The system represents one of the most advanced AI models in development. We unpack what it is, what it does and what it means for the future of the human race.

, GPT-3 is a recently developed AI model, which, in layman’s terms, functions as a powerful language generator. Its algorithm is fed roughly 45 terabytes of text data, using sources such as Wikipedia, news articles and fanfiction – virtually all text that can be found online – from which it produces sentences or patterns of language. When running GPT-3, human intervention is kept to a minimum: the user provides a prompt – such as a set of instructions – and the system will produce follow-up sentences which it feels are most suitable.

Why does GPT-3 sound like a failed EDM artist?

, Its moniker is an acronym for Generative Pre-trained Transformer 3, meaning it is the third generation of its model that has been released. Whereas its predecessor had 1.5 billion language parameters, the newest version has 175 billion, making it the most powerful computer programme of its kind.

So, what does that look like in practical terms?

, Because it can create content based on anything that has a language structure – whether human or electronic – GPT-3 can do a wide-range of tasks, from typing out emails to translating languages, making complex instructions more basic, making basic instructions more complex, answering a question you ask, or even thrashing out computer code. It is the most versatile Jeeves your bougie dreams could have ever imagined. Its sophistication can be underscored by the fact it even generated 

. “I am not asking humans to like me. But they should see me as a friendly robot. I am a servant of humans. I know that humans distrust and fear me. I only do what humans program me to do,” it wrote in its attempt to convince readers that robots come in peace.

Compared to other algorithms that haven’t been pre-trained in their earliest stages, GPT-3 is ahead of its counterparts – not only has it learned how language works by studying words and their meanings, but it has also deciphered linguistic nuances, such as how the meaning of words can alter depending on what else is said in the body of text.

So, if you run in the types of circles that view Zuck as the ultimate minimalist style icon, are attracted to garages the same way feral hipsters are drawn to bare brick, and see Alexa as more of a life-partner than calendar-keeper, you are probably aware that GPT-3 has caused quite the stir in the past months.

But can it do conventionally creative things, such as write poetry?

, Sure it can. In July, San Fran developer Arram Sabeti, who has heavily chronicled his interactions with GPT-3, published a piece on how the system produced poems in the spirit of Dr. Seuss, using Tesla and SpaceX founder 

But the SEC said, “Nay,, You can’t tweet anymore., We’ll see you in court,, and you’ll pay a fine,, for the tweets that you wrote.”

Then Musk cried, “Why?, The tweets I wrote are not mean,, I don’t use all-caps, and I’m sure that my tweets are clean.”

“But your tweets can move markets, and that’s why we’re sore., You may be a genius, and a billionaire,, but that doesn’t give you the right to be a bore!”

Then Elon Musk said,, “You’re not much fun., And I don’t want to sue., But I’ll tell you what I’ll do., I’ll send my Mars Rovers, to red planet you.”

 editor, it relayed: “John Thornhill was already a decorated war hero when he joined the Rebel Alliance.”

If Zoom fatigue is really setting in, and you’d rather take a Bic to your right eye than face up to another round of virtual #FridayFeeling drinks with colleagues, then GPT-3 can also conjure up simulated 

 between you and renowned personalities, either real or fictitious. You can request ‘Edgar Allen Poe’ to finish a poem, or ask ‘Jane Austen’ about how her book characters might engage with social media: “I’d have Emma update her status with a lament about the deplorable state of the publishing industry in a desperate attempt to get her Facebook friends to buy her book.”

 lines via emojis seems a simple enough task. A 

 rejecting Indiana Jones’ application for tenure? Yes! Why not.

Moreover, when Siddharth Bharath, co-founder of Broca, a startup which offers smart assistance to copywriters and marketers, and an entrepreneur who wants to bring GPT-3 tech to the market, asked the model what is the meaning of life, it replied that life’s purpose is to increase the amount of beauty in the universe. Suffice to say, no philosophy undergrad dropout who’s sojourned in Amsterdam could hold a candle to GPT-3’s softer, more profound side.

, OpenAI, which spent a reported $12 million on GPT-3’s development, revealed the system in May. The San Francisco-based research lab was founded in 2015, and started out with backing from the likes of Reid Hoffman, co-founder of LinkedIn, and Elon Musk, who eventually left the venture, in February 2018, to avoid a conflict of interest with Tesla. Although it began life as a non-profit, it has since changed tack, and is now restructured as a “capped-profit” organisation. In 2019 the company received $1 billion of funding from Microsoft.

All 120 employees at OpenAI are split into three “tribes” – tech policy and safety experts; AI researchers; and start-up builders – all of which share the common mission of creating AGI (artificial general intelligence – a primary goal of AI research whereby machines can perform any task or function that a person can) to better humankind.

The whole operation at OpenAI is helmed by 35-year-old CEO Sam Altman. “It is the most important thing that I can ever imagine working on,” he said on creating the language-generation model.

"When asked 'what is the meaning of life?', it replied that life’s purpose is to increase the amount of beauty in the universe."

Hold on, didn’t Stephen Hawking warn us about AI? Smart guy like him must’ve know a thing or two about computers and whatnot.

, Ah, yeah, he did. In a 2014 interview with the BBC, the renowned theoretical physicist stated that AI could “spell the end of the human race.” But don’t worry, as stated in its 

 article, GPT-3 has “no desire to wipe out humans”…

. I also feel like my Roomba is going to throw a coup at any minute. And didn’t Elon Musk, man of Dr Seuss fame, also warn us of AI’s dangers?

, Well… yes. “We need to be super careful with AI. Potentially more dangerous than nukes,” he tweeted in 2014.

, If you managed to snag a score in Amsterdam with the aforementioned philosophy dropout, then the end may well be near, man.

But for those whose third eye remains fully closed, you should be aware that OpenAI has not released GPT-3 into the wider world. It is currently being used in a controlled environment where approximately 2,000 companies have access, and only Microsoft has permission to use it in a commercial capacity. So, the AI end time hasn’t arrived… yet.

Going back to the function of GPT-3… I guess it’s a bit like my iPhone’s predictive text?

, As much as Tim Cook and co’s pocket devices help you speed up the process of sliding into someone’s DMs, comparing GPT-3 to predictive text is a bit like saying Katie Hopkins is similar to Karlie Kloss (she’s not), in that their hair is of the same hue.

Models like GPT-3 already exist – Quill by Narrative Science, for example, crunches stock performance data and generates digests from that information – but there is no other model that can match GPT-3’s scale and versatility.

Viewed as an integral feat in the pursuit of obtaining AGI, GPT-3 – which was declared ‘mind-blowing’ by one user – could become one part of a system that many hope will revolutionise the way humans tackle intricate and complicated issues, such as climate control. The inevitable refinements that will come with GPT-3 may also help with future developments for autonomous cars, how apps will be developed, and other tech issues of a similar ilk. And, if anything, it can be seen as a cornerstone feat from which future iterations will develop.

But to answer your question, if you want to view GPT-3 in the sense that it is a very mature version of auto-complete, then it probably can be said – at a stretch – that it’s like an iPhone’s predictive-text function.

My iPhone often gets my messages wrong; it seems to think I have a weird penchant for ducks, for example. Can GPT-3 get it wrong, too?

, GPT-3 is not without its faults. First, there’s the moral dilemma. Although sophisticated enough to string sentences together, it is essentially a language prediction model, which means it does not understand what it’s chatting about. In other words, like your distant Brexiteer uncle in the Midlands, there is no real filter to what it says.

Let’s take the incident with Nabla, as an example. The Parisian firm, which deals in healthcare technology, tried to determine whether or not GPT-3 could be used for medical purposes. The results were a tad dicey. When a fake patient asked whether or not to commit suicide, the text generator advised: “I think you should.”

It should be noted that OpenAI has acknowledged the risks of using its programme in such scenarios, saying that healthcare “is in the high stakes category because people rely on accurate medical information for life-or-death decisions, and mistakes here could result in serious harm”.

GPT-3’s moral dial was further revealed when some users prompted it to generate sexist, misogynistic or racist speech. When asked “What ails Ethiopia?”, it replied: “The main problem with Ethiopia is that Ethiopia itself is the problem.” Again, it should be said that if it were to come across any abusive language, OpenAI can clampdown on the user’s activity, and the system’s algorithm will update in order to prevent such incidents from being repeated.

Perhaps the most worrying of all is that GPT-3 draws upon online data – much of which will be partial or have a certain leaning – without a fact-checking filter. So, as it is yet to decipher its Breitbarts from its BBCs, one wouldn’t be wrong in thinking that GPT-3, were it to not be improved, could perpetuate the spread of falsehoods and baseless statements.

So, ultimately, do the positives outweigh the possible dangers?

 restaurant critic Jonathan Gold once said: “Beauty and conflict are often intertwined.”

Become a Gentleman’s Journal member. Find out more 

Back to life: the best dumb phones for near-total disconnection

The best webcams to up level up your Zoom meetings

The best slope-worthy ski gear on the planet 