Dear friends,
Since OpenAI released ChatGPT to the public, there has been an interesting debate online about the implications of large language models (LLMs) for the broader tech ecosystem. One of the implications often cited is that LLMs such as ChatGPT will make Google search obsolete. Indeed, there are signs that senior executives at Google have been in panic mode. Not least, Alphabet's stock has been under pressure, and all the buzz around ChatGPT threatening Google's core revenue stream certainly does not help.
Interestingly, the CEO, Sundar Pichai, recently announced their ChatGPT competitor, Bard. However, the launch of Bard was less impressive than ChatGPT. Soon after the public announcement, people started sharing embarrassing shortcomings of Bard, including factual errors in Google's promo video of Bard itself!
Mathematical Performance of LLMs
Like our team, are you curious whether LLMs such as ChatGPT can perform well enough on mathematical-related tasks? Fortunately, a multidisciplinary team of researchers recently published a preprint article entitled 'Mathematical Capabilities of ChatGPT (https://arxiv.org/pdf/2301.13867.pdf). You may find the following quote extracted from the abstract interesting:
We conclude that contrary to many positive reports in the media (a potential case of selection bias), ChatGPT's mathematical abilities are significantly below those of an average mathematics graduate student. Our results show that ChatGPT often understands the question but fails to provide correct solutions. Hence, if your goal is to use it to pass a university exam, you would be better off copying from your average peer!'
Many people have been super positive online about LLMs to the point that they see a possible pathway to the holy grail of AI known as 'Artificial General Intelligence' (AGI). We are very sceptical that LLMs will indeed provide the path for AGI precisely because of its poor mathematical reasoning performance. In our humble opinion, a true AGI will have unprecedented advanced mathematical reasoning capabilities!
Finally, here are some resources on ChatGPT that you may find interesting to read:
[1] A Categorical Archive of ChatGPT Failures.
[3] People tricking ChatGPT "like watching an Asimov novel come to life".
[4] Wolfram Alpha and ChatGPT.
[5] Microsoft in talks to acquire a 49% stake in ChatGPT owner OpenAI.
[6] ChatGPT won't replace search engines any time soon.
[7] Comparing Google and ChatGPT.
[8] ChatGPT produces made-up nonexistent references.
[9] ChatGPT passes the 2022 AP Computer Science A free response section.
Many thanks for reading. We look forward to sharing our deep-tech community & venture building journey with you.
Zaiku Group team
{Twitter} {LinkedIn}
Disclaimer: Any opinions, newsletters, research, analyses, prices, projections or other information offered by ZAIKU GROUP is provided as general market commentary, and does not constitute investment advice. ZAIKU GROUP will not accept liability for any loss or damage, including without limitation to, any loss of profit, which may arise directly or indirectly from use of or reliance on such information.