Roundhill Investments has launched a new exchange-traded fund (ETF) focused on the burgeoning generative AI sector.
Background info
Roundhill is a registered investment adviser and ETF issuer offering thematic and sector-specific funds.
This latest fund, trading under the ticker symbol ‘CHAT’ on the NYSE Arca, consists of holdings in 30 AI- and technology-focused companies.
Those companies are selected across different classifications, including AI platforms (companies developing their own generative AI models and tools), as well as infrastructure providers which supply hardware specifically tailored to AI workflows.
The fund also includes both enterprise and consumer software suppliers, which are building and commercialising generative AI applications for their customers.
What’s in the fund?
Roundhill said the fund is the first ETF globally designed to track the performance of companies involved in generative AI.
The fund’s largest holding is in Nvidia Corporation, a technology supplier which describes itself as the “world leader in artificial intelligence computing”.
The second and third largest holdings in the ETF are Microsoft and Alphabet (which owns Google), followed by processor specialist AMD and software giant Adobe.
Other well-known holdings include Meta, Amazon, Tencent and Snap, as well as Shutterstock, Intel, Salesforce and Alibaba.
The fund also includes several less broadly known tech companies, including “pure play AI companies,” such as the long standing Chinese IT firm iFlyTek, enterprise AI software supplier C3ai and Hong Kong-based AI developer Sensetime.
Why invest in AI?
Roundhill suggested that the total addressable market (TAM) for generative AI enterprise software is estimated at around $121bn.
Goldman Sachs, meanwhile, estimates that AI has the potential to drive some $7tn in global economic growth over the next decade.
OpenAI’s ChatGPT has helped push the conversation around AI technology into the mainstream in recent months, with the service reaching 100 million users in just two months after launching, making it one of the fastest ever applications to reach the milestone.
The CHAT ETF joins several others in Roundhill’s wheelhouse, including the sports betting-focused fund BETZ, which holds shares in more than 40 publicly traded sports betting and iGaming businesses including DraftKings, Flutter and Evolution.
In March, Roundhill also launched its BIGB banking ETF, which consists of just six “too big to fail” US bank stocks.
Hot off the heels of the release of ChatGPT in November 2022, Silicon Valley-based startup OpenAI has released ChatGPT 4, the latest iteration of its artificially intelligent large language model (LLM).
What is ChatGPT? (As if you didn’t know)
ChatGPT has been considered a major step forward in AI technology since its release last year, due to the apparently sophisticated ways in which it can respond to users’ queries.
From writing articles and essays to coming up with poems and song lyrics, writing functional computer code and solving complex mathematical problems, the programme has a broad range of language-based capabilities.
ChatGPT hit an estimated 100 million monthly active users in February, according to Forbes, making it the fastest growing consumer internet application in history.
The LLM has taken the world by storm and helped bring AI applications thoroughly into the mainstream. Its appearance on the cover of TIME magazine last month went to show how the app has rapidly cemented itself front-of-mind for a huge number of businesses and individuals alike.
So, what’s new?
The programme’s latest iteration, ChatGPT 4, brings with it an improved level of sophistication which allows the app to perform increasingly complex tasks and respond to queries with a greater level of accuracy.
According to OpenAI, the app is “more creative and collaborative than ever before,” and significantly outperforms previous iterations on several benchmark tests.
One salient example is its performance in examinations intended to be taken by humans.
When asked to take the Uniform Bar Exam, which is designed to test knowledge and skills that lawyers should be able to demonstrate prior to becoming licensed to practise law, previous iterations of ChatGPT produced results in the bottom 10% of performers.
ChatGPT 4, meanwhile, performs in the top 10%, showing how far the programme’s capabilities have come in a short space of time.
One major concern around the use of LLMs is their tendency to offer up inaccurate information as if it were fact.
ChatGPT 4 is 40% more likely to produce factual responses than the app’s most recent prior iteration, GPT 3.5, according to OpenAI.
However, the knowledge cutoff date for this latest iteration is still September 2021, so some information may be out of date.
Use cases
One especially significant development is ChatGPT 4’s ability to understand image-based inputs as well as text.
That capability brings about innovations such as – perhaps somewhat trivially for most users – the ability to take a picture of a cupboard full of ingredients, and have ChatGPT recommend recipes based on what’s available.
In a more significant and practical scenario, ChatGPT 4 has now been integrated into Be My Eyes, a mobile app which pairs up visually impaired users with human volunteers, who can see a video feed from the visually impaired user’s phone camera and give instructions such as which product to select from a supermarket shelf.
Now, ChatGPT will allow Be My Eyes to integrate a virtual volunteer, which will interpret visual information and give instructions to users without the need for human interaction.
“The implications for global accessibility are profound,” said Michael Buckley, CEO of Be My Eyes.
“In the not so distant future, the blind and low vision community will utilise these tools not only for a host of visual interpretation needs, but also to have a greater degree of independence in their lives.”
What about the gambling sector?
ChatGPT 4 will bring about use cases across perhaps all industries, and should help business owners and employees alike to simplify and automate a variety of tasks.
From helping to write content for affiliate websites, for example, or coordinating diaries to find the best meeting time for a group of colleagues, ChatGPT can work as a virtual personal assistant, and much more.
The determining factor in how useful the app can be will ultimately come down to its users and their ability to properly prompt or command ChatGPT.
Still, OpenAI’s demonstration of the programme yesterday (14 March) also gave a dystopian glimpse into what the future could hold in terms of its use cases.
The firm’s president and co-founder Greg Brockman fed the programme a photo of a hand-drawn, rudimentary sketch of a website, from which ChatGPT automatically created a functional website coded in JavaScript.
The implications of such capabilities could be vast for those in technologically-driven industries such as iGaming.
What are the drawbacks?
OpenAI has not shied away from the fact that ChatGPT remains far from perfect.
While improvements have been implemented in its latest iteration, the LLM still cannot match human reasoning or fact-checking skills, and remains liable to provide information which is inaccurate or out of date.
While it can be trusted to carry out simple tasks with clear instructions, more complex requests may generate less-than-satisfactory responses, or answers which are just downright wrong.
Still, with a $10bn investment from Microsoft earlier this year and the programme now integrated into the software giant’s Bing search engine, it seems that the only way is up for AI-powered LLMs, and their impact on our daily lives will only become more visible from here.
Ethical concerns are more difficult to overlook, however.
Conclusions
AI capabilities are expanding exponentially. Ask anyone to guess where the capabilities of programmes like ChatGPT will be in 10 years’ time, and they are almost guaranteed to get it wrong.
The rate of acceleration of such technologies is only going to increase from here, as an AI revolution sweeps across the globe.
Perhaps the processing power of AI – which of course can already outperform any human in any number of tasks – will bring about fundamental changes to our society, daily lives, and the way we work, and all much sooner than we think. Or perhaps it won’t.
Unlike some technological advancements (NFTs, anyone?), businesses and individuals would do well to keep a close eye on AI and its beneficial use cases.
While no-one quite knows what the world will look like a decade from now, we can say with relative certainty that this is only the beginning for AI.