Instructions to upgrade:<p><pre><code> 1. Log into chat.openai.com
2. Click on "Upgrade plan" on the left slide menu
3. "Your account" popup appears with an option to upgrade to a "Professional Plan" for $42/mo, or stay on the "Free Plan"
</code></pre>
Free Plan benefits:<p><pre><code> - Available when demand is low
- Standard response speed
- Regular model updates
</code></pre>
Professional Plan benefits:<p><pre><code> - Available even when demand is high
- Faster response speed
- Priority access to new features</code></pre>
When asking ChatGPT to write code in language X to do task Y, often it works, for simple Y, but sometimes it produces code that does not even compile. For a professional version, why not integrate ChatGPT with compilers such as g++, so that if it is asked to produce C++ code, it does not return code that does not compile (or at least warns that the code is not compilable). You could go further and have it compile and execute programs and verify that they do what the user wants.
I find it quicker then a google search for things like generating code to initialize a list in a single line in Java when I forgot about Arrays.asList or creating a skeleton rest service in Springboot so I don't have to remember the exact annotations. I won't use it on my work computer though. It's not worth $42 for personal use but if I thought I could get my company to get me the professional plan I would.
People in academia were super excited about chatGPT. But very quickly we found out that it just makes up stuff such as citations. It literally creates a bibliography of fake papers! Not sure anyone I know will pay anything for it even though our research budgets can easily afford this subscription.
I had the same negative reaction to the price, but that's as a hobbyist/explorer. Given the number of people launching products that sit on top of ChatGPT, I understand the high price point, as much as any pre-trained large language model can be said to 'understand' something.
It's a great product. Some day. But now that the honeymoon is over, it needs a whole lot more going on to justify that kind of price. Come back when there's more improvements than a couple months of a public alpha behind it.