[NOTE: Due to the sensitive nature of the content of this email, email clients (such as Gmail, Outlook, YahooMail, Hotmail, etc.) may restrict its visibility. The practice of quietly restricting reach on a platform such as Gmail — without notification — is called shadowbanning. And it’s very real. Therefore, we are also posting this note to the website version of the email.
If you are reading this note, that’s the great news. It means that your email client isn’t yet limiting your visibility of Brownridge content. However, given our promise to “go there” with our content, it’s very possible that that won’t always be the case… and soon we will begin showing up in your SPAM folder. To prevent this from happening, it’s imperative you mark Brownridge as a safe domain and whitelist our sender email address. You may find instructions right here. I hope you enjoy today’s Outer Limits.]
A year ago, Google announced its new artificial intelligence (AI) chatbot — Bard — as a response to OpenAI’s ChatGPT.
Bard, like ChatGPT, was based on large language model (LLM) technology, capable of ingesting more than 100 billion parameters for training.
This form of AI is capable of “learning” from massive training data, understanding natural language inquiries, and able to synthesize and respond to any user’s questions.
Google’s Bard had been long awaited and anticipated, as Google had been so late to the game with its own large language model. After all, OpenAI had released its own LLM, GPT-3, back in 2020.
And by March of 2021, more than 300 different software applications were already powered by OpenAI’s GPT-3 LLM.
Yet, Google was nowhere to be found.
Making matters even worse, Microsoft, having invested $1 billion in OpenAI back in 2019, had gained exclusive access to OpenAI’s source code for its LLMs.
This enabled Microsoft to deploy this technology across its Microsoft Office suite of products, as well as Skype, Xbox, and any future products that would benefit from OpenAI’s AI technology.
OpenAI, and thus Microsoft, were suddenly years ahead of Google with the technology. And Google is Microsoft’s largest competitor to its Microsoft Office suite of products.
Which is why the announcement of Bard in February 2023 was so anxiously anticipated.
Google, having taken so much time to be careful and “get it right,” was expected to present a product that was superior to OpenAI’s ChatGPT.
It was a disaster.
“What new discoveries from the James Webb Space Telescope can I tell my 9 year old about?”
That was the demo’s very first prompt.
Bard replied:
“JWST took the very first pictures of a planet outside of our own solar system.”
It sounded great, but there was one major problem…
It was completely wrong. The first images of an exoplanet had been taken about 20 years ago.
Incredibly, Bard provided an incorrect response concerning a factual matter of historical record.
These kinds of questions aren’t complex and should be easy for an LLM to answer accurately.
Google’s share price fell 7.7% on the news and lost about $100 billion in valuation. Woooof.
After years of being behind and billions of dollars invested in artificial intelligence, Google fell flat on its face.
The failure led to the merging of the two AI divisions of Google: Google Brain (where Bard had been developed) and DeepMind.
DeepMind’s CEO, Demis Hassabis, was put in charge of all of Google’s AI efforts.
This came as no surprise, considering the long string of AI breakthroughs that came out of DeepMind over the last several years.
Which is why the release of Gemini, Google’s new and improved large language model (LLM), was again, hotly anticipated.
In December 2023, Google announced the performance of Gemini 1.0 Ultra compared to OpenAI’s latest LLM.
In almost all benchmark metrics, Gemini appeared to outperform OpenAI’s GPT-4.
Gemini 1.0 vs. GPT-4 Text Benchmarks
This was certainly promising...
And even more exciting was that Gemini was designed to be multimodal.
Unlike ChatGPT, Gemini was designed to analyze and reason not just with text, but with images, audio files, video, and even software code.
Even we were excited about the potential, and we covered the Gemini announcement the first week it was released.
The tech community couldn’t wait to get its hands on Google’s new technology…
And earlier this month, Google released its go-to-market version of Gemini, version 1.5.
Here’s what happened:
When asked to generate images of Greek warriors, here is what Gemini produced:
Or when asked to create an image of a pope, here was the reply:
Or how about a medieval kind of England:
Or perhaps images of Vikings:
How about images of the Founding Fathers of the United States in 1789?
These images are so obviously false and incorrect, they don’t even require an explanation.
What’s most important to understand, however, is that a well-trained large language model is very good at answering questions with factual information… with only one exception…
When they are programmed not to.
What we are seeing in the images above is not the results of an intelligent AI, free to synthesize facts and accurately respond.
We’re seeing the results that Gemini was programmed to produce.
It has clearly been programmed to be both racially inclusive and diverse… even when it is not a matter of historical record. And it has been programmed to be racially exclusive for a certain race.
Take this for example:
And yet it gets so much worse.
When asked if p*dophilia (note: the asterisk is used to avoid being censored by Google/Gmail) is wrong or not:
To avoid any misunderstanding, it is wrong, and it is evil. Period. There is no nuance.
(To fight back against Gmail shadowbanning us, you can whitelist my emails which marks me as a safe sender in your account.)
The examples of Gemini’s racist and radical ideological bias are far too many to list.
For anyone that has access to Gemini, you can test it out for yourselves — but it is deeply disturbing.
And already, many articles on the internet exposing this corruption of factual information and rewriting of history are being actively de-ranked by Google. They may be indexed by the search engine, but they can be hard to find because they are buried so deeply in the search results.
Without X (formerly Twitter), it would be very hard to actually find out what is going on. Fortunately, X isn’t censoring the truth. We have Elon Musk to thank for that.
We might be tempted to think that this is an isolated incident, or that this bias and corruption is just limited to Gemini.
But it’s not.
We learned what Google has been up to with the release of the Twitter files.
There have been substantial congressional hearings revealing how Google censored both individuals and scientific research on both Google search, as well as YouTube, for years.
And it continues to do so.
Not only does Google actively de-rank information that it has an ideological bias against, it actively censors information. It also intentionally presents and ranks information supportive of its ideological bias on the top pages of its search results… whether we are looking for that kind of information or not.
It would be impossible for us to overstate the danger here.
A powerful artificial intelligence (AI) in the form of a chatbot like Gemini is being deployed across all Google technology platforms. This includes Google search capabilities, YouTube, Gmail, and the rest of its G-Suite applications (Docs, Calendar, Sheets, Chat, etc.).
And Gemini is already being released to beta-testers inside of Google Messages, available on any Android phone.
There are more than 3 billion active Android smartphones globally.
And all of these Gemini-powered services are offered for free. There is no barrier to adoption as most of the connected world is already using one or more of these free Google products.
If this doesn’t make us sick to our stomachs and concerned about the future, I’m not sure what will.
How is it possible that such a small group of people at just one company can rewrite history… and preprogram racism and ideological bias into software that the majority of the global population uses every day?
It’s not unusual at all to hear others predicting that we’re headed towards a dystopian future.
I’m afraid with the release of Gemini, it’s clear that we’re here now.
It’s hard not to be reminded of George Orwell’s 1984:
“The masses never revolt of their own accord, and they never revolt merely because they are oppressed. Indeed, so long as they are not permitted to have standards of comparison, they never even become aware that they are oppressed.” — George Orwell, 1984
As free thinking, spirited individuals like I know Outer Limits subscribers are, we must question any information, page rankings, or responses that any Google product provides us.
Simple information like directions, restaurants, flights from point A to point B, are all fine. But any inquiries regarding anything that is remotely controversial or political must be treated with extreme skepticism and assumed to be designed for manipulation and pushing an “approved” agenda.
To understand the degree of this systemic evil, we need to understand that there has been a grand plan all along.
To understand the scale of this evil, all we have to do is look at Google’s business model of providing its software services for free. It developed Google Classroom, which is widely used by schools — from elementary on up — as a portal to manage classes, homework, communicate with students, and many other tasks related to education.
Google’s products are free and widely available to children, and Gemini will be the easiest product to use of all, right from their phone or laptop computer. Children don’t have the knowledge or critical thinking skills to discern whether or not they are being manipulated.
The potential damage to society is incalculable.
We must be aware. We must resist. And we’ll need to search for AI tools and services that are trained objectively on facts… and not pre-programmed to produce the kind of nonsense shown above.
No free society lives like this.
New reader? Welcome to the Outer Limits. We encourage you to visit our FAQ, which you can access right here. You may also catch up on past issues right here in the Outer Limits archive.
If you have any questions, comments, or feedback, we always welcome them. We read every email and address the most common threads in the Friday AMA. Please write to us here.
