago, mandyzines said:

But, AI's so....sexy. It's such a cool tool that will make our lives easier. We simply adapt.

Yeah, I guess we'll simply adapt to writing boring wills just to make our magnificent and smart AI happy in the hope that it will look at us with loving eyes and keep our rating safe, even if we annoy our buyers.

[mandyzines]

44 minutes ago, domenikbrenner said:

Yeah people get lazy and think AI will do the job just fine. The reality is: AI is not as smart as people think it is yet. That is a dangerous mentality, really. 

Well, there are the toys we're allowed to play with and other toys that play with us, but the whole concept of 'AI' is being publicized as sexy and necessary, especially in terms of governance.

This is semi-related, but Hawaii just announced each island will be installing new SMART technology in their (my) electric grids for the sake of predicting and mitigating fires caused by climate change.

Yet, in the same grand announcement, they told us this technology had already been installed and in use in the state. I could write several pages about Lahaina, as I've been following it very closely since 'the event' happened, but the heaviest of themes has been lack of transparency and accountability. We've no solid answers and a lot of what's given formally is conflicting. So far, their only way of publicly addressing it is to throw more AI on it.

It feels eerily similar to what's happening here.

At least here, ...there's no one literally blocking the only escape routes out.

Edited March 9 by mandyzines

[domenikbrenner]

2 minutes ago, mandyzines said:

Well, there are the toys we're allowed to play with and other toys that play with us, but the whole concept of AI is being publicized as sexy and necessary, especially in terms of governance.

This is semi-related, but Hawaii just announced each island will be installing new SMART technology in their (my) electric grids for the sake of predicting and mitigating fires caused by climate change.

Yet, in the same grand announcement, they told us this technology had already been installed and in use in the state. I could write several pages about Lahaina, as I've been following it very closely since it happened, but the heaviest of themes have been lack of transparency and accountability. We've no solid answers and a lot of what's given formally is conflicting. So far, their only way of addressing it is to throw more AI on it.

It feels eerily similar to what's happening here.

I think AI will be a wonderful thing in the future, but god knows why chatgpt became so popular, and now suddenly the whole world thinks it's like in the Science Fiction movies.

"Computer. Analyse human behaviour of chicago and calculate which suburban street is most likely to be attacked by terrorists in the next 3 hours, we will put our guards there and save the day" "Calculating ..... According to public data from the last 25 years, I have identified 23 hotspot locations in this sector *displays satellite image with perfect grid*"

In reality:
"Computer, in this sentence; "The house on the hill has three windows", how many words start with an "h"?"
"6"
"That is wrong, computer"
"You are right, please excuse the misunderstanding. I might have understood your question wrong. If you count the words that start with H you will find that there are 6 words. I hope that helps"
"No, that is still wrong. The answer is 3"
"You are right. The answer is 3."
"Do you know which ones?"
"I am sorry, I am a simple AI and I cannot access the internet for further information. My current knowledge is based on events up until august 2022."
"I literally just told you."
"Please excuse the misunderstanding. Indeed you just told me."
"So which words start with an h?"
"The words containing the letter h are "the", "house", "the", "hill", "has", "three". I hope this information helps you. If you have any further questions, please feel free to ask." 

 

[jefftuts]

 

Looks like reviews for buyers are now showing 3 decimal places!

[mandyzines]

32 minutes ago, domenikbrenner said:

I think AI will be a wonderful thing in the future, but god knows why chatgpt became so popular, and now suddenly the whole world thinks it's like in the Science Fiction movies.

ChatGPT is such a small component of all the technology that's being clumped together into the term "AI." I think of living in a science fiction movie, too, when I see that, for instance, a certain military is using a system to analyze large amounts of data to determine who/where to airstrike--right now, as I type this.

Edited March 9 by mandyzines

[domenikbrenner]

4 hours ago, mandyzines said:

ChatGPT is such a small component of all the technology that's being clumped together into the term "AI." I think of living in a science fiction movie, too, when I see that, for instance, a certain military is using a system to analyze large amounts of data to determine who/where to airstrike--right now, as I type this.

It is right that there are way smarter AIs out there than chatgpt, which, in my opinion, is one of the dumbest ones. The only thing chatgpt is doing better than other AIs is the creative writing part. 

But the point is that everyone just wants AI these days, and the good ones are not public, the rest will be chatgpt clones. It's all carbon copies. And that's the world we live in.

[emmaki]

I mean the internet did evolve from military technology (ARPANET).

Regarding AI, there was a strange story from OpenAI around the time that Altman got fired/rehired in the space of a week as Bill Gates floated omnipresently and mysteriously around the whole affair. Anyway, researchers got very nervous and fidgety about a project they were doing. Lots of speculation in that article, but the people working on it seem anxious enough to leak it while the Fearless Leaders are dismissing concerns and saying it's great. Sound familiar? We still don't know what Q is. One thing we do know is that while programmers and software engineers can control non-AI technology through coding, we still haven't really worked out how to control AI the same. It's also worth noting that LLMs are not meant to be good at math (another thing entirely) which is why its easy to break them with basic calcs. 

It's also a given that this technology has largely been nerfed. GPT 4.5 and the upcoming 5 are no doubt better than past versions, but they are getting worse for users because of all the restriction put on the retail models. Some are saying that 2024 is going to be the year of downloadable LLMs where there are no restrictions where you don't get "sorry I'm not allowed to do that" responses. 

Corporations (and government) will have access to much better tools than regular people. The problem before was that AI was too good for the normal person, which threatens basic systems of power and control in our global society. So now AI is being dumbed down for the masses while simultaneously the bigger budgets of government/corporation allow them to access more powerful models that may not be (as) nerfed due to them being, well, higher on the power structure than the average meatbag. But that's the problem: these organizations might have power, but do they use that power responsibly and with consideration for their users' needs? You only need to look at Fiverr for an example of the reality - corporations are designed for profits, not people. But who is watching the corporations' use of AI? Right now, there are no watchers. That will change soon in the EU. This is an unregulated hype train with plenty of enormously disturbing undertones. 

The speed of AI evolution cannot be ignored though - the video stuff from OpenAI (Sora) is very impressive stuff when you consider that this time last year, AI-produced videos were terrible. Remember the end of Raiders of the Lost Ark where the baddies open up the ark and the mist comes out and hilariously melts their faces? That's exactly what AI videos were like. They're a lot better now. 

That's in one year. 

5 hours ago, domenikbrenner said:

The only thing chatgpt is doing better than other AIs is the creative writing part. 

Eugh, no! In the fast-paced world of creative writing, all ChatGPT can do is delve into a plethora of words and navitage to the same few tired ones that give its presence away everytime (unless you prompt well). Claude AI is much better. Or it was until 6 months ago, when the EU meant I couldn't use it anymore. Claude had/has its own problems, but it's a better writer than ChatGPT. 

This Y Combinator link has some links to some great stories about OpenAI. Highlights

- "OpenAI is not a technology company" 

- Suspicious touches of Theranos/Elizabeth Holmes-like behavior:
 

Quote

 Sutskever led employees in a chant: “Feel the AGI! Feel the AGI!” The phrase itself was popular enough that OpenAI employees created a special “Feel the AGI” reaction emoji in Slack.

If you've watched the Netflix series (The Dropout) on Theranos, you'll no doubt remember that it was the labrats who first realized the dangerousness of the work they were doing. Which brings us right to the top of this article. Who knows what the always-philanthropical Mr. Gates' interest is in this technology? 

So how do Fiverr employees feel about this tech? About the advancements in it? About how it is affecting users? They can't tell us, but they're a company like any other, stuffed with people. Some of them probably have grave doubts privately, but what can they do? 

Bonus credits go to anyone who takes the time to read E.M. Forster's short story The Machine Stops (1909). He didn't just create dull Edwardian romances where repressed English people go through the agony of my emotions! 

I sometimes wonder what amazing technology is hidden from us. But it hardly matters: we're already one foot in a neo-feudal society run by big tech and despite "the people" generically not really liking the direction any of this is going in, we have little choice but to wrap up and follow. 

All I can say is that it won't be pretty when everything snaps. I haven't even touched on 4IR, the addictive nature of so many of the elements of modern tech (see also: gamification from Fiverr) and how doctors are starting to warn gen z that they are (as a cohort) going to get arthritis insanely early due to being addicted to their tech. The evidence is already coming out through studies. Then there's the rise of autoimmune disease. For all the talk of techonology and the modern world, it's pretty clear that the design of technology is making us sick mentally (anxiety is one of the related issues) and physically. Most big tech leaders don't let their kids use smartphones. That alone should have been a warning to the parents of the iPad generation.

Still, it's nice to see that Fiverr reviews, right at the end of testing, are now testing a third number after the decimal point. We really are the lab rats trapped in a maze of experimentation run by people who have no real idea of what they are doing and who are empathically not listening to their userbase.

/sunday soapbox

Edited March 10 by emmaki
added name of netflix series. will not edit anything else. deal with typos and half-finished thoughts.

[melissaharlowvo]

On 3/6/2024 at 7:49 AM, ssj1236 said:

So, I've thankfully had quite a few good ratings recently with no negative experiences yet today I went from a 5.0 rating to a 4.9 rating......what happened? I'm genuinely confused. Fiverr, please, I don't have the stamina to keep up with your mood swings anymore. 

Same here. I noticed all the sellers on the first page have also had their rating dropped. Very few 5 star sellers now. Definitely frustrating and confusing.

[marketingpro_3]

Fiverr added two new rating metrics "Quality of delivery" and "Value of delivery" which will ensure that every seller now will have a 4.9 in their overall rating instead of a 5. which will end up changing nothing.

[creative_kad]

i was dropped to 4.9 stars from 5 last few days and now it is 5 again 🥹

[designelitely]

23 minutes ago, creative_kad said:

i was dropped to 4.9 stars from 5 last few days and now it is 5 again 🥹

Same thing happened with me,too..... 🙂