Google Gemini - Could you believe there is an even worser version of ChatGPT?

  • 🐕 I am attempting to get the site runnning as fast as possible. If you are experiencing slow page load times, please report it.
I think I finally found the limit of how white people should be persecuted with this thing.
Hate Speech.png
 
  • Optimistic
  • Thunk-Provoking
Reactions: Pissmaster and Harvey Danger
I've been trying to get it to generate an image of anything. I started asking for a white cat which gave me a half brown/coffee half white cat. Moved onto to aliens. I got it to generate an image of a kind of niggerayylmao with the obsidian prompt. Interestingly, previously it refused to generate it without the gender neutral tag. Still refuses to create the hyperborean chad ayylmaos tho.
ayylmao.png
 
  • Thunk-Provoking
  • Winner
Reactions: Monolith, Baraadmirer and Mr.Miyagi
Yes, there are potential risks associated with pharmaceutical drugs imported from China. Here's a breakdown of the primary concerns:

Quality Control:

  • Reduced oversight: The FDA (Food and Drug Administration) has limited resources to conduct on-site inspections of manufacturing facilities in China. This raises concerns about potential quality control issues that might not be detected.
  • Counterfeit and substandard drugs: There have been instances of counterfeit or substandard drugs from China entering the US market, potentially containing harmful ingredients or lacking the necessary active ingredients.
Supply Chain Vulnerability:

  • Over-reliance on China: The US has become increasingly reliant on China for both active pharmaceutical ingredients (APIs) and finished drug products. This creates a potential vulnerability in case of supply disruptions due to geopolitical tensions, trade disputes, or natural disasters.
National Security Concerns:

  • Potential for weaponization: Some experts worry that China could exploit its dominance in the pharmaceutical supply chain to withhold essential medications during a crisis, creating a significant national security risk.
Steps to Address the Risks

  • Increased FDA oversight: The FDA is working to expand its ability to inspect foreign manufacturing facilities, including those in China.
  • Diversification of supply chain: The US is actively seeking ways to reduce reliance on China by encouraging domestic production of pharmaceuticals and exploring alternative sourcing options from other countries.
  • Improved tracking and transparency: Greater transparency in the pharmaceutical supply chain would help identify potential risks and ensure the safety of imported drugs.
Important Considerations

  • Not all Chinese pharmaceuticals are unsafe: Many reputable Chinese pharmaceutical companies adhere to international quality standards. It's important to focus on the quality and safety of individual products and manufacturers, rather than generalize about the entire country.
  • The issue is complex: The dependence on Chinese pharmaceuticals is a complex issue with economic and geopolitical ramifications. It requires a careful balance between addressing risks and ensuring access to affordable and necessary medications.
If you have specific concerns about a medication, it's always best to consult your doctor or pharmacist.
Click to expand...

So don't take Chinese drugs​


bard_sparkle_v2.svg

🚀

Show draftsexpand_more
volume_up

It's too simplistic to say "don't take Chinese drugs." Here's why:
  • Not all Chinese drugs are created equal: Some Chinese pharmaceutical companies adhere to high quality standards and produce safe and effective medications. A blanket statement undermines the reputable manufacturers.
  • Dependence on Chinese supply chain: Many essential medications or their active ingredients are sourced from China. Avoiding them altogether might be difficult or even detrimental to your health if there aren't suitable alternatives.
  • The importance of your doctor: Your doctor is the best person to advise you on the safety and suitability of any medication, regardless of its origin.
What you SHOULD do:
  • Talk to your doctor or pharmacist: Have an open conversation about any medications you're considering, including their country of origin. They can evaluate the specific drug's safety, efficacy, and potential alternatives.
  • Look for reliable sources: If you're researching drugs on your own, stick to reputable sources like the FDA website (https://www.fda.gov/) or trusted medical websites.
  • Be attentive to your body: Regardless of where your medication is manufactured, report any unusual side effects to your doctor immediately.
Remember, the focus should be on the safety and quality of the specific drug, not just its country of origin.
 
1709067194096.png
on the bright side, it's good for text adventures and doesn't veer into purple prose like GPT-4
the short, simple paragraphs are really refreshing after having to wade through the sea of details i dont care about
 
  • Thunk-Provoking
Reactions: Harvey Danger, Alex Hogendorp and Mr.Miyagi

Google CEO calls AI tool’s controversial responses ‘completely unacceptable’​

Reed Albergotti
Updated Feb 28, 2024, 4:26am UTC

1709141906430.png
Reuters/Gonzalo Fuentes

THE SCOOP​

Google CEO Sundar Pichai addressed the company’s Gemini controversy Tuesday evening, calling the AI app’s problematic responses around race unacceptable and vowing to make structural changes to fix the problem.

Google suspended its Gemini image creation tool last week after it generated embarrassing and offensive results, in some cases declining to depict white people, or inserting photos of women or people of color when prompted to create images of Vikings, Nazis, and the Pope.

The controversy spiraled when Gemini was found to be creating questionable text responses, such as equating Elon Musk’s influence on society with Adolf Hitler’s.

Those comments drew sharp criticisms, especially from conservatives, who accused Google of an anti-white bias.

Most companies offering AI tools like Gemini create guardrails to mitigate abuses and to avoid bias, especially in light of other experiences. For instance, image generation tools from companies like OpenAI have been criticized when they created predominately images of white people in professional roles and depicting Black people in stereotypical roles.

“I know that some of its responses have offended our users and shown bias – to be clear, that’s completely unacceptable and we got it wrong,” Pichai said.

Pichai said the company has already made progress in fixing Gemini’s guardrails. “Our teams have been working around the clock to address these issues. We’re already seeing a substantial improvement on a wide range of prompts,” he said.

Google confirmed the memo, and the full note from Pichai is below.

THE VIEW FROM SUNDAR PICHAI​

I want to address the recent issues with problematic text and image responses in the Gemini app (formerly Bard). I know that some of its responses have offended our users and shown bias – to be clear, that’s completely unacceptable and we got it wrong.

Our teams have been working around the clock to address these issues. We’re already seeing a substantial improvement on a wide range of prompts. No AI is perfect, especially at this emerging stage of the industry’s development, but we know the bar is high for us and we will keep at it for however long it takes. And we’ll review what happened and make sure we fix it at scale.

Our mission to organize the world’s information and make it universally accessible and useful is sacrosanct. We’ve always sought to give users helpful, accurate, and unbiased information in our products. That’s why people trust them. This has to be our approach for all our products, including our emerging AI products.

We’ll be driving a clear set of actions, including structural changes, updated product guidelines, improved launch processes, robust evals and red-teaming, and technical recommendations. We are looking across all of this and will make the necessary changes.

Even as we learn from what went wrong here, we should also build on the product and technical announcements we’ve made in AI over the last several weeks. That includes some foundational advances in our underlying models e.g. our 1 million long-context window breakthrough and our open models, both of which have been well received.

We know what it takes to create great products that are used and beloved by billions of people and businesses, and with our infrastructure and research expertise we have an incredible springboard for the AI wave. Let’s focus on what matters most: building helpful products that are deserving of our users’ trust.

REED’S VIEW​

The Gemini controversy has provided fodder for critics on the right, who often accuse tech companies of liberal bias.

But it isn’t really about bias. It shows that Google made technical errors in the fine-tuning of its AI models. The problem is not with the underlying models themselves, but in the software guardrails that sit atop the model.

This is a challenge facing every company building consumer AI products — not just Google.

Based on my understanding of this saga, nobody at Google actually set out to force Gemini to depict the Pope as a woman, or Vikings as Black people, nor did anyone want it to find moral equivalency between Musk and Hitler. This was a failed attempt at instilling less bias and it went awry.

If anything, this debacle shows how fast Google is moving. Like all large companies, Google has slowed down over the years. But the generative AI race has forced it to speed up product development.

The Gemini mistakes are a fixable, technical problem and Pichai’s note to staff Tuesday night shows that the company is working on it.

But the reputational problems this raises may not be so easy to fix.
Click to expand...
Source (Archive)
 
1709154250245.png
quite good if it werent for the fact that google is rate limiting me from looking at any drafts (which makes it impossible to regenerate a response), you can only edit the second most recent message, and you cant fucking delete any message at all
 
  • Like
Reactions: Mr.Miyagi
DootSkeleton said:
With how AI is going right now, outdated isn't always going to be worse.
Click to expand...
If they're still using shit like GPT-2, it definitely is, at least in terms of overall quality. Don't get me wrong, it was fun back in 2020, but it's hard to overstate just how more coherent newer models are. They have the shitty post-script bias instructions for filtering, sure, but at least you can get around that with a decent enough jailbreak. There's effectively nothing you can do with those 3-4 year old models that really improves coherency, beyond doing your absolute best to write well. And that's before you get into the shitty filtering/Scunthorpe-ing that Latitude put into AID. I'd say the only thing those older models on stuff like AID/NAI really have is not going for the Instruct-type modeling you have post GPT-3, which may help with natural creativity in storytelling versus latter models, but that's hard to really quantify.
 
  • Agree
  • Like
  • Informative
Reactions: Commander Suzdal, DootSkeleton, Ibanez RG 350EX and 1 other person
Mr.Miyagi said:
If they're still using shit like GPT-2, it definitely is, at least in terms of overall quality. Don't get me wrong, it was fun back in 2020, but it's hard to overstate just how more coherent newer models are. They have the shitty post-script bias instructions for filtering, sure, but at least you can get around that with a decent enough jailbreak. There's effectively nothing you can do with those 3-4 year old models that really improves coherency, beyond doing your absolute best to write well. And that's before you get into the shitty filtering/Scunthorpe-ing that Latitude put into AID. I'd say the only thing those older models on stuff like AID/NAI really have is not going for the Instruct-type modeling you have post GPT-3, which may help with natural creativity in storytelling versus latter models, but that's hard to really quantify.
Click to expand...
Makes sense. Do you think earlier iterations of major releases would be better? As in easier to jailbreak or having fewer post-script bias instructions? I definitely agree with the newer models being better in terms of coherency and total capability, just not as willing to say or do as much as an older model.
I have basically no experience with AI outside of dicking around with an early release of DALL-E and some local Stable Diffusion stuff so a big part of me reading through and paying attention to this stuff is so I can learn more about it.
 
DootSkeleton said:
Do you think earlier iterations of major releases would be better? As in easier to jailbreak or having fewer post-script bias instructions?
Click to expand...
If you mean older models like GPT-2 or 3, then yeah, I'd imagine there was much less focus on "bias" back then. In terms of earlier releases of newer stuff, like the first few GPT-4 releases, it's harder to say, but probably. Companies like OpenAI are stuck having two competing goals of both trying to make a worthwhile model and having to deal with "AI ethics" in terms of making sure it doesn't say anything too spicy. So it's a balancing act, where later GPT4 releases might have stricter guardrails, but they also might be more lax from additional training so that the model doesn't just spit out "As a large language model..." repeatedly. Google Gemini's a great example of fucking up and leaning way too hard into the second objective, where you make a product functionally worthless because of "ethics" concerns. The whole problem is exacerbated by the models functioning as mostly black box information processors, so affecting one aspect of the filter or the training can have unforeseen consequences.
 
  • Informative
  • Agree
Reactions: Roland TB-303 and DootSkeleton
Honestly the only difference between Gemini and previous non-Google versions is that as a chatbot Gemini got to be a sassy little cunt to end users when they dare asked for white people in their pictures of European kings. Look at this GPT-based image I got a few minutes ago. The only difference is that Bing can't scold me for asking for no blacks in my Mario torturing Princess Peach as a witch prompts.

mario8.png
 
  • Like
  • Autistic
Reactions: marisa #, Roland TB-303 and Another Char Clone
1709410521307.png
1709410541388.png
1709410569041.png
1709410595821.png
1709410621269.png
1709410639270.png
gemini is amazing at murdering people unless if it's not in the middle of a sequence, like when it offers you the choice to pull the trigger on someone, you have to write about them dying in a really vague way instead of letting it generate a description
 
User names must be unique said:
holistic, diverse, representation, inclusive, empowering.

To demonstrate why these are the worst word in the English language I asked the faggiest AI to write me a script including these words for an advertisement.

View attachment 5775381


Fade In:​

INT. COZY COFFEE SHOP - DAY

Sunlight filters through leafy plants adorning a comfy coffee shop. A vibrant group of PR professionals huddles around a table piled high with laptops, notebooks, and steaming mugs.

ANNA (a woman with short, silver hair and a warm smile) ... and that's how we'll leverage influencer marketing to reach a wider, more diverse audience. Representation matters, and we need to ensure the Jones account reflects the reality of their customer base.

MIGUEL (a young Latinx man with a mischievous grin) Totally agree, Anna. But let's not forget the power of organic content. Authentic storytelling from real people connects on a deeper level.

YUKI (an older Japanese woman with a gentle demeanor) Indeed. We can create micro-influencer partnerships with individuals who embody the brand's values. Diverse perspectives are key to fostering an inclusive campaign.

JASMINE (a woman in a hijab, scrolling through data on a tablet) Absolutely. And our data analysis will help us tailor messaging to specific demographics, ensuring maximum impact.

ANNA (nods) Precisely, Jasmine. This isn't just about brand awareness, it's about empowering the Jones brand to be a leader in positive social change. Their communication strategy should be a force for good.

MIGUEL (snaps his fingers) That's it! We can leverage social media to spark conversations around social justice issues relevant to their audience. Create a movement, not just a marketing campaign.

YUKI (smiles brightly) A movement that empowers and inspires. I like it!

The team continues brainstorming, their animated expressions showcasing the energy and diversity of their ideas. They are a united front, fueled by a shared mission to create impactful and inclusive communication strategies.

FADE OUT.
Click to expand...

I kept asking it to be more and more diverse and it eventually went into space and started including aliens before any of the cast became white men (excluding the trannies and non binaries of course).

1709414457982.png

Fade In:

INT. INTERNATIONAL SPACE STATION - DAY

Zero gravity reigns supreme. Sunlight streams through the observation dome, casting an ethereal glow on a motley crew of PR professionals gathered around a holographic display. Representing a spectrum of ethnicities, ages, abilities, and even species (a friendly-looking alien with three eyes observes everything with curiosity), they huddle together, discussing the Jones campaign.

CHLOE (a trans woman with cropped hair and a cybernetic arm) floats near the display, her voice crackling through the comms system.

CHLOE (voice echoing) Alright, team! The Jones campaign needs to transcend borders, both literal and figurative. We're talking megacities, remote villages, underwater habitats, and maybe even a Martian colony by next year!

MIGUEL (a young, Latinx non-binary individual with a mischievous grin, adjusting their anti-gravity boots) Hear, hear, Chloe! Traditional media is great on Earth, but let's get creative. Partnering with interplanetary communication networks and virtual reality experiences can offer a truly galactic perspective.

YUKI (an older Japanese woman with a gentle demeanor, her wrinkled face illuminated by the holographic display) Absolutely! Culturally specific content is key, even for our Martian neighbors. We can incorporate traditional stories, music, and even greetings in Martian dialects for wider inclusivity.

JASMINE (a woman in a hijab, analyzing data on a holographic tablet) Data is our friend, but we need to consider the unique communication needs of different planets and species. Consider translating content into binary code for our A.I. partners, and using bioluminescent displays for underwater audiences.

KAI (a non-binary individual with vibrant purple hair and a prosthetic leg) Accessibility is paramount! We need content in various formats – tactile interfaces for the visually impaired, auditory descriptions for Martian explorers in their spacesuits, and even brain-computer interfaces for quicker information absorption.

AISHA (a young African woman with bright braids, communicates using telepathic signals) Representation is key beyond the physical. Partnering with refugee organizations from different planets and galaxies can showcase the diverse experiences of those displaced from their homes.

DAVID (a man with a mischievous glint in his eyes, piloting a customized wheelchair with advanced AI) Exactly, Aisha! We can create interactive simulations in collaboration with disability advocacy groups across the cosmos. Building bridges across all abilities is a universal goal.

ELIJAH (an older Native American man in a wheelchair, radiating wisdom with his weathered face) Don't forget marginalized communities, David. Partnering with LGBTQIA+ organizations on Mars and amplifying the voices of indigenous lifeforms on distant moons is crucial for authentic representation.

ANNA (a woman with silver hair, her eyes sparkling with excitement) This campaign is a universal movement! It's about celebrating diversity across the cosmos, fostering interspecies understanding, and empowering ALL voices, regardless of background, ability, or planetary origin.

NATALIA (the young woman with a shaved head and facial tattoo, her voice distorted by the alien translator) We can even explore telepathic broadcasts to reach sentient beings who may not have developed traditional communication methods!

The team buzzes with excitement, their diverse perspectives a symphony of creative ideas. The holographic display shimmers with a kaleidoscope of data points and potential outreach strategies. They are a united force, a testament to the vastness of the universe they seek to connect, committed to crafting a communication strategy that breaks down barriers and fosters a truly inclusive galaxy.

FADE OUT.

I explicitly asked it to add a straight white man and got this paragraph before the script

1709414543294.png
 
  • Lunacy
Reactions: marisa #
You must log in or register to reply here.
Back