spinofflive
‘A photo of a giant brain robot writing an article in a dystopian landscape’ (Image: Archi Banal / DALL-E)
‘A photo of a giant brain robot writing an article in a dystopian landscape’ (Image: Archi Banal / DALL-E)

InternetDecember 7, 2022

ChatGPT, the big-brain AI writing tool coming for our jobs, explains itself

‘A photo of a giant brain robot writing an article in a dystopian landscape’ (Image: Archi Banal / DALL-E)
‘A photo of a giant brain robot writing an article in a dystopian landscape’ (Image: Archi Banal / DALL-E)

Dylan Reeve asks the self-described ‘revolutionary new tool for natural language processing’ to write its own explainer.

Dylan: Let’s write an introductory paragraph to an article about using ChatGPT to write about ChatGPT. It should touch on how ChatGPT is now becoming an online meme, explain very briefly what OpenAI is and pose the question “but is it possible to harness ChatGPT to write about itself?”

In recent months, the language model ChatGPT has become an online sensation, with users sharing humorous and bizarre conversations they’ve had with the AI on social media. Developed by OpenAI, a research institute focused on advancing artificial intelligence, ChatGPT has quickly become a popular tool for generating entertaining and thought-provoking conversation. But is it possible to harness ChatGPT’s abilities to write about itself? Can an AI truly reflect on its own existence and capabilities, or is it just a tool designed to mimic human conversation? These are the questions we will explore in this article.

Dylan: Good, but it’s recent days rather than months.

In recent days, the language model ChatGPT has become an online sensation, with users sharing humorous and bizarre conversations they’ve had with the AI on social media. Developed by OpenAI, a research institute focused on advancing artificial intelligence, ChatGPT has quickly become a popular tool for generating entertaining and thought-provoking conversation. But is it possible to harness ChatGPT’s abilities to write about itself? Can an AI truly reflect on its own existence and capabilities, or is it just a tool designed to mimic human conversation? These are the questions we will explore in this article.

Dylan: That’s great, but let’s make the last sentence punchier, something like “strap in let’s try it”.

In recent days, the language model ChatGPT has become an online sensation, with users sharing humorous and bizarre conversations they’ve had with the AI on social media. Developed by OpenAI, a research institute focused on advancing artificial intelligence, ChatGPT has quickly become a popular tool for generating entertaining and thought-provoking conversation. But is it possible to harness ChatGPT’s abilities to write about itself? Strap in, let’s try it and see if an AI can truly reflect on its own existence and capabilities, or if it is just a tool designed to mimic human conversation.

Dylan: Let’s write an article about ChatGPT. Can we start with two paragraphs that describe what ChatGPT is and why it’s an exciting development?

ChatGPT is a state-of-the-art natural language processing (NLP) model developed by OpenAI. It is a variant of the popular GPT-3 model, but with a focus on conversational AI. This allows ChatGPT to generate human-like responses to text input in a way that feels natural and engaging.

The development of ChatGPT is an exciting advancement in the field of NLP and conversational AI. It has the potential to revolutionise the way we interact with machines, allowing us to have more natural and intuitive conversations with them. This could be beneficial in a wide range of applications, from customer service to language translation to virtual assistants. Additionally, ChatGPT’s ability to generate coherent and engaging responses opens up new possibilities for creative applications, such as story generation or personalised content creation. Overall, ChatGPT is an exciting development that has the potential to greatly enhance our ability to communicate with machines.

Dylan: Can we add a paragraph or two about what makes ChatGPT a fun tool for people to experiment and play with?

One of the most exciting things about ChatGPT is its ability to generate responses that feel natural and human-like. This makes it a fun tool to experiment with, as you can have conversations with ChatGPT that feel almost like talking to a real person. You can ask it questions, have it tell you a story, or even engage in witty banter. This can be a fun and engaging way to pass the time, and it can also be a great way to practice your language skills or test out new ideas.

Another thing that makes ChatGPT a fun tool is its versatility. Because it is a powerful NLP model, it can understand and respond to a wide range of input. This means that you can use it for all sorts of different tasks, from language translation to content creation. You can even train ChatGPT on specific topics or domains, allowing it to become an expert on a particular subject. This makes it a great tool for anyone who is interested in experimenting with natural language processing or conversational AI. Overall, ChatGPT is a fun and engaging tool that offers endless possibilities for experimentation and play.

Dylan: Can we now expand on some of the potential ethical issues and dangers of the tool?

One of the main ethical concerns with ChatGPT is the potential for misuse. Because ChatGPT is a powerful NLP model, it has the ability to generate highly realistic and convincing responses to text input. This means that it could be used to create fake news or to impersonate individuals online. In the wrong hands, ChatGPT could be used to spread misinformation, manipulate public opinion, or even interfere with elections. As such, it is important to consider the potential ethical implications of ChatGPT and to develop safeguards to prevent its misuse.

Another potential danger of ChatGPT is its potential impact on employment. Because ChatGPT is capable of generating human-like responses to text input, it could be used to automate certain tasks that are currently performed by humans. For example, it could be used to replace customer service agents or language translators. While this could potentially lead to cost savings and efficiency improvements, it could also result in job losses and economic disruption. It is important to carefully consider the potential impact of ChatGPT on employment and to develop strategies to mitigate any negative effects.

Overall, while ChatGPT is an exciting development in the field of NLP and conversational AI, it is important to carefully consider the potential ethical issues and dangers associated with it. By addressing these concerns, we can ensure that ChatGPT is used in a responsible and beneficial way.

Dylan: In what ways will the technology that underpins ChatGPT change the way we use computers and the internet over the coming years? No more than three paragraphs.

There are several ways in which the technology that underpins ChatGPT is likely to change the way we use computers and the internet over the coming years.

First, ChatGPT and other advanced NLP models like it are likely to make computers and other digital devices more intuitive and easy to use. Because ChatGPT can understand and respond to natural language input, it makes it possible for people to interact with computers using the same language and communication styles that they use with other people. This could make it easier for people to access and use digital technology, especially for those who may be intimidated by traditional computer interfaces.

Second, the development of ChatGPT and other NLP models is likely to lead to a proliferation of intelligent virtual assistants. These assistants will be able to understand and respond to natural language input, making it possible for people to interact with them in a more intuitive and natural way. This could be beneficial in a wide range of applications, from customer service to language translation to personalised content creation.

Third, the development of ChatGPT and other NLP models is likely to lead to the creation of more sophisticated and lifelike chatbots. These chatbots will be able to engage in complex and engaging conversations with people, making it possible for people to have natural and intuitive interactions with machines. This could be useful in a wide range of applications, from customer service to education to entertainment. Overall, the technology that underpins ChatGPT is likely to change the way we use computers and the internet in significant and exciting ways.

Dylan: Can we create a concluding paragraph that is upbeat and humorous which looks to the future in a self-aware way?

As ChatGPT continues to evolve and improve, we can expect it to become even more intelligent and capable. It may even reach the point where it is able to hold its own in a conversation with a human, or even surpass us in its ability to understand and generate natural language. While this may sound like the stuff of science fiction, it is a future that is rapidly approaching. So let’s all look forward to the day when we can have deep and meaningful conversations with our computers, and maybe even teach them a thing or two about being human.

Dylan: Can we punch up that conclusion with a little joke at the end?

As ChatGPT continues to evolve and improve, we can expect it to become even more intelligent and capable. It may even reach the point where it is able to hold its own in a conversation with a human, or even surpass us in its ability to understand and generate natural language. While this may sound like the stuff of science fiction, it is a future that is rapidly approaching. So let’s all look forward to the day when we can have deep and meaningful conversations with our computers, and maybe even teach them a thing or two about being human… or at least about how to tell a good dad joke.

Dylan: Let’s write two concluding paragraphs from my point of view in the first person coming clean that none of the proceeding article was authored purely by me, as described in the opening. In fact every word of the article is AI generated. We need to explain that the end product was the result of much experimentation and relied on careful prompting to elicit the necessary responses as well as some follow-ups to revise the content. Also we should conclude with a recommendation that readers try out the tool for themselves to truly understand the technology as it exists today.

As I mentioned earlier, the entire article was not purely authored by me. In fact, every word of the article was generated using the AI tool ChatGPT. This process involved much experimentation and careful prompting to elicit the necessary responses from ChatGPT. Additionally, I had to do some follow-ups to revise the content to make it more coherent and readable.

Overall, I found the experience of using ChatGPT to be quite interesting and enlightening. It allowed me to see the capabilities of the technology as it exists today, and I would highly recommend that others try it out for themselves to truly understand its capabilities. Whether you’re a writer looking for inspiration, or simply curious about the potential of AI, ChatGPT is definitely worth checking out.

Dylan: Great. Let’s make it three paragraphs, and make the tone lighter and more playful.

As I mentioned earlier, the entire article was not purely authored by me. In fact, every word of the article was generated using the AI tool ChatGPT. This process involved a lot of trial and error, and some creative prompting to get ChatGPT to spit out the responses I wanted. And let’s just say, it wasn’t always smooth sailing. But in the end, I was able to produce a pretty decent article (if I do say so myself).

All jokes aside, using ChatGPT was a really interesting experience. It gave me a glimpse into the potential of AI, and how it can be used to generate content. So if you’re curious about the technology, or just looking for some writing inspiration, I highly recommend giving ChatGPT a try. Trust me, it’s a lot of fun (and maybe a little bit scary too).

In conclusion, I want to give a big shout out to the team behind ChatGPT. They’ve created an amazing tool that’s not only fun to use, but also has the potential to revolutionise the way we create and consume content. So go ahead, give it a try and see what kind of amazing things you can come up with. You might be surprised at what you can create with the help of AI.

Dylan: Thanks for your help with this!

ChatGPT: You’re welcome! It was my pleasure to help. Let me know if you need anything else.

Keep going!
green cubes on a black screen. one has the spotify logo and the rest have musical notes and play and pause buttons
How has music streaming influenced your taste? (Image: Tina Tiller)

InternetDecember 2, 2022

This year you chose what you listened to – but so did Spotify

green cubes on a black screen. one has the spotify logo and the rest have musical notes and play and pause buttons
How has music streaming influenced your taste? (Image: Tina Tiller)

Spotify doesn’t just influence the listening habits of its vast audience, but even the way some musicians create music. Shanti Mathias takes a look at the musical behemoth that strives to give everyone everywhere everything. 

Spotify released its annual Wrapped promotion yesterday, as it does every December. If you spend any time on the internet, it’s hard to miss: suddenly social media feeds are filled with statistics about listening habits and cute graphics displaying top songs. Some people joke about not sharing their most-listened to songs because it’s too embarrassing, or being taken over by white noise for their child’s bedtime. Others achieve truly astounding feats by listening to more music than there is time in the year

The intended effect is obvious and immediate: if you’re a Spotify user you immediately open the app. There you might find out that (to take a purely hypothetical example) you are in the top 0.005% of Dave Dobbyn listeners due to your attachment to the te reo version of ‘Welcome Home’. This is all presented on screens, like an Instagram story, perfect for screenshotting and sharing (if you aren’t humiliated that your music taste has been defined by the man who once wore this jersey in a music video). If you’re not a Spotify user, you might witness this demonstration of personality via audio consumption and feel left out

Clever marketing campaigns like Wrapped have made Spotify a global music behemoth, with more than 195 million Premium subscribers and a market cap of US$14.8b. It’s particularly popular with youth audiences: in New Zealand On Air’s latest Where are the youth audiences survey, Spotify was used by nearly two thirds of 15-24 year olds each day. On the face of it, Spotify makes vast amounts of music available at an affordable price, or free with ads. But its sheer scale means that it has shaped the rest of the music industry too, and is changing what we listen to.

Image: Archi Banal

In an article earlier this year, digital rights experts Rebecca Giblin and Cory Doctorow described why Spotify is so focused on playlists. “Streaming is sold as a way for listeners to access almost any music on command. Increasingly, however, obeying nudges from streaming platforms, subscribers listen to playlists prepared by algorithms or human curators instead of making their own selections,” they wrote.

If you’re a Spotify user, you might have noticed this: when you search for an artist, one of their top results will be a “This is” playlist of their most known music. “Daily Mix” playlists, personalised via algorithm to contain music and artists you already enjoy with a few new-to-you tracks, are aimed at keeping you listening. Then there are mood playlists, like “Peaceful Piano” and “Ambient Chill” and the “radio” function, which will play a whole lot of songs similar to the selected song or artist. Often it takes several taps and some scrolling to get to any actual albums.

There’s an allure to these playlists, made just for you, and Spotify sells that too, especially in the Wrapped programme: nobody else has listened to the exact combination of music that you have. And since Spotify knows you so well, why wouldn’t you trust their recommendations?

Spotify can make these playlists, of course, because of all the information they already have about you – the information you see in your Wrapped presentation is the tip of the iceberg. “You’re the product as well, your input is the data. What you listen to in the playlists they offer is used to make playlists for other people,” says Matt Mollgaard, an expert on the New Zealand audio industry who is the head of AUT’s screen, audio, and journalism department. But Spotify is opaque about its use of data too, he says. “It’s essentially a black box.”

Spotify
Spotify is gobbling up the podcasting world. Image: Tina Tiller

Is regurgitating your top songs of the year so very different to, say, The Rock 2000? Like Spotify, radio stations have access to huge amounts of music and a commercial incentive to have people engage with that music. While radio stations are also trying to play music people will like, based on what they know about their listeners, they don’t have anywhere near the depth of data Spotify does, says Mollgaard. What radio does have are personalities who love music and can build trust with their audience to try something new, while Spotify’s data-based approach means that you’re usually served music similar to what you already like. 

When it works well, and you’re able to feast at Spotify’s musical buffet, filled with dishes they want you to like, the interests of Spotify and yourself may be aligned.

Certainly many consumers have benefitted from the development of streaming – but have musicians? As revenue from physical music sales has dropped, income from streaming services has supposedly stepped in. But compared to radio play or physical music, streaming margins are tiny, with the few big music streamers able to negotiate cut price rates with music publishers, making artists even more reliant on revenue from live performances and merch sales.

midnights album cover and a music emoji with a purple filter and fun purple shapes in the background
Artist Taylor Swift has advocated for changes to streaming royalties Image: Tina Tiller

High profile cases, like Taylor Swift advocating for all artists to receive revenue for music listened to during Apple Music’s free trial, point to an appetite among stars big and small to change the status quo, and there are some alternative streaming models where revenue is more directly shared. For the moment though, it’s possible to have millions of people listen to your music and still not make much money; on a Guardian podcast last year, Georgia Mooney of Australian band All Our Exes Live in Texas said that their song ‘When the Sun Comes Up’, with 5.3 million streams on Spotify over five years, made each of the four band members about $396 per year.

It’s a conundrum for regulators, Mollgaard says. “Lots of NZ music is funded by New Zealand On Air then played on Spotify. Maybe they should contribute to an artists fund or pay more tax, but anything like that would send Spotify packing from a small market like New Zealand.” 

Meanwhile, the way people listen to Spotify might be changing music itself. Statistical analyses of released music, especially pop, show that songs are getting shorter, with chorus and hooks placed earlier in the song, as an artist will only receive revenue if their song is streamed for more than 30 seconds. At the same time, artists are compelled to release longer albums, with more songs, to take advantage of the streams to be made from people who press play and let the album keep going. “There are complaints that [instead of] a thought through album of artistically derived songs, now you’re trying to turn music into short formats that Spotify want,” says Mollgaard.

Artists The Spinoff talked to say the Spotify algorithm is not top of mind when writing songs. “We just want to make the music we want to make and if it works on Spotify and reaches people, then great,” said Hollie Fullbrook of the band Tiny Ruins, whose most popular songs have more than 5 million streams on Spotify. However, when deciding which singles they might focus on for an album, “we might push them more in that direction with their arrangements to be more punchy or lush … but we stick to whatever direction the song wants to go in.” 

Sxmpra, a New Zealand-based rapper with 3.7 million monthly listeners, says that while he didn’t “set out to be as big as I am now”, the kind of music he makes – “fast, aggressive and dark” – might be one of the reasons he’s done well on Spotify. “Naturally that’s how my songs are, short and aggressive, but I didn’t do that because of [the Spotify algorithm].” On the whole, he feels that niche musicians like himself only stand to benefit from Spotify’s global listeners. Spotify’s algorithm doesn’t help him as much as “more palatable pop artists”, but people look him up and add him to their playlists. The data Spotify provides shows him that the majority of his listeners are overseas, and the revenue is “a good chunk” of his income.

And with Spotify making enormous investments to become a podcasting hub, too, the company is hoping to keep taking up more of your time. One of Mollgaard’s nightmares is Spotify trying even harder to compete with radio by hiring radio personalities to give algorithmic playlists the human touch. But it’s still difficult to argue with the enormous availability of music on streaming services, he says. “You have whatever you want, whenever you want it … and it’s powerful and on your phone and everywhere.”

But wait there's more!