Summary
An incorrect analysis by the investment bank Goldman Sachs on the collapse in traffic at ChatGPT caused unrest on the stock markets. But the data was misinterpreted. Demand for OpenAI’s services continues to grow strongly.
A report by Peter Oppenheimer, an analyst at Goldman Sachs, has caused a stir in financial circles. According to a graphic contained in the report, data traffic to OpenAI’s popular chatbot ChatGPT has plummeted in recent months. The graphic found its way into the Financial Times and contributed to the poor sentiment surrounding AI stocks.
The faulty graphic from Goldman Sachs shows a drop in traffic to the ChatGPT website because a domain change was not taken into account. | Image: Goldman Sachs
Share article
Recommend our article
Split
But it was an embarrassing mistake by the renowned investment bank. Goldman Sachs had collected data from Similarweb for the analysis, but did not take into account that OpenAI had recently changed ChatGPT’s domain from chat.openai.com to chatgpt.com. Traffic to chat.openai.com dropped accordingly.
In fact, Similarweb offers a monthly updated analysis of the use of AI services on its own blog. The most recent evaluation from the beginning of August shows a completely different picture: ChatGPT recorded year-on-year growth of 66.2 percent and remains by far the most popular application for generative AI. Competitors such as Anthropics Claude and Perplexity are catching up, but have not yet been able to catch up with ChatGPT.
Advertisement
THE DECODER Newsletter
The important AI news directly to your email inbox.
✓ 1x weekly
✓ free
✓ Can be cancelled at any time
ChatGPT records rapid increase in global user numbers: The graph shows an impressive growth trend in desktop and mobile visits from November 2022 to July 2024, with a peak of almost 2.5 billion visits in June and July. | Image: Similarweb
Other indicators also point to continued demand for OpenAI’s services. The company itself announced a few days ago that To have 200 million active weekly users – a doubling since November 2022. Added to this is a growing business in the enterprise sector as well as the use of OpenAI models via APIs and integration with Microsoft.
Declining AI demand is not OpenAI’s problem
Table of Contents
Table of Contents
A recently published analysis shows that OpenAI could generate revenue of between $3.5 billion and $4.5 billion this year – a remarkable achievement for a company that has only been seriously commercially active since the end of 2022. However, the revenue will be offset by costs of up to $8.5 billion.
So OpenAI may have challenges to overcome, from sometimes naive communication to frustrated security researchers to high expenses, but a demand problem is clearly not one of them.
Goldman Sachs’ AI Blunder: Misinterpreting ChatGPT’s Traffic Data
In a recent report, Goldman Sachs analyst Peter Oppenheimer caused a stir in financial circles by suggesting that traffic to OpenAI’s popular chatbot ChatGPT had plummeted in recent months [[1]]. The graphic, which was featured in the Financial Times, contributed to the poor sentiment surrounding AI stocks. However, it has come to light that Goldman Sachs made a critical error in their analysis, leading to a misinterpretation of the data.
The Mistake: Ignoring Domain Name Change
Goldman Sachs collected data from Similarweb for their analysis, but failed to take into account that OpenAI had recently changed ChatGPT’s domain from chat.openai.com to chatgpt.com [[3]]. As a result, traffic to chat.openai.com dropped accordingly, leading to the mistaken conclusion that ChatGPT’s traffic was declining.
The Real Picture: ChatGPT’s Continued Growth
In reality, Similarweb’s analysis shows a completely different picture. ChatGPT recorded year-on-year growth of 66.2% and remains the most popular application for generative AI by a significant margin [[1]]. Competitors such as Anthropics Claude and Perplexity are catching up, but have not yet been able to surpass ChatGPT.
Indicators of Continued Demand for OpenAI’s Services
Other indicators also point to continued demand for OpenAI’s services. The company itself announced that it has 200 million active weekly users, a doubling since November 2022 [[1]]. Additionally, OpenAI’s growing business in the enterprise sector and the use of its models via APIs and integration with Microsoft further highlight the company’s growth trajectory.
Goldman Sachs’ AI Experimentation
Interestingly, Goldman Sachs has been experimenting with ChatGPT-like AI tools to help its
How has Goldman Sachs’ analysis on AI and job losses been misinterpreted in the context of employment and economic impact?
The Impact of AI on Employment: A Misinterpreted Analysis by Goldman Sachs
A recent report by Goldman Sachs has caused a stir in financial circles, suggesting that ChatGPT, a popular chatbot, and other AI systems could threaten 300 million jobs worldwide [1]. However, this analysis has been called into question, and its findings may have been misinterpreted.
According to the report, ChatGPT and similar AI systems could contribute to a 7% increase in GDP, but at the cost of massive job losses [2]. This prediction has sparked concerns about the impact of AI on employment, with many wondering if the benefits of technological advancement outweigh the costs.
However, a closer look at the data reveals that the analysis by Goldman Sachs may have been flawed. The report failed to account for a domain change made by OpenAI, the company behind ChatGPT, which led to a drop in traffic to the old domain [3]. This oversight has led to an incorrect interpretation of the data, with ChatGPT’s traffic actually experiencing strong year-on-year growth of 66.2% [similarweb[similarweb].
The misinterpretation of data has significant implications for our understanding of the impact of AI on employment. While AI systems like ChatGPT are undoubtedly changing the nature of work, the extent of this change may not be as drastic as initially suggested. In fact, AI has the potential to create new job opportunities and improve productivity, rather than simply replacing human workers [[World Economic Forum](https://www.weforum.org/agenda
