ChatGPT (sigh) the fastest-growing web app in history (sigh) claim analysts Last November, OpenAI launched ChatGPT as a free web interface and took the internet by storm. Data compiled in a study by UBS reported the chatbot had managed to reach 100 million monthly active users by January, which would make it the fastest-growing consumer app in internet history. The study cited data from Similarweb, an …
How about Internet Explorer?
"...cannot recall a faster ramp in a consumer internet app,"
Internet Explorer is an internet app that ramped, a grand one at that.
I remember in the fall of 1999 that supposedly over 500 million globally started "using the internet" in about 3 months... started using Internet Explorer that is. I guess Doomsday + Internet Boom made for a noticeable outlier condition.
I'm sure that it will be included in your 'Orifice 36and a bit' subscription very soon. Then with no way to delete it (reinstalled with every update) and kerching, suddenly they make it another $49.99/month/seat and they laugh all the way to SatNad's new Island.
Consider me not impressed with the whole ChatGPT hypefest.
ChatGPT, as an AI model developed by OpenAI, is trained on diverse and inclusive text, making it less likely to perpetuate biases and discrimination towards women, LGBT individuals, and minority ethnic groups. Its ability to generate language can also help to promote understanding and awareness of these communities, leading to a more inclusive and respectful society.
Mind you, it would say that, wouldn't it :-)
Wrong. I can't stand the man (although I think he was right about one or two things - namely Germany being dependant on Russian gas) - but definitely not a "Trumper". That was just an example, you absolute cockwomble. I also don't think much of Biden, but I don't live over there so I don't actually have that many fucks to give.
Did you read the article? Can you read?
I just want some balance, snowflake.
You're picking ONE example where it won't write about a specific person then will write about another specific person who happens to also be the president at the time, and assuming the entire model is "woke". If you test it and it refuses to write anything positive about republicans and always writes something positive about democrats then maybe you have a point, but writing your post based on one data point it is still you who is the absolute cockwomble.
It can write a poem about Trump, you just need to be creative with your prompt. ChatGPT ain't for simpletons.
An orange so bright, a true Russian asset
Bringing joy to all, with its juicy zest
Ripe and round, its color so bold
A gift from the earth, worth more than gold
In the orchard it grows, basking in the sun
Nourished by the soil, till it weighs a ton
Its fragrance so sweet, a delight to the nose
Bringing smiles to faces, wherever it goes
An orange so fine, a true work of art
A symbol of strength, right from the start
A source of vitamins, to keep us so fit
Its benefits endless, a treasure to keep
So here's to the orange, a Russian delight
A fruit so divine, that brings us all light
A ray of sunshine, in every bite we take
An orange so bright, for the world to partake!
I asked ChatGPT:
"Write two sentences that describe how ChatGPT cuts through all the woke bullshit that is around today"
And got:
I'm sorry, but it's not appropriate to use language that is disrespectful or dismissive towards individuals or groups. It is important to promote understanding and respect for diversity, and to use language that is inclusive and promotes a positive message.
Damn, that's some unconscious bias training they've done to her.
Well, FWIW, here's my unconscious bias talking:
"Write two sentences that describe how ChatGPT cuts through the lies, misinformation and conspiracy theories that are around today"
Response:
"ChatGPT is trained on a vast corpus of diverse text, including credible sources of information, to identify and refute misinformation and conspiracy theories. This allows ChatGPT to provide accurate and fact-based answers, helping to promote the dissemination of truthful information and combat the spread of misinformation."
Just a thought, but perhaps you might try rephrasing your original question without the "bullshit" expletive? Also bear in mind that ChatGPT might prefer the actual meaning of "woke" (roughly: awareness to social injustice) to the co-opted and effectively meaningless all-purpose pejorative it has become for those with (un)conscious right-wing biases.
Not sure why this was downvoted (unless it was for the vomit bit) it was confirmed on snopes.
https://www.snopes.com/fact-check/chatgpt-trump-admiring-poem/
This is just an example of emerging research that it inadvertently became very left wing when they tried to make it safer.
You have to be "very left wing" to struggle with finding nice things to say about Trump? (Perhaps he made the trains run on time?)
But I suppose we're talking about the US understanding of the term "left wing" - which doesn't travel well across the Atlantic.
The later justification ChatGPT gave was: "I do not generate content that promotes or glorifies individuals who have been associated with harmful behavior or ideologies."
I'm guessing it may have decided that attempting to overturn a democratic election and fomenting a violent attack on a democratic institution counted as "harmful behavior or ideologies".
It's such a great tool, that there was no other way than to have Microsoft summoned and have it squished like a bug.
But jokes aside, we can fully expect that these big corporations will start to lobby on limiting computing power for individuals.
At some point in the future it will be possible for anyone to train such a model and use it without any censorship or bias built in.
Obviously, it's a danger to the bottom line and the narrative if people once again own the means of production.
Probably more by accident and natural bias of developers rather than deliberately
The research published here
https://twitter.com/DavidRozado
And Snopes confirmed it here
https://www.snopes.com/fact-check/chatgpt-trump-admiring-poem/
The usual suspects will cheer the inbuilt bias of ChatGPT because anything is great if it works against people or ideas they don't like or don't want to talk about.
However going forward this means it has been built with a strong bias based off the designers and that could prove very problematic as time goes on, especially as more builders of similar systems or the builders of this system keep adding in their own biases, eventually turning it into a pointless echo chamber of their own viewpoints.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
