5 Creative uses for AI - Generative AI pt. 3
An AI chatbot that gives tax advice
Depending on where you live, taxation is a complex and multi-faceted process where you have to know the ins and outs of the system. Do you need to file crypto purchases? What can you deduct as personal expenses? A former colleague of mine, Mr. Tianyi Pan, made an AI chatbot that can answer questions regarding taxation. The kysykarhulta.com service is only available for the Finnish tax system, but Mr. Pan has previously posted a guide on how to build custom AI chat services yourself, so perhaps you can create one for yourself.
A special mention goes to CatGPD. Another generative chat service built by Mr. Pan that gives replies in cat speak with paired cat portraits.
Make ChatGPT a CEO with 1,000$ capital and 1h/week.
One of LinkedIn’s most followed story arcs has been that of João Ferrão dos Santos. He asked Chat GPT to act as a CEO with US$ 1,000 of start-up capital and 1h of daily work with one goal: Build a business as profitable as possible in 30 days.
The first post detailing the decisions made by ChatGPT got more than 2 million views in one day. The following 29 days detailed the journey of Mr. Santos setting up a clothing store using Shopify, looking for investors, touring the world meeting business people and media representatives. At the end of the 30 days, Mr. Santos posted some numbers regarding the project:
8,538,906 - impressions on LinkedIn.
120,000 - capital raised from investors (EUR).
39,541 - new followers.
12,535 - in sales (fixed double-counting error).
6,000 - connection requests.
2,678 - Substack subscribers, 59% open rate.
~90 - guests in our AI meetups.
~40 - requests to invest in Aisthetic.
~30 - offers to work with us for free.
~20 - requests to start a franchise.
14 - podcast + keynote speaker invites.
5 - projects born from our post.
4 - requests to write a book on this.
3 - law firms guiding us.
2 - features in top media publishers.
1 - feature in a documentary on AI.
1 - ask for permission to pitch Netflix.
1 - fresh LLC created in Denmark.
1 - mother building a company with her son
Ace a job interview with the help of an AI interviewer
Eric Gu is an engineer and an AI enthusiast working on a doctorate in AI creativity, AR/VR, and conversational interfaces.
He recently applied to the Antler Accelerator in Amsterdam, and as part of the application process, he built a prompt to have ChatGPT simulate the interview process. To create the prompt, he used a tool he has co-created called Promptly, which allows efficient prompt creation for generative AI.
He provided an AI-generated summary of the company’s investment philosophy in the prompt and a format for the questions based on the interview invite. He asked the prompt to ask one question at a time and give feedback on the answers and the reasoning behind each question which helped him to ace the interview.
Eric has also been working on an AI project called AskBria. This AI-powered chatbot is a virtual doula designed to provide expecting parents personalized support and guidance throughout pregnancy and the birthing process! Read more at www.askbria.fyi
Use AI to filter out background noise on calls.
SONY announced the latest version of the popular WF-1000 XM series headphones this summer, and one of the impressive additions was how SONY had solved the problem of background noise. I have a pair of the headphones’ predecessors, the WF-1000 XM4, and they are almost unusable for calls in busy streets or in windy situations. The noise from cars and people gets picked up by the microphones, and the other side cannot hear my voice.
With the WF-1000 XM5, SONY appears to have solved this using two methods: Firstly, the headphones use bone-conducting sensors to detect when you are speaking. Secondly, SONY used AI machine learning to analyze 500 million voice samples to separate a human voice from environmental noise. I have not yet had the pleasure to try these puppies out, but I look forward to them once they are available in Hong Kong.
Creative solutions on how AI can filter out background noise from calls through machine learning by SONY
Create selling points based on competitor reviews.
I went to an e-commerce meetup and met a person working in the Venture Capital/e-commerce space. We ended up discussing how AI is changing e-commerce, and he told me of a colleague of his who utilized ChatGPT to crunch through customer reviews to find out what customers thought of a competitor’s product. What they liked and disliked.
Here is where I toot my own horn as I blatantly copied the idea and used ChatGPT to do that. A well-established competitor of a client of mine has hundreds of reviews on Amazon. While I could have spent 30-60 minutes reading through a bunch of them, I had ChatGPT spit out a quick five-point summary of likes and dislikes for the product. I then used this summary to formulate the structure of the product page with clear, fast-to-read titles that matched the key areas that people were praising and identifying how the newly launched product was better than the competitor in the areas that people were not happy with. Especially now that ChatGPT has access to the internet and is no longer limited by the September 2021 timeline, ChatGPT can help with similar research.